z . .... ”...". 2

June... 1116
Owl-Fwy
. #33“

‘41. . Huh“,
7 2..
, Lunar
a

 

 

 

 

 

{\I 7‘",
...... 4.;
1.11”»... z-..
Juno)» .... .1
.' ”up. ‘ u. .Ijl ‘
Wayrﬁ...
.0 .‘l

..., 2 1.. u.

..V ...L
.1.

5.....1

 

 

 

LIBRARY

Michigan State
University

 

 

 

This is to certify that the
dissertation entitled

Verification of a Monoclonal Antibody-Based

ELISA for Monitoring Lipid Oxidation in Chicken

Myof ibrils . presented by

Carolyn Felicity Ross

has been accepted towards fulﬁllment
of the requirements for

Ph.D. degree in Food Science

 

JW Jﬂuﬂi,

Major professor

 

Date December 21 2001

 

MS U is an Afﬁrmative Action/Equal Opportunity Institution 0-12771

 

PLACE IN RETURN Box to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cJClFiC/DateDuep65op. 15

 

VERIFICATION OF A MONOCLONAL ANTIBODY-BASED ELISA FOR
MONITORING LIPID OXIDATION IN CHICKEN MYOFIBRILS

By

Carolyn Felicity Ross

A DISSERTATION
Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of
DOCTORATE OF PHILOSOPHY
Department of Food Science and Human Nutrition

2001

UMI Number: 3053802

®

UMI

 

UMI Microform 3053802

Copyright 2002 by ProQuest Information and Learning Company.

All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.

 

ProQuest Information and Learning Company
300 North Zeeb Road
PO. Box 1346
Ann Arbor, MI 48106-1346

ABSTRACT

VERIFICATION OF A MONOCLONAL ANTIBODY-BASED ELISA FOR
MONITORING LIPID OXIDATION IN CHICKEN MYOFIBRILS
By

Carolyn Felicity Ross

Lipid oxidation is a major quality deterioration problem in all muscle foods,
affecting both the organoleptic value and toxicological status of the food. Detection of
lipid oxidation needs to be accomplished in a very speciﬁc, sensitive and rapid manner.
The standard method for measuring lipid oxidation is the 2-thiobarbituric acid reactive
substances (TBARS) assay. While this assay is simple, criticisms arise from its lack of
speciﬁcity. Hexanal. a secondary product of lipid oxidation, has often been used as an
indicator of lipid oxidation. Headspace-gas chromatography has typically been used to
measure hexanal; however, this method is expensive and tedious. As a sensitive, rapid
and reproducible alternative, an enzyme-linked immunosorbent assay (ELISA) has been
previously described for the detection of hexanal in meat products. The objective of this
study was to verify the use of the monoclonal ELISA for monitoring lipid oxidation in
chicken protein through the quantiﬁcation of hexanal.

A solid-phase microextraction (SPME) method combined with gas
chromatography/mass spectrometry (GC/MS) was developed and evaluated for its
suitability in the detection of hexanal in chicken myoﬁbrils. Results indicated that

recovery of hexanal from spiked myoﬁbrils was 95% using a 5 min sampling time. with a

total analysis time of about 12 min/sample. Using SPME, the limit of detection for
hexanal in myoﬁbrils was 10 ppb with a coefﬁcient of variation ranging from 1 to 13%.
and a working linear response from 10 ppb to 10 ppm (r3=0.995). The ability of the
hexanal-speciﬁc monoclonal CI-ELISA to monitor lipid oxidation in freeze-dried chicken
protein was compared to the thiobarbituric acid reactive substances assay (TBARS) and
the solid-phase microextraction- gas chromatography/mass spectrometry (GC/MS-SPME)
method for hexanal. Protein extraction and solubilization procedures were optimized to
maximize the formation of hexanal-protein adducts and percent extractable protein.
Freeze—dried myoﬁbrils (MF) with added methyl linoleate (0.6 mmol/g protein) were
stored for 5 days at 50°C at a“. of 0.30 and 0.75. Hexanal was measured by GC/MS-
SPME and CI-ELISA, and malonaldehyde was measured by TBARS. Lipid oxidation
reached a maximum after 4 days of storage at both water activities. then decreased. At
low water activity, 34.7 and 39.7 ppm hexanal were detected by GC/MS-SPME and CI-
ELISA respectively, after 4 days of storage. The Cl-ELISA was well correlated with the
GC/MS-SPME (r=0.78) and TBARS (r=0.87) methods. The correlation of the hexanal-
speciﬁc CI-ELISA to both GC/MS-SPME and TBARS veriﬁed the ability of the Cl-
ELISA to be used as an index of lipid oxidation. offering the convenience for use in a kit

to be utilized within a food processing facility.

My (Days go On

I praise ’Iﬂee “cu/iii}: my days go on;
I ﬂwe 771cc w/iilc my days go on;
r17iroug/i d’ar/{and’déarti t/irougﬁfire and frost,
’Wz’t/i emptied'arms and treasure ﬁrst,

I Man/{ﬂee wﬁile my days go on

~ £[1Lza5et/i @arrett Crowning

iv

ACKNOWLEDGMENTS

I gratefully acknowledge the support, patience, kindness and wisdom extended to me by
my doctoral advisory committee; Dr. Denise Smith, Dr. James Pestka, Dr. Randy

Beaudry, and Dr. Gale Strasburg.

For their support both in the laboratory and out, I would like to thank my labmates.
Jennifer Maurer, Sarah Smith, Jin-Shan Shie, Virginia Vega-Warner and Mannee
Vittayanont. For technical support, I would like to thank Chuck Allison, Dr. Winnie

Chiang, Chris Warsow, and the CANR Statistical Consulting Centre.

For their continued support and friendship through the years, Melanie Johnson, Sarah

Warner, John Vink, Denise Bailey and Kim Green.

To all my Michigan friends who supported me throughout the eventful journey to my
Ph.D.. especially Nicole Zarb. And to Laura Clifford, without whom, classes and
comprehensives would not have been nearly as fun. To my BFF Christie Eppler, I extend
my eternal gratitude for all the advice, pep talks, and bad movies. Also, thank you to
Betty and Charlie Downs (and Bob) for giving me such a welcome home away from

home.

One person who has been an essential part of my Ph.D. experience is Alicia Orta-

Ramirez. Her support, humour and friendship have meant a great deal to me during these

past 4 years.

I would like to thank my brothers Eric and Hugh who never doubted the abilities of their
baby sister (und Nils - Danke schon). To all the strong, intelligent and kind women in my
family, especially my sister, Colleen Ross, my mom, Felicity Ross, my grandma, Vernice
Ross, and my late grandma, Margaret Neave, whose examples have motivated and

inspired me... thank you.

Heart-felt thanks to my parents, David and Felicity Ross, whose love, encouragement.

support, motivation, and passion for learning gave me the tools to pursue my Ph.D.

And ﬁnally... I would like to thank my husband Larry Goodridge for all his love and

support, interesting insights and scientiﬁc ideas and for always having conﬁdence in me.

vi

TABLE OF CONTENTS

List of Tables ................................................................................................................ xi
List of Figures ............................................................................................................. xiii
Chapter 1. Introduction ............................................................................................ 1
Chapter 2. Literature Review ................................................................................... 5
2.1 Mechanism of Lipid Oxidation .............................................................. 5
2.2 Toxicity of Lipid Oxidation Products .................................................... 9
2.3 Measurement of Lipid Oxidation ......................................................... 1 1
2.4 Hexanal ................................................................................................ 13
2.5 Volatile Analysis .................................................................................. 15
2.6 SPME ................................................................................................... 16
2.7 lmmunoassays ...................................................................................... 22
2.8 ELISAs in Use ..................................................................................... 24
Chapter 3. Solid-phase Microextraction for the Measurement of Hexanal in

3.1

3.2

3.3.

Chicken Myoﬁbrils

Abstract ................................................................................................ 31
Introduction .......................................................................................... 32
Materials and Methods ......................................................................... 35
3.3.] Preparation of Chicken Myoﬁbrils .......................................... 35
3.3.2 GC/MS-SPME Procedure ........................................................ 36

vii

3.3.3 Limit of Detection and Response Linearity of
GC/MS-SPME ......................................................................... 37

3.3.4 Hexanal Loading and Repeated Sampling of Spiked Chicken

Myoﬁbrils ................................................................................ 38

3.3.5 Hexanal Partitioning ................................................................ 40

3.3.6 Accelerated Oxidation Study ................................................... 41

3.3.6.1 Conjugated Diene Assay .............................................. 42

3.3.6.2 Thiobarbituric Acid-Reactive Substances Assay ......... 42

3.3.6.3 Accelerated Oxidation Study Data Analysis ................ 43

3.4 Results and Discussion ........................................................................ 44
3.4.1 Validation of the GC/MS-SPME Procedure ........................... 44

3.4.2 Hexanal Loading and Repeated Sampling of Spiked Chicken

Myoﬁbrils ................................................................................ 46
3.4.3 Hexanal Partitioning between Headspace and Myoﬁbrils ...... 46
3.4.4 Accelerated Oxidation Study ................................................... 52
3.4.4.1 Hexanal in Headspace ................................................ 52

3.4.4.2 Effect of Water Activity on Lipid Oxidation during

Accelerated Studies ..................................................... 54
3.4.4.3 Nine-day Accelerated Oxidation Study ....................... 55

3.4.4.4 Fifty-hour Accelerated Oxidation Study ..................... 59

3.4.5 SPME Applications in Lipid Oxidation Detection ................ 61

3.5 Conclusions .......................................................................................... 64

viii

Chapter 4.

4.1
4.2

4.3

Veriﬁcation of a Monoclonal Antibody-Based ELISA for the

Monitoring of Lipid Oxidation in Chicken

Abstract ................................................................................................ 65
Introduction .......................................................................................... 66
Materials and Methods ......................................................................... 70
4.3.1 Materials ..................................................................................... 70

4.3.2 Optimization of Extraction procedures to Solubilize

Myoﬁbrils ................................................................................. 71
4.3.2.1 Effect of KCl on CI-ELISA .................................................... 72
4.3.2.2 Use of Protease to Solubilize Myoﬁbrils ................................ 72
4.3.2.3 Trypsin Optimization and Inactivation .................................... 73
4.3.3 Growth and Maintenance of Hybridomas ................................... 74
4.3.4 Protein-HexanalConjugation for Standard Preparation .............. 75
4.3.4.1 Use of Dialysis during Protein-Hexanal Conjugation .............. 75
4.3.5 Trinitrobenzenesulfonic Acid Assay (TNBS) ............................. 76
4.3.6 CI-ELISA Optimization .............................................................. 76
4.3.6 1 Indirect Competitive ELISA ................................................... 76
4.3.6.2 Calculation of Hexanal Concentration from the C I-ELISA ..... 77
4.3.6.3 Precision of the CI-ELISA ....................................................... 79
4.3.7 Bradford Protein Assay ............................................................... 79
4.3.8 Accelerated Oxidation Study ..................................................... 79
4.3.9 Data Analysis ............................................................................. 8]

ix

4.4 Results and Discussion ........................................................................ 82

4.4.1 CI-ELISA Optimization ........................................................... 81

4.4.1.1 Limit of Detection and Reproducibility ....................... 81

4.4.1.2 Effect of Dialysis ........................................................ 82

4.4.2 Optimization of Extraction Procedures ................................... 85

4.4.2.1 Effect of Extraction Buffer ........................................ 85

4.4.2.2 Effect of Proteases ....................................................... 88

4.4.3 Optimization of Trypsin Digestion .......................................... 91

4.4.4 Accelerated Oxidation Study ................................................... 97

4.5 Conclusion ......................................................................................... 1 02
Chapter 5. Conclusions ........................................................................................ 104
Chapter 6. Future Research .................................................................................. 106
References .................................................................................................................. 108

Table

3.1.

3.2.

3.3.

4.1.

4.2.

4.3.

4.4.

LIST OF TABLES

Log peak area of hexanal alone and hexanal incubated with chicken
myoﬁbrils following the achievement of steady state equilibrium. Hexanal
standards (1.5 mL) were either incubated alone in a 15-mL vial or with
chicken myoﬁbrils (0.03 g) for 24-hours at ambient temperature and
sampled ............................................................................................................ 50

Correlation coefﬁcient matrix (r) between methods used to measure the extent
of lipid oxidation in freeze-dried chicken myoﬁbrils over 9-day

storage at 50°C. Tests used were GC/MS SPME to measure hexanal.

TBARS and CD to measure conjugated dienes ............................................... 58

Correlation coefﬁcient matrix (r) between methods used to measure the

extent of lipid oxidation in freeze-dried chicken myoﬁbrils over 50-hour
storage at 50°C. Tests used were GC/MS SPME to measure hexanal,

TBARS and CD to measure conjugated dienes ............................................... 62

Reproducibility of the CI-ELISA as described by the coefﬁcient of
variation for within-assay and between-assay reproducibility ......................... 84

The effect of papain, fungal protease and alkaline protease on the solubility

of chicken myoﬁbrils. Myoﬁbrils (0.30 g/mL) were mixed with 10 mL of
buffer and protease and incubated at 50°C for 2 hours in a shaking water

bath. Following straining through cheesecloth. % soluble protein was
determined. Determinations were made in triplicate ...................................... 89

Effect of alkaline protease and endoproteinase lys-C on the soluble

protein and hexanal detection in chicken myoﬁbrils. Myoﬁbrils (30 mg/mL)
were mixed with 10 mL of buffer and protease and incubated in a 50°C
circulating water bath for 2 hours. Following straining through cheeseclotlz.
protein concentration was determined. Myoﬁbrils were conjugated to 200
mM hexanal, Schiff bases reduced, and hexanal concentration was determined
using the CI-ELISA; the amount of spiked hexanal was compared to the
amount of recovered hexanal. Determinations were made in triplicate .......... 90

Effect of typsin and endoproteinase lys-C on the protein and hexanal
recovery in spiked chicken myoﬁbrils. Myoﬁbrils (0.6 g) or BSA

were mixed with 10 mL of buffer (pH 7.7). Treatments included the
addition 0.25% (w/w) trypsin or 0.005% (v/v) endoproteinase lys-C .

xi

4.5.

4.6

Samples were incubated at 50°C for 2 hours in a shaking water bath,

strained through cheesecloth, soluble protein determined. The conjugation
and reduction reactions were performed and hexanal concentration was
determined using the CI-ELISA. Recovery of hexanal was determined by
comparing the amount of hexanal used for spiked and the amount determined
using the CI-ELISA. Determinations were made in triplicate ........................ 92

Protein solubility as affected by trypsin concentration (0.05, 0.1 0.25 % w/w).
pH level (8.0 and 9.0) and incubation time at 50°C (2, 4, 5.5 hours).
Myoﬁbrils (0.15g/mL) were mixed with 10 mL of 0.6 M KCl

in 0.05 M phosphate buffer, trypsin, and hexanal, making a ﬁnal hexanal
concentration of 200 mM. Solutions were conjugated, reduced and

incubated at 50°C for the appropriate time. Solutions were analyzed for

% soluble protein using Bradford Protein Assay, and % modiﬁcation

using the TNBS Assay. Determinations were made in triplicate .................... 93

Correlation coefficient matrix (I) between methods used to measure the

extent of lipid oxidation in freeze-dried chicken myoﬁbrils over 5 days of
storage at 50°C. Tests used were GC/MS-SPME and CI-ELISA to measure
hexanal and TBARS ...................................................................................... 100

xii

Figure

2.1.

2.2.

2.3.

3.1.

3.3.

LIST OF FIGURES

Solid-phase microextraction - extraction desorption process. Adapted
from Supelco - The Reporter 19:1 (2001) ........................................................ l7

Diagram of indirect competitive enzyme-linked immunosorbent assay (C l-
ELISA). An antigen-coated well is prepared, followed by the addition of the
monoclonal antibody and sample. Enzyme-conjugated secondary antibody
and substrate are then added. Adapted from Kuby (1997) ............................. 23

Reaction in the formation of a Schiﬁ base between a primary amine and
hexanal: 1) unstable conjugate (Schiff Base) 2) stable conjugate ................... 29

Hexanal loading and recovery from chicken myoﬁbrils. A) Hexanal

loading: 0.03 g of myoﬁbrils incubated over 1.5 mL of hexanal standard

B) Recovery of hexanal from spiked myoﬁbrils following 24 hours of
incubation ....................................................................................................... 39

Linearity of GC/MS-SPME response to hexanal concentrations (r2=0.995)..
Standards were prepared in squalene and sampled using SPME Line
represents a ﬁtted regression line (y=963142x+168333.2). Determinations
were made in triplicate ..................................................................................... 45

Steady state equilibrium in chicken myoﬁbrils with hexanal loading.
Myoﬁbrils (0.03 g) were incubated over 1.5 mL hexanal standard (0.01 to

10 ppm hexanal) prepare in squalene. The headspace above the myoﬁbrils

in the chamber was sampled every hour up to 5 hours and following 24 hours
of incubation. Open symbols represent the ﬁrst desorption of hexanal from
spiked chicken myoﬁbrils following transfer from hexanal loading vials to 4
mL vials. Determinations were made in triplicate .......................................... 47

Effect of repeated sampling on headspace above chicken myoﬁbrils

(0.03 g) incubated over 1.5 mL of 0.01 to 10 ppm hexanal standard

prepared in squalene. Headspace was sampled 6 times over 5 hours.
Determinations were made in triplicate ........................................................... 48

Hexanal formation in chicken myoﬁbrils containing 0.6 umol methyl
linoleate/g protein as affected by storage time at 50°C at aw of 0.30 and

0.75. The headspace above the myoﬁbrils was sampled using GC/MS-SPME
for 50 hr storage period (a) and 9 day storage period (b). Error bars

represent standard deviation ............................................................................. 53

xiii

3.6.

4.1.

4.2.

4.3.

4.4.

Comparison of the headspace hexanal concentration determined by
GC/MS-SPME and TBARS concentration of chicken myoﬁbrils stored for
9-days at different water activities at 50°C. Chicken myoﬁbrils contained 0.6
umol methylinoleate/g protein. Inset: Effect of water activity and length of
storage at 50°C on the production of conjugated dienes during the same 9-day
storage period. Determinations were made in triplicate. Error bars represent
standard deviations ........................................................................................... 56

Comparison of the hexanal concentration determined by GC/MS-SPME and
TBARS concentration of chicken myoﬁbrils stored for 50-hours at different
water activities at 50°C. Chicken myoﬁbrils contained 0.6 umol methyl
linoleate/ g protein. Inset: Effect of water activity and length of storage at
50°C on the production of conjugated dienes during the same 50-hour storage
period. Determinations were made in triplicate .............................................. 60

The percent binding inhibition produced in the CI-ELISA from
concentrations of hexanal. A stock solution of 1 mg/mL CSA was prepared,
reduced, dialyzed and diluted. Standards between 0 and 690 ng hex/mL were

q

prepared. Each point represents 6 replicates (r2=0.975) .................................. 83

Effect of overnight dialysis following conjugation/reduction on the

CI-ELISA standard curve. The CSA-hex conjugate (1 mg/mL) was

prepared, reduced and dialyzed (control was not dialyzed) and standards from
3 to 69 ng hex/mL were prepared. Each point represents three replicates ...... 86

Effect of the addition of different concentrations of 0.6 M KCl in 0.05 M
potassium phosphate on the CI-ELISA standard curve. From a stock of 1
mg/mL C SA-hex, standards from 3 to 69 ng hex/mL were prepared in
different dilutions 0.6 M KCl in 0.05 M potassium phosphate. The control
standard curve was prepared in 0.05 M phosphate buffer. Determinations
were made in triplicate using the CI-ELISA .................................................... 87

Trypsin inactivation through heating and the effect on the CI-ELISA

standard curve. Solutions of BSA (0.3 mg/mL) in 10 mL of 0.6 M KC1

in 0.05 M potassium phosphate were mixed with 0.25% (w/w) trypsin,
hexanal and sodium cyanoborohydride. Final reactant concentrations were
200 mM hexanal and 80 mM sodium cyanoborohydride. The mixtures

were incubated at 50°C for 2 hours and subjected to a heat treatment.
Following cooling and centrifugation, the solutions were used to prepare
standard curves ranging from 3 to 69 ng hexanal/mL and analyzed using

the CI—ELISA. Determinations were made in triplicate ................................ .95

xiv

4.5.

4.6.

4.8.

Flow chart of the optimized CI-ELISA method ............................................... 96

Comparison of the hexanal concentration determined by GC/MS-SPME and
CI-ELISA, and TBARS concentration of chicken myoﬁbrils stored for 5 days
at 50°C. Chicken myoﬁbrils contained 0.6 umol methyl linoleate/g protein
and were stored at aw of 0.30. Determinations were made in triplicate .......... 98

Comparison of the hexanal concentration determined by GC/MS-SPME and
CI-ELISA, and TBARS concentration of chicken myoﬁbrils stored for 5 days
at 50°C. Chicken myoﬁbrils contained 0.6 umol methyl linoleate/ g protein
and were stored at aw of 0.75. Determinations were made in triplicate .......... 99

XV

CHAPTER 1

INTRODUCTION

Lipid oxidation has long been recognized as a leading cause of quality
deterioration in muscle foods. Changes in quality are manifested by adverse changes in
ﬂavor, color, texture and nutrient value. Lipid oxidation is often the decisive factor in
determining the useful processing and storage life of both fresh and frozen meat products
(Frankel, 1993). In addition, consumer acceptance of meat products depends to a major
extent on ﬂavor quality, and ﬂavors generated by lipid oxidation are generally undesirable
(Ramarathnam et al., 1991 ). All muscle foods are susceptible to lipid oxidation; however.
the muscle foods of most concern are those with high concentrations of unsaturated fatty
acids. such as chicken. As reported by the USDA (2001), chicken consumption has
surpassed pork and beef, at 82 pounds of chicken consumed per person per year. Chicken
consumption is expected to continue to increase over the next several years (National
Chicken Council, 2001).

The production of off-ﬂavor and odors were considered to be the most imp 'thant
consequences of lipid oxidation in the past; however, increased attention is now being
given to the health risks that lipid oxidation may impose. In animal studies, thermally
oxidized lipids were found to be precursors of colon cancer (Yang et al., 1998; Corpet et
al., 1990). The genotoxicity and cytotoxicity of the secondary oxidation products.

malonaldialdehyde and 4-hydroxynonenal. has been established through several studies

(Esterbauer, 1993; Bird et al., 1982). Research has also linked the formation of these
secondary products to disease states such as atherosclerosis (Kondo et al., 2001; Salomon
et al., 2000), autoimmunity (Khan et al., 2001), and cancer (Bruce et al., 1993). Due to
health concerns, measuring the extent of lipid oxidation is becoming increasingly
important due to the possible health risks associated with the consumption of lipid
oxidation products.

From the perspective of the meat industry, detection of lipid oxidation needs to be
accomplished in a very speciﬁc, sensitive and rapid manner. While organoleptic
evaluation provides the most useful consumer information, it is time consuming and often
has low reproducibility (Frankel, 1993). Chemical methods attempt to improve on these
limitations but ultimately, the results of chemical methods must support sensory
measurements. The most commonly used method for measuring lipid oxidation is the
2-thiobarbituric acid reactive substances test (TBARS), which was originally thought to
measure only malonaldehyde, a secondary product of lipid oxidation. Criticisms of
TBARS arise from its lack of sensitivity and its speciﬁcity to malonaldehyde.

The measurement of hexanal, another secondary product of lipid oxidation, has
been used as an indicator of lipid oxidation. Hexanal is the dominant oxidation product
of linoleic acid, being generated by both the 9 and 13-hydroperoxide of linoleate. Due to
its low odor threshold, hexanal is also a major contributor to the rancid odors and ﬂavors
associated with oxidized meat (Shahidi‘and Pegg, 1994). Gas chromatography was used
to establish strong correlations (0.995) between hexanal content, sensory scores and

TBARS values in a number of meat products (Spanier et al., 1992; Ahn et al., 1998). Gas

chromatography has many drawbacks, including cost, time and the need for trained
personnel. The majority of studies in the area of lipid oxidation have employed the
traditional headspace GC techniques as a standard against which other methods are
evaluated. Solid-phase microextraction (SPME) is a new adsorption technique that has
gained widespread acceptance in volatile analysis during recent years (Steffen and
Pawliszyn, 1996). SPME involves the extraction of analytes from the headspace above a
sample onto a fused silica ﬁber; the analytes are then desorbed into a GC for analysis.
While SPME combined with gas chromatography/mass spectrometry has been used for
volatile analysis in various fruits and vegetables, its application has been limited it: meat
products. In the area of lipid oxidation, SPME has recently been employed for the
detection of aldehydes in turkey (Brunton et al., 2000) and sunﬂower oil (Keszler et al.,
1998). SPME overcomes the difﬁculties associated with traditional GC techniques,
including lengthy analysis time, use of solvents, and equipment dedication.

Hexanal has been detected using both polyclonal (Smith et al., 1999) and
monoclonal antibodies (Zielinski et al., 2001) in a competitive indirect enzyme-linked
immunosorbent assay (CI-ELISA) . The advantages of a CI-ELISA over conventional
techniques have been well described in the literature as rapid, reproducible, and highly
sensitive (Samarajeewa et al., 1991). In addition, the CI-ELISA offers the convenience
for use in a self-contained kit, which can be easily used within a food processing facility
to monitor products as a routine component of a quality assurance program. This kit
would also allow analyses in different laboratories to be performed under standard

conditions in different laboratories.

Though the use of a monoclonal-based CI-ELISA for detection of lipid oxidation
seems ideal, previous studies still have not established its use in a food system. The
overall goal of this project is to verify the use of a monoclonal CI-ELISA for monitoring
lipid oxidation in chicken, through the quantiﬁcation of hexanal. To achieve this goal.
the following objectives were investigated:

1) develop and optimize a GC/MS method for hexanal detection

2 optimize extraction procedures to maximize formation of hexanal-protein

adducts and percentage extractable protein on preparation for the C I-
ELISA

3) optimize the monoclonal CI-ELISA for hexanal detection

4) conduct an accelerated oxidation study in which results from the

monoclonal CI-ELISA are correlated with results from the GC/MS and

TBARS

CHAPTER 2

LITERATURE REVIEW

2.1 Mechanism of Lipid Oxidation

The process of lipid oxidation, as outlined by Asghar et a1. (1988), proceeds via a
free radical mechanism termed autoxidation and involves initiation, propagation and
termination steps. Initiation is the abstraction of a hydrogen atom from the methylene
carbon in the side chain of an unsaturated fatty acid to form a lipid free radical.
Propagation involves the molecular rearrangement of the lipid free radical followed by a
reaction with molecular oxygen (in an aerobic system) to form the hydroperoxy radical.
The hydroperoxy radical can remove hydrogens from other fatty acid side chains, thereby
propagating the reaction. Termination occurs when two radicals react to form
nonpropagating (non-reactive) intermediates. The major primary products of this
reaction. hydroperoxides. are relatively unstable and decompose into a wide range of
carbonyl compounds, hydrocarbons, ketones and other materials that contribute to the
ﬂavor deterioration of the food.

The decomposition of lipid hydroperoxides involves further reactions and the
production of low molecular weight compounds such as alkanes, alkenes, aldehydes,
ketones, alcohols, esters and acids. Homolysis of lipid hydroperoxides to hydroxy and
alkoxy radicals is followed by cleavage of the fatty acid chain adjacent to the alkoxy

radicals. Cleavage, or scission, occurs via one of two routes. In the ﬁrst scission route,

an unsaturated aldehyde and an alkyl radical are formed, which can react with a hydroxyl
radical to form an alcohol. In the second scission route, a saturated aldehyde and vinyl
radical are formed, which in turn can react with a hydroxyl radical to form an aldehyde.
Alternative pathways of hydroperoxide decomposition are further oxidation, yielding
epoxides, cyclic and bicyclic peroxides, or condensation into dimers and polymers.
which may oxidize and decompose into volatile breakdown products (Ladikos and
Lougovois, 1990). Aldehydes are the most important breakdown products as they
possess low threshold values and are the major contributors to the development of rancid
off-ﬂavors and odors (Ladikos and Lougovois, 1990).

The interactions of hydroperoxides and their secondary products with proteins or
amino acids have a considerable impact on ﬂavor stability during processing, cooking and
storage. The form of these interactions is determined by the point at which the proteins
enter the lipid oxidation reaction chain, when either the radicals or secondary prod .cts
predominate (Ladikos and Lougovois, 1990). Radical induced cross-linking of proteins
may be responsible for most nutritional losses, including lysine (up to 71%) and histidine
(up to 57%) in a whey protein-methyl linoleate system (Nielsen et al., 1985). At high
temperature (160°C), hydroperoxides begin to decompose spontaneously and the radical
concentration can become relatively high, creating a greater likelihood of radical-radical
interactions. In addition, color changes in beef during lipid oxidation are due to the loss
of oxygen from oxymyoglobin (bright-red), together with an electron transfer. to produce
metmyoglobin (brown).

Lipid oxidation in foods may be initiated by a number of mechanisms. however,

the majority of biological lipid peroxidation reactions involve transition metal ions
present within the system (Kanner, 1994). This can occur through the homolysis of weak
bonds, such as the CO bond in a peroxide (Kaur and Perkins, 1991). The classic
example is the reduction of hydrogen peroxide by iron (11), known as the Fenton reaction.
The resulting hydroxyl free radical is extremely reactive and is capable of removing a
hydrogen atom from an unsaturated lipid (Kubow, 1993). Following initiation, metals
may also play a role in the propagation of lipid oxidation by catalyzing the breakdown of
lipid hydroperoxides. The source of these metal catalysts in food is controversial;
however, heme and various iron proteins including hemoglobin and myoglobin, both in
muscle tissues, have been suggested as sources of iron (Gutterridge and Halliwell, 1990).
Ferric heme pigments have been implicated as the major prooxidants in tissue lipid
oxidation (Love and Pearson, 1971). In heated pork, heme proteins were reported to have
a greater prooxidant effect than inorganic iron (Monahan et al., 1993). All muscle foods
contain iron and are therefore susceptible to lipid oxidation.

Another important internal factor inﬂuencing lipid oxidation is the unsaturated
fatty acid composition of the food (Shahidi and Pegg, 1994). In general, the reaction rate
of autoxidation increases as the number of double bonds increases. The oxidation of
oleic, linoleic, linolenic and arachidonic acid, the main unsaturated fatty acids in animal
lipids, gives rise to a number of different hydroperoxides, which in turn leads to the
formation of a large number of volatile compounds (Ladikos and Lougovois, 1990).
Oleic acid oxidation generates octanal, nonanal and decanal, while the oxidation of

linolenic and arachidonic acids generate propanal and hexanal, respectively. Linoleic

acid oxidation produces two hydroperoxides, which both undergo scission to produce
hexanal (Ladikos and Lougovois, 1990).

In addition to internal factors, a number of external factors. including storage
temperature, time, light, oxygen availability and water activity inﬂuence the rate of lipid
oxidation. The term “warmed-over ﬂavor” (WOF) was ﬁrst introduced by Tims and
Watts (195 8) to describe the rapid onset of rancidity in meat products during refrigerated
storage and can develop within 48 hours post-processing. While this term was originally
used to describe the development of rancid off-odors and ﬂavors in cooked, refrigerated
meat, research has indicated that WOF also occurs in raw meat (Jakobsen and Bertelsen.
2000; Gray et al., 1996). Recently, the term WOF has been questioned on the basis that it
does not adequately describe the ﬂavor changes in cooked meat during storage (Gray et
al., 1996). The term, meat ﬂavor deterioration, has been developed to better describe the
chemical reactions that contribute to the overall increase in oxidative off-ﬂavors wth a
simultaneous deterioration of the desirable meat ﬂavor (Spanier et al., 1992).

Heat affects lipid oxidation through the decrease of activation energy required for
oxidation; the fraction of total collisions that can lead to the reaction increases rapidly
with increasing temperature (Gillespie et al., 1989). High temperatures also accelerate
the break down of hydroperoxides to free radicals, which propagate lipid oxidation and
the development of off-ﬂavors (Kanner, 1994). Increased light and oxygen availability
(J akobsen and Bertelsen, 2000) have also been identiﬁed as factors that increase the rate
of lipid oxidation.

In model lipid systems, research has shown that oxidation rate depends strongly

on water activity (Karel, 1980). At very low water activity. oxidation proceeds very
rapidly. Increasing the water activity to aw of 0.30 retards lipid oxidation, as this water
activity is near the monolayer moisture content. The monolayer represents a theoretical
moisture content where the protein binds one molecule of water per polar group on the
surface of the protein (Kinsella and Fox, 1986). More recently, the monolayer has been
deﬁned as the ﬁrst water molecules which interact with functional groups of solutes
(Chinachoti, 2000). The protective effect of water is believed to occur by reducing the
catalytic activity of metal catalysts, by quenching free radicals, by promoting non-
enzymatic browning and/or by impeding the access of oxygen to the food. At high er
water activities, the rate of oxidation increases again as a result of increased mobilization
of the catalysts present (Nawar, 1985). These factors, both internal and external.
inﬂuence the rate of lipid oxidation and by extension, the production of primary and
secondary lipid oxidation products.
2.2 Toxicity of Lipid Oxidation Products

The production of off-ﬂavors and odors were considered to be the most important
consequence of lipid oxidation in the past; however, increased attention is now being
given to the health risks that lipid oxidation may impose. Therefore, considerable interest
exists in the role played by lipid oxidation in human disease and toxicology. While there
has been no direct evidence implicating meat lipid hydroperoxides as a possible health
hazard, there are numerous studies that report the toxic effects of oxidized fats in animals.
Two studies (Yang et al., 1998; Corpet et al., 1990) reported thermally oxidized lipids to

be precursors of colon cancer. In both studies, the aberrant crypt foci (ACF) assay was

used as a short-term test of colon carcinogens based on early putative steps in colon
carcinogenesis. In the Corpet et al. (1990) study, rats were fed oxidized beef tallow (20%
of the diet) for 100 days while in the Yang et al. (1998) study. rats were fed oxidized beef
tallow for 28 or 100 days. Results from both studies indicated that oxidized beef tallow
produced more (p<0.05) ACF than did the control diet. F urthermore. Fouad et al. (1995)
conducted a histological and pathological study that demonstrated the effect of thermally
oxidized lipids on the development of cirrhosis of the liver in rats fed oxidized butter oil
for 100 days at 1.05 mL per day.

Several studies have been conducted examining the toxicity of aldehydes,
secondary products of lipid oxidation. Bird et al. (1982) examined the chronic toxicity of
malonaldehyde in female mice. Malonaldehyde, administered in the drinking water for
one year at levels of 0. 1 , l and 10 mg/g body weights, resulted in a dose-dependent
(p<0.05) increase of neoplasms and liver lesions. A study conducted by Yao (1979,
reported that 24-hour exposure of mammalian cells to 20 mm malonaldehyde produced
detectable cytotoxicity. Esterbauer et al. (1993) also reported toxicity of prepared
4-hydroxynonenal in primary cultures of rat hepatocytes at a concentration of 100 mm
following 48-hour incubation. Micronuclei frequency, chromosomal aberrations and
sister-chromatid exchanges were also observed following incubation.

Hydroperoxides readily decompose, leading to the production of secondary
compounds. Due to health concerns, measuring the extent of lipid oxidation in foods is
becoming increasingly important due to the possible health risks associated with the

consumption of these secondary lipid oxidation products.

10

2.3 Measurement of Lipid Oxidation

Lipid oxidation is a complex process and a number of methods exist by which
oxidative changes can be measured. Sensory methods provide the most useful
information related to consumer acceptance of the food; however, these methods are
time-consuming and display poor reproducibility (Frankel, 1991). Chemical methods
include the measurement of hydroperoxides, the colorless, tasteless and odorless primary
products of lipid oxidation. The breakdown of these products yields aldehydes, ketones
and esters, and other low molecular weight compounds, which impart rancid and other
off-ﬂavors to the meat and the toxic effects mentioned above.

Lipid oxidation methods have been divided into those that measure primary
change and those that measure secondary change (Gray and Monahan, 1992). Methods
that measure primary change quantify the loss of reactants (unsaturated fatty acids or
oxygen) or the formation of primary lipid oxidation products (hydroperoxides, conjugated
dienes). A new method of hydroperoxide determination in dark chicken meat was
proposed by Grau et al. (2000). This method consisted of the peroxide-mediated
oxidation of ferrous ions in an acidic medium containing the dye xylenol orange; the dye
bound to the ferric ions in the chicken and produced a blue-purple complex that was
measured spectrophotometrically. Results indicated that this method was highly speciﬁc
for hydroperoxides and showed a linear relationship between volume of meat extract and
absorbance. Peroxide is a common measurement of lipid oxidation; however. because
peroxides decompose to secondary products relatively quickly, this value can result in an

underestimation of the degree of oxidation. These methods are thought to be more suited

11

to measure low levels of oxidation in uncooked products stored at low temperature
(Frankel, 1991).

In situations where oxidation occurs at an accelerated rate, as in cooked meats. an
accumulation of secondary products quickly occurs. Lipid oxidation is best assessed
through the measurement of these secondary products (Frankel, 1991). Secondary
changes include the formation of carbonyls and hydrocarbons, and the formation of
aldehydes such as malonaldehyde, measured by the 2-thiobarbituric acid reactive
substances assay (TBARS) or the increasingly popular HPLC, and hexanal, usually
measured by gas chromatography.

TBARS is one of the oldest and most commonly used methods for assessing lipid
oxidation in foods (Ladikos and Lougovois, 1990). This method is based on the
spectrophotometric determination of extracted malonaldehyde, a minor product of
oxidation, and can be performed either on the food product or on a steam distillate of the
food. In this assay, one mole of malonaldehyde reacts with 2 moles of thiobarbituric acid
to form a pink complex, which can then be measured spectrophotometrically at 532 nm.
The extent of lipid oxidation is reported as a TBARS value, or milligrams of
malonaldehyde equivalents per kilogram of sample. TBARS measurements have been
reported to correlate well with sensory scores of oxidized ﬂavors in muscle foods and
WOF development (St. Angelo et al., 1987; Igene et al., 1985).

Due to the simple procedure and high correlation with sensory scores (Igene et al..
1979), many studies have relied on TBARS for lipid oxidation detection. In several

recent studies in chicken (Tang et al., 2001a; Tang et al.. 2001 b; Alasnier et al.. 2000:

12

Sista et al., 2000; Li and King, 1996), freeze-dried beef (Wilkinson et al., 2001), and beef
(Jakobsen and Bertelsen, 2000), TBARS was the principal method used for lipid
oxidation assessment. Criticisms of the TBARS method include wide variations in
procedures between laboratories, leading to large differences in results. Several methods.
such as direct heating of the sample with the TBA solution (Sinnhuber and Yu, 195 8) and
distillate reaction with the TBA solution (Tarladgis et al., 1960) require a heating step.
which may cause the formation of additional lipid oxidation products. The major
disadvantage of the TBARS reaction is that it is not speciﬁc for malonaldehyde (Gray and
Monahan, 1992). Other lipid oxidation compounds, or compounds not related to livid
oxidation, can react with thiobarbituric acid to form colored complexes that also absorb at
532 nm, thereby leading to an overestimation of the extent of lipid oxidation (Gray and
Monahan, 1992). Sun et al. (2001) reported that in freeze-dried beef stored at 49°C,
propional, butanal and S-hydroxymethyl-Z-ﬁrrﬁrral all produced strong TBARS. For
these reasons, the TBARS procedure may be used to assess the extent of lipid oxidation
in general. rather than to quantify malonaldehyde.
2.4 Hexanal

Due to the limitations of TBARS, the measurement of hexanal, a six-carbon
aldehyde, has become a popular indicator of lipid oxidation. Hexanal, one of the
secondary products formed during the oxidation of linoleic acid, dominates among the
volatile aldehydes. Hexanal is the only aldehyde that arises from both the 9 and 13
hydroperoxides of linoleate, and from other unsaturated aldehydes formed during the

oxidation of linoleate (Shahidi and Pegg, 1994). Dupuy et al. (1987) reported that in

13

cooked beef, hexanal, pentanal, nonanal and total volatiles increased following ﬁve days
of storage at 4°C. Speciﬁcally, hexanal increased from 0.1 to 15 ppm in white meat and
0.9 to 11 ppm in dark meat during storage. In comparison, in white meat, pentanal‘
increased from 0.07 to 0.8 ppm and 0.03 to 0.9 ppm in dark meat. Another study found
hexanal to increase in cooked white chicken meat from 0.064 ppm at Day 0 to 0.314 ppm
following 15 days of storage at 4°C (Ajuyah et al., 1993).

In order for any chemical method to be adopted as a valid measure of lipid
oxidation, the method must support sensory measurements (Gray, 1978), as odor and
ﬂavor evaluations relate directly to consumer acceptance (Frankel, 1991). A strong
correlation of 0.995 between hexanal content and TBARS values was reported in ground
pork (Shahidi et al., 1987). Flavor acceptability decreased as hexanal and TBARS?
number increased. Correlations between hexanal, TBARS and sensory scores were also
reported in cooked roast beef (Dupuy et al., 1987), ground beef (St. Angelo et al., 1987).
cooked pork patties (Ahn et al., 1998), cooked beef patties (Spanier et al., 1992). and
chicken nuggets stored at -20°C for 6 months (Lai et al., 1995).

In sensory tests, hexanal is often described as “painty”, “grassy” and “ﬁshy”
(Brewer and Vega, 1995). Sensory tests using descriptive analyses and category scales
identiﬁed the detectable odor threshold for hexanal to range between 2.8 and 36.8 ppm in
lean beef (Brewer and Vega, 1995) and between 5 and 10 ppm in cereal (Fritsch and
Gale, 1977). Hexanal was also reported to correlate well to the development of WOF and
sensory scores in refrigerated cooked beef (Kerler and Grosch, 1996; St. Angelo et al.,

1987; Dupuy et al., 1987). Du et al. (2001) employed a dynamic purge-and-trap GC

14

technique for hexanal detection in irradiated cooked chicken meat, in addition to using
TBARS and a l6-member trained sensory panel. Although correlations were not
published, similar trends were observed between the hexanal, TBARS and sensory
methods. Based on these studies, hexanal determination has shown to be a robust and
accurate means of assessing lipid oxidation. Several researchers have suggested that
hexanal represents the lipid oxidation of meat better than any other volatile component
(Brunton et al., 2000; Ahn et al., 1998).
2.5 Volatile Analysis

Gas chromatography has been used extensively for the separation and
identiﬁcation of volatile lipid oxidation products (Gray, 1978). Traditionally, the two
major gas chromatography methods used for volatile analysis in food are static headspace
and dynamic purge-and-trap techniques (Frankel, 1991). With the static headspace
method, the sample is closed in a vessel and equilibrated; the headspace is then sampled
and injected directly into the GC. Using static headspace, a correlation of 0.856 was
reported between TBARS and hexanal in broiler breast patties (Su et al., 1991). Static
headspace GC was also employed for the measurement of volatile compounds in cooked
chicken (Ang and Young, 1989) and beef (Wu et al., 1998). While static headspace
techniques are simple and rapid, insufﬁcient quantities of hexanal are often obtained,
leading to a decrease in sensitivity (Miller and Stuart, 1999; Reineccius, 1996).

To overcome the low sensitivity of the static headspace technique. dynamic
headspace or purge-and-trap techniques utilize an adsorbent material, such as Tenax. to

collect volatiles over a period of time. The volatiles are then applied directly to the

15

column. The use of dynamic headspace GC for the analysis of lipid oxidation volatiles
has been widespread, including cooked chicken (Ang et al., 1994; Ajuyah et al., 1993).
beef (Spanier et al., 1992) milk products (Park and Goins, 1992) and irradiated pork (Ahn
et al., 2001; Ahn et al., 2000). More recently, a correlation of 0.98 was reported between
TBARS and dynamic headspace GC in the detection of lipid oxidation in cooked chicken
thigh (Zielinski et al., 2001). While these dynamic headspace techniques are extremely
sensitive, disadvantages include excessive preparation time. Some methods require as
long as 10 hours to trap volatiles (Ahn et al., 2001). Additional disadvantages include
difﬁculty of use, equipment dedication and inconvenience for analyzing large number of
samples quickly (Steffen and Pawliszyn, 1996).

2.6 Solid-Phase Microextraction (SPME)

Solid-phase microextraction (SPME) is a new adsorption analytical technique that
has gained widespread acceptance in many areas during recent years. SPME overcc mes
the difﬁculties experienced with headspace methods, including high cost, excessive
preparation time and difﬁculty of use (Steffen and Pawliszyn, 1996). SPME was also
reported to be more sensitive to ﬂavor volatiles and allow for higher recovery than
traditional static headspace techniques (Miller and Stuart, 1999; Garcia et al., 1998).

In SPME, analytes are ﬁrst adsorbed from a sample (by headspace extraction)
onto the stationary phase, a fused silica -poly(dimethylsiloxane)-coated ﬁbre (Figure 2.1).
The analytes are then thermally desorbed into the injection port of a GC for analysis
(Arthur and Pawliszyn, 1990). Whether the analytes migrate into the coating of the ﬁber

or remain at its surface depends on the diffusion coefﬁcient of the analytes (Gorecki et al.

16

 

 

Extraction 2. Exgqigciaber/
Procedure 3. Retroct
F'b z’
1. Pierce Sample A i Rémeorpe
Septum 1’!
l ' ——-> —>
2'41) ‘
D o pti 2. Expose Fiber/ I
es r on
Procedure Desorb 3 Retroct
l Fiberx’
1. Pierce CC ; Remove
Inlet Septum i; i
———> ——> i; -+ .
j,

 

 

 

 

Figure 2.1. Solid-phase microextraction-extraction process
Adapted from Supelco - The Reporter 19:1 (2001)

17

1999). During analysis of headspace by GC-SPME, two equilibria are reached by the
analytes: between the matrix and the headspace, and between the headspace and the
coating of the ﬁber (Zhang and Pawliszyn, 1993). Using SPME. l7 analytes commonly
found in orange juice ﬂavor volatiles, were calculated to be within the same range as
those reported using headspace techniques, with high reproducibility (below 10% relative
standard deviation) (Steffen and Pawliszyn, 1996). In addition to being used for volatile
detection in food, SPME has also been employed for other uses, including the analysis of
volatiles in wire coating varnishes (Hinz et al., 1999), tobacco (Clark and Bunch. 1997)
and pesticides (Betran et al., 2000; Chen etal., 1998).

SPME has successfully been applied to the analysis of a number of products. The
use of SPME for wine characterization based on aroma proﬁles has been demonstrated by
many researchers. While the use of SPME and GC provides identiﬁcation of known
compounds, the use of GC/MS can provide tentative identiﬁcation of unknown
compounds based on a spectral library of identiﬁed compounds. Garcia et al. (1998) used
GC/MS-SPME for the varietal characterization of 90 German wines using aroma proﬁles
or aromagrams. Vas et al. (1998) reported similar ﬁndings in the screening of wine.
ﬁnding that SPME was capable of distinguishing between different wine types and
wines produced in different regions. In the comparison of SPME with solvent extraction.
the researchers also found that GC-SPME was more sensitive. In the detection of
pesticides in wine, the GC/MS-SPME was reported to compare favorably to an ELISA for
the pesticide procymidone. A collaborative study coordinated by The American

Chemical Society found a correlation of 0.995 between the GC/MS-SPME detection of

18

procymidone in wine and their ELISA results. Other researchers have used GC-SPME
for the detection of pesticides in wine including Corrcia et al.(2001) and Vitali et al.
(1998). Researchers have also used GC-SPME for the veriﬁcation of the decaffeination
process in a variety of products (Yang et al., 1997) and for the analysis of ﬂavor volatiles
present in vanilla (Sostaric et al., 2000).

The use of GC/MS-SPME in cereals and grains have been limited in the past but
is becoming a more popular technique. Grimm et al. (2001) used GC-SPME for the
detection of the volatile compound 2-acetyl-l-pyrroline in rice and found that GC-SPME
compared favorably to the traditional solvent extraction technique. In oats, Sides et al.
(2001) used GC-SPME to survey changes in the volatile proﬁle induced by processing.
Both the trained taste panel and SPME reported similar ﬁndings and were able to
differentiate between oats processed under various heating conditions. Biswas and Staff
(2001) used GC/MS-SPME with thermogravimetric analysis in distiller grains to follow
volatile changes during processing.

Detection of off-ﬂavors and product characterization have been the predominate
uses of SPME in the dairy industry. In milk, SPME was employed for shelf-life
prediction of processed milk through the quantiﬁcation of volatile bacterial metabc lites
(Marsili. 2000; Marsili, 1999). A number of researchers have used SPME for the
characterization of cheese based on volatile proﬁles, ﬁnding the GC-SPME patterns
distinctly different among cheese varieties (Peres et al., 2001; Chin et al., 1996). SPME
has also been employed for the detection of free fatty acids and lactones in cheese

(Wijesundera et al., 1998) and acetaldehyde in milk (van Aardt et al., 2001). In both

19

studies, the limit of detection of GC-SPME was greater than using solvent extraction and
similar to sensory scores.

SPME and GC/MS have been used for volatile ﬂavor compound analyses in many
fruits and vegetables. In tomato and strawberry fruit (Song et al., 1998), the ﬂavor
compounds detected and their relative abundance determined using GC/MS-SPME were
reported to compare favorably to those values published using traditional methods.
Steffen and Pawliszyn (1996) reported similar ﬁndings in orange juice. GC-SPME has
been used to follow growth stage and stage of ripeness in sweet oranges (Alonzo et al..
2001) and cantaloupe (Beaulieu and Grim, 2001). The volatiles in strawberry drink
(Siegmund et al., 2001) and apple fruit (Song et al., 1997) have also been studied using
GC-SPME. In vegetables, GC-SPME was employed to detect pesticide residues and
compared with the traditional solvent extraction method, it showed a strong agreement
(Volante et al., 2000).

In the meat industry, the use of SPME for many applications has been growing.
Dry-cured hams have been analyzed for volatiles using SPME-GC/MS, while smoked
hams have been analyzed for N-nitrosodibutylamine and N-nitrosobenzylamine using
SPME and GC-thermal energy analysis (Sen etal., 1997). For both compounds, high
recoveries were achieved (86% and 70%, respectively), and the GC~SPME method
showed a strong correlation with the conventional solvent method (r2==0.97). In dry-cured
Iberian ham, Ruiz et al. (1998) used GC-SPME to assess the inﬂuence of extraction time
and temperature on the volatile proﬁle. Results indicated a time and temperature effect

on the chromatographic area of extracted compounds at 40 and 60°C at 20, 40 and 60

20

min. In most of the compounds studied, the largest peak areas were found in samples
extracted at 60°C for 60 min. Heating can release bound hexanal, improving hexanal
recovery. However, some volatile compounds, such as hexanal, are thermally generated
and heating a sample produces additional hexanal. In a variety of meat products. smoked
chicken, ham and corned beef, GC-SPME analyses for lipid oxidation products were
compared to the analyses of supercritical ﬂuid extraction and traditional purge and trap
GC (Snyder et al., 1998). Results from each method showed the pattern of oxidation
products formed to be similar and consistent, but GC-SPME was able to measure
compounds at lower concentrations than supercritical ﬂuid extraction. GC-SPME has
also been used for volatile analysis in pork (Elmore et al., 2000) and as an indicator of
food safety, to measure volatile compounds produced by bacteria in processed poultry
(Arnold and Senter, 1998).

Recently, SPME ﬁbers have been used for the quantitative measurement of
hexanal in cooked turkey (Brunton et al., 2000). A strong correlation between SPME/GC
and TBARS (r2 of 0.994) was reported when measuring lipid oxidation in cooked turkey
during storage. Currently, this is the highest reported correlation between hexanal and
TBARS. In previous studies in muscle foods, lower correlations were found betwc en
hexanal and TBARS when using different methods. Using a headspace method, a
correlation of 0.856 was reported in broiler breast patties (Su et al., 1991). When a steam
distillation method was used for hexanal quantiﬁcation, a correlation of 0.93 was reported
in pork patties (Ahn et al., 1998). Due to this high correlation, SPME GC/MS is a

promising method for quantiﬁcation of lipid oxidation through measurement of hexanal.

21

2.7 Immunoassays

The increased regulation of foodstuffs requires analytical methods which are easy
to perform, sensitive, and speciﬁc (Fukal, 1991). Immunoassays, including radio
immunoassays (RIA), and enzyme-linked immunosorbent assays (ELISA), employ the
speciﬁcity of an antibody and corresponding antigen to permit rapid detection or
measurement of the antigen or hapten. Because the ELISA protocol can be carried out by
unskilled analysts using laboratory-prepared or commercially available reagents or
standardized kits, the highly sensitive and speciﬁc ELISA continues to gain popularity
(Fukal, 1991).

Currently the most widely used immunoassay format is the ELISA (Swaminathan
and F eng, 1994). Three common types of ELISAs are the competitive direct, sandwich
and indirect. In the competitive direct ELISA, the sample and an enzyme-linked
competitor are added to an antibody-coated solid support. A substrate is added whi ‘h will
bind to the enzyme-linked competitor and the reaction is stopped. As this assay is based
on competition for antibody binding sites, free antigen concentration is inversely related
to antibody-bound enzyme conjugate (Pestka, 1988). The sandwich ELISA begins with
an antibody-coated solid support followed by antigen addition. The antigen and antibody
are allowed to react and after washing, the enzyme-conjugated secondary antibody is
added. The ﬁnal step is substrate addition and the measurement of the colored reaction
product. In the indirect ELISA, a sample containing primary antibody is allowed to bind
to the antigen-coated solid support (Figure 2.2). After any free antibody is washed away.

the enzyme-conjugated secondary antibody is added to bind to the primary antibody. The

22

 

Q

E

E
2
a

ﬁx

/‘\

ll
Add speciﬁc
antibody

A
vw

Antigen-
coated well

 

Figure 2.2. Diagram of indirect competitive ELISA. An antigen-coated
well is prepared, followed by the addition of the monoclonal antibody and

sample. Enzyme-conjugated secondary antibody and substrate are then
added. Adapted from Kuby (1997)

23

substrate for the enzyme is added and the colored reaction product is measured.

Production of both polyclonal and monoclonal antibodies begins with the
introduction of the antigen into the host. In order to elicit an immune response. the
compound must be immunogenic. Small molecules alone, generally less than 5000
daltons, are non-immunogenic and must be conjugated to a larger protein; this conjugate
is then large enough to elicit a response (Fukal, 1991). Following injection of the
antigen, the host produces a number of polyclonal antibodies, each speciﬁc to the
different epitopes on the antigen. For most research and therapeutic purposes. the
polyclonal antibodies produced are not speciﬁc enough and a monoclonal antibody
speciﬁc for a single epitope on the antigen is preferred (Kuby, 1997).

Direct puriﬁcation of a monoclonal antibody from a polyclonal antibody is not
feasible. In 1975, Koehler and Milstein devised a method for preparing monoclonal
antibodies. In this method. antibody-producing spleen cells are isolated from the host and
fused with a myeloma cell in order to produce a hybrid cell. This cell, called a
hybridoma, possesses immortal growth properties and secretes the antibody. The
resulting clones of the hybridomas can be cultured indeﬁnitely. with regular harvesting
and subsequent puriﬁcation of the antibody (Kuby. 1997).

2.8 ELISAs in use

ELISAs have found speciﬁc applications in the meat industry, including
adulteration detection, thermal processing of food proteins, and food safety. To detect
species adulteration in ground meat products, a monoclonal antibody was produced

speciﬁc to a heat stable protein in cooked mammalian meats (Hsieh et al.. 1998).

24

Adulteration of the chicken with as little as 0.5% pork, beef, horse and lamb could be
detected using the monoclonal antibody in the ELISA. In order to detect adulteration of
sausage with soy protein, Medina (1988) employed a polyclonal antibody to soy protein.
Using this antibody in an ELISA, soy concentrations of less than 2% were detected in
sausage. A sandwich ELISA using polyclonal antibodies for the indicator protein lactate
dehydrogenase was developed to verify endpoint processing temperatures in turkey roll
(Wang et al., 1993). Lactate dehydrogenase concentration, determined by the ELISA.
differed in turkey rolls processed to 70°C and 71.1°C, while lactate dehydrogenase
activity and extractable protein results showed no difference between those temperatures.
The use of LDH as an endpoint for heating was also studied in ground beef with similar
results (Wang et al., 1995). In the area of meat food safety, the ELISA format has been
adopted for the detection of bacterial toxins, including Salmonella and Staphylococcal
enterotoxins (Morissette et al., 1990) and other contaminants of meat. such as drugs.
pesticides and hormones (Fukal, 1991).

In the medical ﬁeld, antibodies have also been produced against many
compounds, including several aldehydes. Israel et al. (1986) produced antibodies to
acetaldehyde, a biochemical marker of alcohol consumption that has been shown to form
adducts with a number of proteins (Donahue et al., 1983). The antibodies reacted
speciﬁcally with erythrocyte protein-acetaldehyde but not with the control erythrocyte
proteins. Speciﬁcally, the adduct formed between the lysine group on the protein and the
acetaldehyde was recognized by the antibody. No recognition was observed when

adducts were prepared with the amino acids valine and tyrosine. Both stable (non-

25

reduced) and unstable adducts (Schiff bases) were formed when acetaldehyde bound to
protein. Using sodium cyanoborohydride as the reducing agent, Klassen et al. (1994) and
Thiele et al. (1994) raised monoclonal antibodies capable of detecting reduced
acetaldehyde-protein adducts formed with the N-ethyl lysine group on the protein.

In medicine, the cytotoxicity of several products of lipid peroxidation has
prompted researchers to develop antibodies for their rapid detection, replacing such
traditional methods as TBARS, spectrophotometry and high performance liquid
chromatography (Raharjo and Sofos, 1993). Antibodies have been raised to 4-hydroxy-2-
nonenal protein adducts (Uchida et al., 1995). This aldehyde is believed to be largely
responsible for cytopathological effects observed during oxidative stress and has been
implicated in Parkinson's disease and cardiovascular disease. In a study of people
afﬂicted with Parkinson’s Disease and using appropriate controls, researchers
successfully used the 4-hydroxy-2-nonenal antibody to detect the presence of oxidative
stress within the nigral neurons in the brain in those affected with the disease (Yoritaka et
al.. 1996). Another research group produced antibodies to a 2-oxo-nonanal protein
adduct (Salomon et al., 2000). Using an ELISA, the concentration of 2-oxo-nonan-il were
elevated in renal failure and atherosclerosis patients when compared to those in healthy
volunteers.

Antibodies have also been produced against malonaldehyde, an aldehyde thought
to play a role in the development of atherosclerosis (Swets et al., 2001; Kondo, 2001).
Results from a study in mice by Khan et al. (2001) indicated that an early appearance of

anti-malonaldehyde antibodies in the blood showed increased oxidative stress in the

26

animal. In another study, malonaldehyde antibodies were used to determine lipid
peroxidation damage following the administration of the industrial chemical aniline
(Khan et al., 2000). Rats were gavaged with different doses of aniline and a dose-
response relationship was evident, with increased concentrations of malonaldehyde in the
blood observed as the dose increased.

A food allergy study also employed 4- hydroxynonenal and malonaldehyde
antibodies in the examination of the increased allergenicity of roasted peanuts over raw
peanuts (Chung and Champagne, 2001). The authors postulated that an increase in I gE
binding of peanut allergy sufferers was attributed to increased levels of protein-bound
adducts in the blood, such as malonaldehyde and 4-hydroxynonenal. For this stud; ,
antibodies to adducts of 4- hydroxynonenal and malonaldehyde were produced, and the
ability of these antibodies to bind to IgE from serum samples of patients with peanut
allergies was studied. No binding between IgE and either 4- hydroxynonenal or
malonaldehyde was observed, indicating the limited involvement of 4- hydroxynonenal
and malonaldehyde in the allergenicity of peanuts.

Recently, a CI-ELISA using polyclonal (Smith et al., 1999) and monoclonal
antibodies (Zielinski et al., 2001) was developed for the detection of hexanal. While
some compounds are antigenic and can elicit an antibody response alone, smaller
molecules such as hexanal rarely act as antigens. These molecules are referred to as
haptens and require conjugation with a large protein molecule, such as chicken serum
albumin, to become antigenic and capable of inducing an antibody-antigen response. For

hexanal, conjugate production involves the reaction of the carbonyl group in hexanal with

27

a free amino group to form an unstable Schiff base (Figure 2.3). These bases readily
dissociate when exposed to dialysis, gel ﬁltration or treatment with weak acids or bases.
The Schiff bases can be stabilized however, by reduction to secondary amines using
sodium cyanoborohydride, a strong reducing agent, resulting in a stable conjugate (Jentoft
and Dearbom, 1979). Reductive alkylation, through the addition of sodium
cyanoborohydride, appears highly speciﬁc for amino groups of the protein, primarily
lysine (Means and Feeney, 1968).

Both the polyclonal and monoclonal CI-ELISAs for hexanal were sensitive and
showed potential for use in a meat system to follow lipid oxidation. The limit of
detection of the polyclonal antibody was 7.4 ng hexanal/mL, with coefficients of variation
ranging from 4 to 13.9 % (Smith et al., 1999). The antibody was able to detect hexanal
conjugated to several proteins, including bovine and chicken serum albumin, but did not
recognize free hexanal. In an accelerated oxidation study in a chicken thigh muscle
homogenate system, the polyclonal CI-ELISA for hexanal was found to correlate v ith the
dynamic headspace GC (r2 = 0.79) method and with lipid oxidation determined by
TBARS (r2= 0.72). While the polyclonal antibody was hexanal sensitive, its also
displayed cross reactivity to pentanal (11.8%) and heptanal (86.3%) when conjugated to
chicken serum albumin.

In order improve the speciﬁcity of the ELISA, monoclonal antibodies to hexanal
were developed. The monoclonal antibody did not recognize free hexanal but did
recognize hexanal when conjugated to a protein. The sensitivity of the monoclonal

antibody was determined to be 1 ng hexanal/mL and had coefﬁcients of variation ranging

28

gaming 633m 3 Aommm tEomV Suwanee 6382: 2
“3:88; 28 65:8 basin a 56389 6me tEom a .3 228::8 2: E cozomom .m.~ 95%..—

026383330
sewage 039m paeom cam team
AHV N m -
.mus:u-.:u-.:u-.:o-:u-zz-m 4529.2 + .5. :o- mutatzuiuuzd
oEE<
68m caEOm 633%.: E55: basin
AHV

om:+.mu-£u-.:o-.:9.5-5122 .mo-.:u-.:u-.:u-£u-muuo + £22

29

from 4.15 to 7.71 % (Zielinski et al., 2001). The monoclonal antibody displayed
moderate cross-reactivity with other aldehydes, reacting at 76.6%, 45.0% and 37.9% to
modiﬁed heptanal, 2-t-hexenal and pentanal, respectively. In the detection of hexanal in
an oxidized chicken thigh homogenate accelerated oxidation study. hexanal contents
determined using the CI-ELISA were not different from values determined using dynamic
headspace GC except at one time point (r2 = 0.66). These initial studies indicated that the
use of hexanal speciﬁc monoclonal antibody was feasible for accurate assessment of lipid

oxidation.

30

CHAPTER 3

SOLID-PHASE MICROEXTRACTION FOR THE MEASUREMENT OF

HEXANAL IN CHICKEN MYOFIBRILS

3.1 ABSTRACT

Solid-phase microextraction (SPME) combined with gas chomatography/mass
spectrometry (GC/MS) was evaluated for its suitability in the detection of hexanal in
chicken myoﬁbrils. Recovery of hexanal from spiked myoﬁbrils was 95% using a
5-minute sampling time, with a total analysis time of ~12 min/sample. Hexanal
partitioning into the chicken myoﬁbrils was also examined. Using SPME, the limit of
detection for hexanal in myoﬁbrils was 10 ppb with a coefﬁcient of variation rangi.1g
from 1 to 13%, and a working linear response from 10 ppb to 10 ppm (r2=0.995).
Freeze-dried chicken myoﬁbrils with added methyl linoleate (0.6 mmol/g of protein) were
stored at 50°C at water activities of 0.30 and 0.75. Samples were removed at two time
intervals, in Trial 1, Day 0, 3. 6 and 9 and Trial 2, H 0, 12, 27 and 50. Hexanal was
measured by SPME coupled with gas chromatographic separation and time-of—ﬂight mass
spectrometry (GC/MS-SPME). Samples were also analyzed using the thiobarbituric acid
reactive substances assay (TBARS) and a conjugated diene assay (CD). In both trials.
sampling time had an effect on lipid oxidation (p<0.001). In the 50 h study, differences
in oxidation were observed between the two water activities. A strong correlation

(r=0.975 and 0.938) existed between SPME and the TBARS assay in both the 9 day and

31

50 h studies. SPME offers a sensitive, rapid and solvent-free method of detecting hexanal
as an indicator of lipid oxidation in chicken myoﬁbrils.
3.2 INTRODUCTION

Lipid oxidation has long been recognized as a leading cause of quality
deterioration in muscle foods and is often the decisive factor in determining food product
storage life (Frankel, 1993). All muscle foods are susceptible to lipid oxidation: however.
the muscle foods of most concern are those with high concentrations of unsaturated fatty
acids, such as chicken (Tamura et al., 1991). The oxidative deterioration of food lipids
involves autoxidative reactions of unsaturated fatty acids, accompanied by various
secondary reactions. Hydroperoxides, the major initial products, are colorless, tasteless
and odorless. The degradation of hydroperoxides into a complex mixture of low
molecular weight alkanes, alkenes, aldehydes and ketones lead to oxidative rancidity and
the objectionable off-odors and ﬂavors characteristic of rancid food. Consumer
acceptance of meat products depends to a major extent on ﬂavor quality, and the ﬂavors
generated by lipid oxidation are generally undesirable (Ramarathnam et al., 1991).
Associated changes in quality are manifested by deterioration in texture, nutritional
quality and the production of potentially toxic compounds (Kanner, 1994).

From the perspective of the meat industry, detection of lipid oxidation needs to be
accomplished using a very speciﬁc, sensitive and rapid method. A wide spectrum of lipid
oxidation tests exists, ranging from simple organoleptic evaluations to more complex
chemical methods. The most commonly used method for measuring lipid oxidation is the

2-thiobarbituric acid reactive substances test (TBARS), which was originally thought to

32

quantify only malonaldehyde, a secondary product of lipid oxidation (Ladikos and
Lougovois, 1990). Criticisms of this test arise from its lack of sensitivity and speciﬁcity
to malonaldehyde. Other lipid oxidation compounds can react with thiobarbituric z‘.cid to
form colored complexes. leading to an overestimation of the extent of lipid oxidation.

For this reasons, the TBARS procedure may be used to assess the extent of lipid
oxidation in general, however, it should be complemented with sensory scores and a valid
analytical procedure (Gray and Monohan, 1992).

Hexanal has become a popular indicator of lipid oxidation in foods. Hexanal, one
of the secondary products formed during the oxidation of linoleic acid, dominates among
the volatile aldehydes. Hexanal is the only aldehyde that arises from both the 9 and 13-
hydroperoxide of linoleate, and from other unsaturated aldehydes formed during the
oxidation of linoleate (Shahidi and Pegg, 1994). Dupuy et al. (1987) reported that in
cooked chicken, all aldehydes increased following 5 days of storage at 4°C, with hexanal
showing the greatest increase from 0.1 to 15 ppm in white meat. Hexanal, often
described as “painty”, “grassy” and “ﬁshy”, contributes notably to undesirable off-odors
created during lipid oxidation and has a low odor threshold. In sensory tests, the
detectable odor threshold of hexanal in lean ground beef was reported to be 5.9 ppm
(Brewer and Vega, 1995), and between 5 and 10 ppm in cereal (Fritsch and Gale, 1977).

Gas chomatography (headspace GC) was used to establish strong correlations
between hexanal content, sensory scores and TBARS values in a number of meat

products, including chicken (Ahn et al., 2000), beef (Spanier et al., 1992; St. Angelo et

al., 1987), irradiated pork (Ahn et al., 2001) and ground pork (Shahidi et al., 1987).

33

Several GC techniques have been employed for the analysis of hexanal, including static
headspace techniques and dynamic purge-and-trap techniques. Using a static headspace
method, a correlation (r2) of 0.856 was reported between TBARS and hexanal in broiler
breast patties (Su et al., 1991). In pork patties, using a steam distillation method, a
correlation of 0.93 was reported between hexanal and TBARS (Ahn et al., 1998) and a
correlation of 0.97 in chicken thigh was found between a TBARS and a dynamic
headspace method (Zielinski et al., 2001). Static headspace techniques are simple and
rapid, however lack sensitivity as insufﬁcient quantities of hexanal are often obtained
(Reineccius, 1996). To overcome the decreased sensitivity of the static method, dynamic
headspace or purge-and-trap techniques utilize an adsorbent material to collect volatiles
over a period of time. While these dynamic headspace techniques are extremely
sensitive, disadvantages include excessive preparation time, difficulty of use. and
inconvenience when analyzing a large number of samples quickly (Steffen and Pawliszyn.
1996).

Solid-phase microextraction (SPME) is a new adsorption technique that has
gained widespread acceptance in analysis of volatile compounds as it overcomes the
difﬁculties experienced with traditional headspace methods (Steffen and Pawliszyn,
1996). SPME involves the extraction of analytes directly from a sample, or from the
headspace above a sample in a closed vial, onto a fused silica ﬁber coated with a
polymeric liquid phase. Following equilibration, the ﬁber containing the analytes is
removed from the vial and the analytes are desorbed in the injection port of a GC for

analysis. Using GC-SPME, analytes were calculated to be within the same concen'ration

34

range as those reported using headspace techniques, and the SPME method displayed
high reproducibility (Steffen and Pawliszyn, 1996). SPME has been used for defection
of volatile compounds in a number of products, including coffee (Yang and Peppard,
1994), tomato and strawberry ﬁ'uit (Song et al., 1998), vanilla extract (Sostaric et al.,
2000), cheese (Peres et al., 2001), oats (Sides et al., 2001), milk (van Aardt et al.. 2001).
and meat products such as cooked pork (Elmore et al., 2000) and ham (Sen et al., 1997;
Ruiz et al., 1998). Recently, SPME ﬁbers have been used for the quantitative
measurement of hexanal in cooked turkey as an indicator of lipid oxidation. Strong
correlations (r2: 0.994) between SPME/GC and TBARS were reported (Brunton et al..
2000). I

The primary objective of this study was to develop a method for the application of
SPME combined with gas chomatography/mass spectrometry (GC/MS) for the detection
of hexanal in chicken myoﬁbrils as an indicator of lipid oxidation. The speciﬁc
objectives were to 1) develop a precise and accurate GC/MS SPME method for the
quantiﬁcation of hexanal, 2) study the partitioning of hexanal between chicken
myoﬁbrils and headspace, and 3) compare hexanal concentration in chicken myoﬁbrils
determined by GC/MS SPME to malonaldehyde concentration determined by TBARS
and establish a correlation between the two. i
3.3 MATERIALS AND METHODS
3.3.1 Preparation of Chicken Myoﬁbrils

The procedure used for myoﬁbril isolation was described by Smith (1987).

Chicken breast was obtained from a local supermarket and ground in a Hobart

35

Kitchen-Aid food grinder with a 4.5 mm plate (Model KF-A, Troy, OH). The ground
chicken was blended for 90 s in 4 volumes of 0.05 M potassium phosphate buffer (0.1 M
NaCl, pH 7.0) at maximum speed in a Waring blender (Model 1120, Winsted, CT). The
slurry was stirred for a two-1 h intervals at 4°C. Following each interval. the myoﬁbrils
were centrifuged at 2000 x g for 15 min. The ﬁnal pellet was washed twice in 10
volumes of distilled water, and centrifuged at 6000 x g to precipitate the protein.
Protein content in the ﬁnal pellet was determined using the Bradford Protein Assay
(Bradford, 1976). If required for an experiment, methyl linoleate (ML) (Sigma Chemical
Co., St. Louis, MO) was stirred in at 0.6 mmol/g of protein. The spiked myoﬁbrils are
referred to as chicken myoﬁbrils-ML. The myoﬁbrils were freeze-dried (Labconco.
Kansas City, MO), vacuum packaged in polyethylene-laminated pouches (Butcher and
Pack Supply, Detroit, MI) and stored at —80°C until needed.
3.3.2 GC/MS-SPME Procedure

A SPME ﬁber assembly (57310-U, Supelco Co., Bellefonte, PA) coated with 100
um polydimethylsiloxane (PDMS) was used to absorb volatiles. The ﬁber assembly was
inserted into a ﬁber holder for manual sampling (5733OU, Supelco) and preconditioned
at 250°C for 1 h prior to initial use. All SPME samplings were conducted at 22°C. For
all experiments, the ﬁber was introduced into the headspace of the clear glass screw top
vial (Supelco), through a Mininert valve screw cap ﬁtted for the vial (33301, Supelco).
The sampling time of the ﬁber to the headspace was 5 min, after which, the ﬁber was
retracted. The desorption time in the glass-lined, splitless injection port of the GC (HP-

6890, Hewlett Packard Co., Wilmington, DE) was 2 min.

36

A Supelcowax -10 capillary GC column (15 m x 100 um, ﬁlm thickness of 0.25
mm; 24077, Supelco) and helium carrier gas at a ﬂow rate of 0.6 mL/min were used for
volatile detection. The ﬁrst 20 cm of the column was cooled with liquid nitrogen tc
cryofocus the volatiles from the ﬁber. Oven temperature was increased from 40°C to
250°C at a rate of 60°C/min, and held for 3 min. The ﬂow was maintained at 0.6
mL/min, the GC/MS transfer line was at 200°C and the inlet temperature was 250°C.
Volatiles were detected using time of ﬂight mass spectrometry (TOF MS), with electron
impact ionization (Pegasus 11, Leco Corp., St. Joseph, MI) and an ionization energy of 70
eV. Mass spectra were collected at a rate of 40 spectra/second over the mass range of m/z
33 to 250. Data were analde using LECO deconvolution software and identiﬁcation of
hexanal was authenticated by comparison with reference spectra found in the National
Institute for Standards and Technology (NIST) Mass Spectral Library (Search Version,
1.5). Total analysis time was ~12 min/sample.

The hexanal vapor standard was prepared by diluting liquid hexanal (Sigma) 10
times in squalene (Sigma) to a concentration of 83 ng/mL. An aliquot of the diluted
hexanal (0.2 ML) was applied to a piece of ﬁlter paper and placed into a 4.4-L Erlenmeyer
ﬂask. ﬁtted with a Mininert valve (Supelco). The ﬂask was sealed and the hexanal was
allowed to vaporize to provide a headspace concentration of 0.9 ppm. The hexane' vapor
standard was analyzed as above. The peak area of the standard was determined. Peak
identities were confirmed using mass spectroscopy at m/z 56. Hexanal concentrations
were expressed in ppm (uL hexanal/L).

3.3.3 Limit of Detection and Response Linearity of GC/MS-SPME

37

The linearity of the response was calculated. Vapor phase hexanal standards of
10, 5, 1, 0.1, 0.01 and 0.001 ppm were prepared in squalene. To prepare a standard
curve, an aliquot of each standard (1.5 mL) was added to a 15 mL clear glass screw top
vial (27162, Supelco) sealed with a Mininert valve (33301, Supelco) and allowed to
equilibrate for 10 min. Each standard was then sampled and analyzed using the
procedure described above. The concentration of the standard was plotted against the
resulting hexanal peak area. The coefﬁcient of variation of each standard was determined
using 6 replicates. For each standard, the mean response and standard deviation were
calculated, and from those values, the coefﬁcient of variation was determined (Kvanli,
1988).

The limit of detection of the GC/MS-SPME was estimated as the concentration of
analyte to produce a signal that was 10x that of the noise (Song et al., 1997). To
determine the limit of detection, a series of hexanal standards were prepared, ranging
from 1 ppb to 10 ppm. Six replicates of each standard was tested using the procedure
described above. The limit of detection was considered the concentration at which the
hexanal signal became less than 10x that of the noise.

3.3.4 Hexanal Loading and Repeated Sampling of Spiked Chicken Myoﬁbrils

Recovery of hexanal from spiked myoﬁbrils was determined by incubating
chicken myoﬁbrils over 0.01, 0.1, l and 10 ppm hexanal standards (Figure 3.1a). Each
standard was tested in triplicate. Brieﬂy, 1.5 mL of each hexanal standard was added to a
lS-mL screw top vial (27162, Supelco) containing 2 g of glass beads (33201, Fisher

Chemicals, Pittsburgh, PA). A 0.03 g sample of myoﬁbrils was weighed in a 4-ml-. screw

38

q ....i 1:“
_ 4 Mininert Valve

.5 . Chicken Myoﬁbrils
3i"- , , ¢“~-- Glass Beads

Hexanal Standard

 

 

I <-——— ,__--,_--_ SPME Fiber Assembly

B

’ I
l
i

l 4 - Coated-Fused Silica Fiber

- 4 . Chicken Myoﬂbrils

Figure 3.1. Hexanal loading and recovery from chicken myoﬁbrils.
A) Hexanal loading: 0.03 g of myoﬁbrils incubated over 1.5 mL of
hexanal standard B) Recovery of hexanal from spiked myoﬁbrils
following 24 hours of incubation

39

top vial (271 l 1, Supelco), and these vials were nested into the glass beads in the larger
15-mL vials. A Mininert valve (33301, Supelco) was used to seal the lS-mL screw top
vial and allow for sampling. At hourly intervals, the headspace in the vial was sampled
using SPME and GC/MS analysis was conducted. Following 24 h of incubation, the
myoﬁbrils were removed from the small vials, placed in new 4-mL screw tOp vial, sealed
with a Mininert valve (33300, Supelco) and sampled hourly by GC/MS-SPME using the
procedures described above (Figure 3.1b). Brieﬂy, the SPME ﬁber was exposed to the
headspace above the myoﬁbrils for 5 min, after which, the ﬁber was desorbed into the
injection port of the GC/MS. GC/MS analysis was conducted and recovery of hexanal
from the myoﬁbrils was determined.
3.3.5 Hexanal Partitioning

Hexanal standards from 0.01 ppm to 10 ppm were prepared in squalene. Each
standard was tested in triplicate. Aliquots of 1.5 mL of each standard were added to 15-
mL screw tOp vials containing glass beads. The vials were sealed with a Mininert valve
and allowed to equilibrate for 24 h at ambient temperature. To a second set of vials, 1.5
mL of each standard was added to a 15-mL glass vial with glass beads. Small 4-mL
screw top vials containing 0.03 g myoﬁbrils were nested into the glass beads; the V ral was
sealed with a Mininert valve and allowed to equilibrate for 24 h. Following 24 h of
incubation, each vial was sampled using the GC/MS-SPME procedure described in
Section 3.3.2. The coefﬁcients of variation were determined for each standard, and
overall. Differences in hexanal partitioning were calculated by dividing the hexanal peak

area of the standards alone by the hexanal peak area of the standards incubated with

40

myoﬁbrils.
3.3.6 Accelerated oxidation study

Freeze-dried chicken myoﬁbrils, spiked with 0.6 mmol/g of protein, prepared as
described in Section 3.3.1, were used for two separate oxidation trials. For both trials. the
samples were stored at 50°C, an accelerated storage temperature typical of military
provision storage studies (Sun et al., 2001). Two storage chambers were used for each
time point at each water activity, and replicate determinations were made for each
sample. A 0.5 g and 0.05 g sample of myoﬁbrils were weighed out, evenly distributed on
an aluminum weigh dish and transferred to a l-L Teﬂon chamber sealed with a screw lid
(Berghof of America, Coral Spring, FL). Each chamber held one-0.5 g and one-0.05 g
sample myoﬁbril sample. The samples were pre-weighed to account for any changes in
weight that may have occurred upon storage in either high or low water activity

At the bottom of the Teﬂon chambers were 10 mL of saturated salt solutions. The
water activities used represented the maximum and minimum lipid oxidation rate
observed by Smith et al. (1990) in freeze-dried chicken myoﬁbrils. Saturated salt
solutions used for humidity control were sodium chloride (aw=0.75) and magnesium
chloride (a,,=0.30) and were prepared at 50°C using the procedure described by Labuza
(1984). In each trial, the myoﬁbrils were separated into two groups (for each aw) aid
incubated in the dark. For all time points, prior to removal of the myoﬁbril sample, the
chamber headspace was sampled with the SPME ﬁber through a septum in the lid of the
chamber. Following sampling of the headspace in the chamber, myoﬁbrils were removed

and sampled.

41

In the ﬁrst trial, samples were removed and analyzed at day 3, 6, and 9. In the
second trial, samples were removed and analyzed at 12, 27 and 50 h. At each sampling
time, 0.05 g of myoﬁbrils was removed and assayed using the conjugated diene (CD)
assay. Also at each time, 0.5 g of myoﬁbrils was transferred to a 15-mL glass screw top
vial, sealed with a Mininert valve and the hexanal in the headspace was analyzed
following equilibrium using GC/MS-SPME as described in Section 3.3.2. The sample
that was analyzed using GC/MS-SPME was then analyzed using the TBARS assay.
3.3.6.1 Conjugated Diene Assay (CD)

Conjugated dienes were quantiﬁed using the method of van Ruth et al. (2000)
except with a smaller sample size. In this method, 0.05 g of myoﬁbrils were mixed with
5 ml cyclohexane and centrifuged for 3 min at 1500 x g, after which the absorbance of
the supernatant was read at 234 nm. Absorbances were calculated as hydroperoxides in
mmol per g protein, using a molar absorptivity of 26,000 for linoleate peroxides (AOC S.
1992). Results were expressed as mmol/g protein. Each sample was tested in tripl‘cate.
3.3.6.2 Thiobarbituric Acid-Reactive Substances Assay (TBARS)

The TBARS procedure of Tarladgis et al. (1960) was used with modiﬁcations of a
decreased sample size and the addition of antioxidant. Speciﬁcally. following accelerated
oxidation, myoﬁbrils (0.5 g) were homogenized for 1 minute with 48.75 mL of distilled
water containing 100 ppm Tenox 2 antioxidant (Eastman Chemical Company, Roebuck,
SC). Hydrochloric acid (4 N), glass beads and 3 sprays of antifoam (Arthur H. Thomas
Company, Philadelphia, PA) were added, and using the highest possible heat on the

distillation apparatus, 15 mL of distillate was collected. A 5 mL sample of distillate was

42

reacted with 5 mL of 0.02 M thiobarbituric acid. and held in a boiling water bath for 35
min. Upon removal and cooling for 10 min, absorbance of the solution was read at 532
nm. Each sample was tested in triplicate.

The standard used for TBARS determination of malonaldehyde was
tetramethoxypropane (TEP); under acid conditions, TEP hydrolyzes to yield
malonaldehyde (Sinnhuber and Yu, 1957). The malonaldehyde standard curve was
prepared using 0 to 9 x 10" M tetramethoxypropane (TEP). Total recovery was
established through the addition of each TEP standard to 0.5 g of chicken myoﬁbrils and
recovery of malonaldehyde following the TBARS procedure. The TBARS number was
calculated by multiplying the absorbance of the sample by a constant (Tarladgis et 11..
1960). The constant was determined as follows:

K (distill) = cone in mol/5 mL ofdistillate x MW of MA x 107 x 100

optical density wt of sample % recovery
The value of the ﬁrst term from the standard curve was found to be 3.09 x 108. With a
molar extinction coefﬁcient of IO7 (Sinnhuber and Yu, 1958), a sample size of 0.5 g and
an average 75.8% recovery, the K value was calculated to be 58.7. Results of the TBARS
assay, often expressed as mg MA/kg sample, were converted to ppm (w/w).
3.3.6.3 Accelerated Oxidation Study Data Analysis

Accelerated oxidation study statistical analyses were preformed using SAS
(Version 6.1; SAS Institute, Inc., Cary, NC). A two-way ANOVA on the log transform of

the CD. TBARS and GC/MS-SPME data was conducted. Pairwise differences were

43

tested and correlation coefﬁcients were determined between GC/MS-SPME, TBARS and
CD. Measurements for each method were taken in triplicate and were all included in the
analyses. The signiﬁcance level used in these analyses was deﬁned as p<0.05.

3.4 RESULTS AND DISCUSSION

3.4.1 Validation of the GC/MS-SPME Procedure

The response of the GC/MS-SPME procedure to hexanal concentration was linear
over the ﬁve hexanal standards tested (r2=0.995) (Figure 3.2). The limit of detection for
this method was 0.01 ppm (10 ppb). This value is comparable to the 7 ng hexanal/g (7
ppb) protein limit of detection reported in turkey by Brunton et al. (2000), and to the limit
of quantiﬁcation of 36 ng hexanal/g (36 ppb) sample reported by Song et al. (1998) in
tomato and strawberry fruit .

The precision of the GC/MS-SPME, as reﬂected in the coefﬁcients of variation of
hexanal peak areas, was found to range between 1 and 13%, with the majority of
responses falling below 10%. The higher coefﬁcient of variation was found to be at the
lowest hexanal concentration. Elmore et a1. (1997) made a similar observation in their
studies. in which compounds with small peak areas had a higher coefﬁcient of variation.
Gutheil and Bailey (1992) also reported greater coefﬁcients of variation at lower
concentrations of hexanal. The precision of this GS/MS-SPME method was comparable
to the SPME method developed for orange juice, in which coefﬁcients of variation for 17
common ﬂavor volatiles ranged from 1 and 18%, with the majority of volatiles falling
below 10% deviation (Steffen and Pawliszyn, 1996). Song et al. (1997) reported

coefﬁcients of variation of hexyl acetate, l-butanol, 2-methylbutyl acetate and alpha-

44

 

 

 

 

3
V?
N
e ..
CU
Q)
8-:
63
ax:
CU
Q)
Q.
an O
2
5...
4 I I I I
-2 -1 0 1

log hexanal concentration (ppm)

Figure 3.2. Lin 'ty of GC/MS-SPME response to hexanal
concentrations (r =0.995). Standards were prepared in squalene and
sampled using SPME. Line represents a ﬁtted regression line
(y=963 142x + 1683332). Determinations were made in triplicate.

45

famesene to range from 2 to 10% in apple ﬁuit.
3.4.2 Hexanal Loading and Repeated Sampling of Spiked Chicken Myoﬁbrils

The headspace concentration of hexanal increased during the ﬁrst 4 h of
incubation of the chicken myoﬁbrils with the hexanal standards (Figure 3.3). The
headspace concentration of hexanal in the vial increased for each of the four
concentrations tested. Following 4 h of incubation, an equilibrium within the chamber
was reached, and maintained up to 24 h.

Following 24 h of incubation with hexanal standards, the chicken myoﬁbrils were
removed and sampled to determine the amount of hexanal that the myoﬁbrils had sorbed
during the incubation time. Recovery of hexanal from the spiked myoﬁbrils was greatest
with no equilibration following transfer to the 4 mL vial, and with a 5-min SPME
sampling time (Figure 3.4). Recovery of hexanal from myoﬁbrils at the ﬁrst sampling
time ranged from 91 to 95.3%. Due to equilibration in hexanal between the myoﬁbrils
and headspace, hexanal concentration decreased with repeated sampling as it was
desorbed from the chicken myoﬁbrils into the headspace and removed from the system.
Hexanal was removed from the headspace in the vial with repeated sampling, ranging
from 29.2% to a 56.7% loss in hexanal over 6 sampling times. An even greater
percentage of hexanal might have been removed from the system had sampling continued
beyond 6 h. Volatile materials are generally held by solid substrates and the extent of
volatile loss as a function of time at a given set of conditions is mainly determined by the
energies of the interaction between the volatile and the substrate (Saleeb and Schenz.

1978).

46

 

 

+ 0.01 ppm
+ 0.1 ppm
+ lppm
+ 10 ppm

 

 

 

 

 

 

log hexanal peak area
0)
b
191

 

 

 

0 5 10 15 20 25 30
time (h)

Figure 3.3. Steady state equilibrium of chicken myoﬁbrils with hexanal
loading. Myoﬁbrils (0.03 g) were incubated over 1.5 mL hexanal
standard (0.01 to 10 ppm hexanal) prepared in squalene. The headspace
above the myoﬁbrils in the chamber was sampled every hour up to 5
hours and following 24 hours of incubation. Unﬁlled shapes represent the
first desorption of hexanal from spiked chicken myoﬁbrils following
incubation. Determinations were made in triplicate.

47

log hexanal peak area

 

 

 

 

 

r 4

 

 

 

sampling number

Figure 3.4. Effect of repeated sampling on headspace above chicken
myoﬁbrils (0.03 g) incubated over 1.5 mL of 0.01 to 10 ppm hexanal
standard prepared in squalene. Headspace was sampled 6 times over 5
hours. Determinations were made in triplicate.

48

3.4.3 Hexanal Partitioning Between Headspace and Myoﬁbrils

Hexanal partitioning was examined by comparing the hexanal concentration found
in the vial headspace above the chicken myoﬁbrils following 24 h of incubation to the
hexanal concentration of standard alone following 24 h of incubation (Table 3.1). For all
standards, a low coefﬁcient of variation was reported between peak areas, ranging from
1.6 to 7.8% in hexanal standard (mean of 5.16%), and in hexanal plus myoﬁbrils,
ranging from 1.1 to 5.5 (mean of 1.77%). The hexanal standards incubated with
myoﬁbrils had a lower coefficient of variation than the hexanal standards alone. Gutheil
and Bailey (1992) reported similar ﬁndings in their study examining the binding of
hexanal to myosin and actin. The linearity of the hexanal standard curve was unaffected
by the presence of myoﬁbrils, with a correlation coefﬁcient of 0.998 with hexanal
standard and 0.999 with hexanal standard plus myoﬁbrils.

Hexanal partitioned between the headspace and the chicken myoﬁbrils. The
slopes of the standard curves prepared from hexanal standard only. and hexanal standard
plus myoﬁbril, displayed little variation; a greater variation was observed between
intercepts of the two standard curves. The standard curve prepared from hexanal standard
plus myoﬁbril had a lower intercept, indicating decreased detection of hexanal in the
system. This may be attributed to the partitioning of the hexanal into the myoﬁbrils.
removing it from the system and from detection using the GC/MS-SPME.

Addition of the myoﬁbrils to the vials containing the hexanal standard reservoir
caused a reduction in headspace hexanal. This reduction may be accounted for by the

physical and chemical bonding of hexanal to the myoﬁbrils. as the myoﬁbrils present in

49

Table 3.1. Log peak area of hexanal alone and hexanal incubated
with chicken myoﬁbrils following equilibrium. Hexanal standards
(1.5 mL) were either incubated alone in a 15-mL vial or with chicken

myoﬁbrils (0.03 g) for 24-h at ambient temperature and sampled.

 

 

 

 

 

 

Hexanal Hexanal log % Hexanal plus % Percent loss in
concentration peak area CV myoﬁbril CV hexanal due to
(ppm) following 24-h hexanal log peak the addition of
incubation area following myoﬁbrils in

24-h incubation incubation vial

0.01 4.5 :1: 0.031 7.1 4.4 i 0.05 1.1 19.3

0.1 5.3 i 0.034 7.8 5.25 :l: 0.02 5.5 10.6

1 6.4 i 0.02 4.1 6.39 i 0.003 0.62 3.3

10 7.2 i 0.007 1.6 7.19 d: 0.004 0.90 1.8

 

 

 

 

 

 

50

 

 

the vial became a secondary repository for the hexanal. The concentration of the hexanal
standard impacted the percentage of hexanal retained by the myoﬁbrils versus the
percentage detected in the headspace. At 0.01 ppm standard, a greater percentage
difference (19.3%) was observed between the hexanal peak area of the standard alone
compared to the hexanal peak area of the standard plus chicken myoﬁbrils. This
indicated that a higher percentage of the hexanal present in the system was being retained
by the myoﬁbrils. As hexanal concentration was increased, the percentage concentration
of hexanal retained by the myoﬁbrils decreased. In 10 ppm hexanal, a 1.8% difference
existed between the standard alone and the standard plus the myoﬁbrils. Thus, while the
myoﬁbrils still sorbed hexanal and removed it from the system, the percentage of
hexanal removed from the system was lower at higher hexanal concentrations. This may
be due to the saturation of the limited number of hexanal binding sites (O’Keefe et al..
1991)

The partitioning of hexanal vapors into the chicken myoﬁbrils may be attributed
to speciﬁc binding and sorption. When a matrix sorbs a compound, the molecules are
adsorbed or absorbed in the free volume which is always present in the amorphous
regions (Hernandez-Munoz et al., 1999). While several studies in the packaging area
have examined the relationship between hexanal and its sorption into different polymer
ﬁlms (Hernandez-Munoz et al., 1999; Hernandez-Munoz et al., 1998), few studies have
studied the sorption of hexanal by a food matrix.

Aldehyde binding has been studied in numerous protein slurries, including soy

glycinin and beta-conglycinin (O’Keefe et al. 1991), soy protein (Aspelund and Wilson.

51

1983), lysozyme (Karel et al., 1975), bovine serum albumin and textured vegetable
protein (Franzen and Kinsella, 1974), whey-based and egg-based fat replacers (Schirle-
Keller et al., 1994), and myoglobin (Lynch and Faustrnan, 2000). Gremli (1974)
conducted a study using a high vacuum-shell freezing system to determine whether
different ﬂavor compounds reacted reversibly or irreversibly with soy protein. In a 5%
soy protein solution, 37-44% of hexanal was found to be reversibly bound, while less
than 5% was irreversibly bound. Franzen and Kinsella (1974) studied the binding of
carbonyl ﬂavor compounds to soy proteins using headspace gas chomatography. Results
showed that 10-21% of the hexanal added to the soy protein solution was retained by the
protein solution. In a later study, Gutheil and Bailey (1992) examined the relationship of
hexanal with the proteins myosin and actin. Their studies were conducted with high
concentrations of hexanal, ranging from 100 to 1600 ppm. Using these concentrations.
myosin was reported to bind less than 10% of the hexanal. When studying the binding of
actin to hexanal (200 to 1000 ppm), an average of 25 to 30% of the added hexanal bound
to the protein. In determining binding of hexanal to protein, the published studies all
used a hexanal solution, while this study was the ﬁrst to examine binding using hexanal
vapors.
3.4.4 Accelerated Oxidation Study
3.4.4.1 Hexanal in Headspace

During both the 9-day and 50-h oxidation study, the concentration of hexanal
increased in the chamber headspace above the myoﬁbrils (Figure 3.5). The hexanal

concentration during the 9-day study at both aw showed a plateau at day 6. In the 9-day

52

 

 

l

hexanal comentration (ppm)
0\ \O

 

 

 

 

I + Aw0.30 I
l I - Aw0.75 J
0 l
l
10 .
0
8%

 

hexanal concentration (ppm)
A

 

 

storage time at 50°C (days)

Figure 3.5. Hexanal formation in chicken myoﬁbrils containing 0.6 umol
methyl linoleate/g protein as affected by storage time at 50°C at aw of 0.30
and 0.75. The headspace above the myoﬁbrils was sampled using
GC/MS-SPME for 50 hr storage period (a) and 9 day storage period (b).
Error bars represent standard deviation.

53

storage study, the maximum hexanal concentration in the chamber (8.3 ppm) was reached
at day 6 at an a“. of 0.30. A similar trend was seen at a, of 0.75. with a maximum hexanal
concentration of 6.9 ppm also at day 6. During the 50-h oxidation study, the maximum
hexanal concentration for both water activities was reached at 50 h of storage. At a“. of
0.30, the hexanal concentration was 8.3 ppm , whereas the hexanal concentration was 1 1
ppm at a“. of 0.75.

Equilibrium is reached when the rate of volatiles moving out of a sample equals
that of volatiles moving back into the sample (Taylor, 1998). Combinations of
physicochemical parameters such as the partition and mass transfer coefﬁcients, together
with dynamic factors, determines the relative distribution of the volatile compound
between the food and air phases (Marin et a1, 1999). In a complex food matrix. true
equilibrium is rarely achieved and only under certain storage conditions. When foods
undergo a dynamic process, such as lipid oxidation, the gas phase is often not constrained
and a true equilibrium is not attained (Taylor, 1998).
3.4.4.2 Effect of Water Activity on Lipid Oxidation during Accelerated Studies

In the 9-day lipid oxidation study, no differences (p<0.05) were observed between
myoﬁbrils stored at a, 0.30 and aw 0.75. Lipid oxidation measured by TBARS
proceeded very slowly at aw of 0.30, changing little over the 50 h storage period. Hexanal
and conjugated diene data indicated that oxidation proceeded at the low water activity.
but not as rapidly as it did at the high water activity.

Lipid oxidation proceeded very slowly in myoﬁbrils stored at an a“, of 0.30.

indicating that this water activity is near the monolayer value. Theoretically. a fool is

54

least susceptible to lipid oxidation when stored at relative humidities corresponding to
monolayer moisture content. The monolayer represents a theoretical moisture content at
which the protein binds one molecule of water per polar group (Kinsella and Fox, ?. 986).
Smith et al. (1990) also made a similar ﬁnding in chicken myoﬁbrils treated with methyl
linoleate. Following 3 weeks of storage at 20°C. these researchers found that TBARS
values were lower at aw of 0.85 and 0.43

In myoﬁbrils stored at aw of 0.75, hexanal and TBARS increased (p<0.05) at
almost all time points over the 50 h storage period, indicating an increase in oxidation.
Smith et a1. (1990) observed that lipid oxidation increased most rapidly when myoﬁbrils
with added methyl linoleate were stored at aW of 0.65.
3.4.4.3 Nine-Day Accelerated Oxidation Storage Trial

The formation of conjugated dienes during storage at 50°C conﬁrmed that lipid
oxidation proceeded in the myoﬁbrils. Conjugated diene formation was affected by water
activity, storage time. and the interaction of the two factors (p<0.05). The conjugated
diene concentration reached a maximum at day 3. decreased at day 6 and did not change
at day 9, indicating the usual pattern of primary oxidation product formation during lipid
oxidation (Figure 3.6). Orlien et al. (2000) reported a similar pattern of conjugated diene
formation in rapeseed oil stored at 25°C for 25 days.

Chicken myoﬁbrils contained increasing amounts of TBARS and hexanal over the
ﬁrst 6 days of storage, followed by a decrease in both products at day 9. Hexanal and
TBARS were not affected by water activity (p<0.05). TBARS increased over time, but

no differences were observed between sample day 3 and 9. The maximum concentration

55

 

 

 

 

 

 

  

 
 
 

 

  

 

 

 

 

 

 

 

12 35
+ GCIMS-SPME (350.30)
10 - + GCJMS-SPME (V0.75) 30
—e— TBARS (a~=0.30)
g —e— TBARS (a~=0.75)
3- 8 ~ 25
t:
.2
E
S 6 r 20
O
t:
O
U
a 4. 15
g A 86-5 W
'3 £50 6e-5 l l
a 2 . E l 4 10
% E, 4e-5 1 A
13 u j / + w0.30 l
E 5, 2e-5 I / I Aw0.75 i
0 - D 1 . 5
o o . - -
‘ 0 2 4 6 8 10
storage at 50°C (days)
I 1 I l I 0
0 2 4 6 8 10

storage time at 50°C (days)

TBARS concentration (ppm)

F igure. 3.6. Comparison of the headspace hexanal concentration determined by
GC/MS and TBARS concentration of chicken myoﬁbrils stored for 9-days at
different water activities at 50°C. Chicken myoﬁbrils contained 0.6 umol methyl
linoleate/g protein. Inset: Effect of water activity and length Of storage at 50°C

on the production of conjugated dienes (CD) during the same 9-day storage

period. Determinations were made in triplicate. Error bars represent standard

deviations.

56

of TBARS was reached at day 6. In myoﬁbrils stored at aw of 0.75 and 0.30, TBARS of
29.6 ppm and 27.4 ppm, respectively, were observed. TBARS declined at day 9 to
concentrations of 20.4 and 22.4 ppm for aw of 0.75 and 0.30, respectively. Hexanal
concentration increased over time, reaching a maximum at day 6 of 7.6 at 0.75 a“. and 9.2
ppm at a“. 0.30. No differences (p<0.05) were reported between sample day 6 and 9.

The minor decrease in hexanal was attributed to the decomposition of hexanal to
hexanoic acid (Shahidi and Pegg, 1994). The results from the 00le conﬁrmed this
increase in hexanoic acid over storage time. The decrease in TBARS is attributed to
increased polymerization of the secondary products (Wen et al., 1997)

At day 6 of storage, the concentrations of TBARS observed were in general
agreement with those values presented by other researchers. Rhee et al. (1996) followed
oxidation in cooked chicken thigh stored at 4°C for 6 days. TBARS values were re ported
to increase from 0.5 to 14 ppm (w/w) by the last day of storage. In turkey thigh,
oxidation was followed over 8 days of storage at 4°C, during which TBARS values
increased from 0.5 to 1 1 ppm (w/w) (Whang and Peng, 1987). In roast chicken white
meat, Dupuy et al. (1987) reported hexanal concentrations of 14.6 ppm (w/w) following 5
days of storage at 4°C. Brunton et al. (2000) found that in a turkey homogenate stored at
4°C for 6 days, the maximtun malonaldehyde and hexanal concentrations reached were 7
and 4 ppm (w/w) respectively. Few studies have followed lipid oxidation in freeze-dried
meats stored at high temperatures (Wilkinson et al., 2001; Sun et al., 2001).

Correlations between the methods used to measure lipid oxidation are listed in

Table 3.2. As conjugated dienes are a measure of primary oxidation, poor correlations

57

Table 3.2. Correlation coefficient matrix (r) between methods used

to measure the extent of lipid oxidation in freeze-dried chicken myoﬁbrils
over 9-day storage at 50°C. Tests used were GC/MS SPME to measure
hexanal, TBARS, and CD to measure conjugated dienes.

 

 

 

 

Method GC/MS-SPME TBARS
TBARS 0.974a - -
CD 0.343 0.557
" Signiﬁcant at p<0.05

58

were observed between conjugated diene concentration with the methods used to quantify
secondary products, hexanal (0.343) and malonaldehyde (0.557). Other researchers have
reported low correlations between CD and secondary product measurements, including
Aubourg (1999) who reported a correlation of 0.57 in ﬁsh stored at -30°C for 1 year.

A strong correlation (r) of 0.974 (p<0.05) was established between hexanal and
TBARS, when both water activities were averaged. Using linear regression. individual
correlation coefﬁcients of 0.954 and 0.976 were calculated for aw of 0.30 and 0.75,
respectively.
3.4.4.4 F ifty-Hour Accelerated Oxidation Storage Trial

Conjugated diene concentration increased (p<0.05) during the 50-h study,
indicating that lipid oxidation was proceeding (Figure 3.7). Water activity, sampling time
and the interaction of the two had an effect (p<0.001) on conjugated diene formation.

During the early stages of oxidation, differences between myoﬁbrils stored at high
versus low water activity were not detected using TBARS. whereas differences in hexanal
concentration were detected. TBARS were (p<0.05) affected by water activity, time and
the interaction of the two factors. TBARS values were not affected (p<0.05) by water
activity at 12 and 27 h of storage. At 50 h of storage, water activity affected (p<0.05)
TBARS formation.

Hexanal production during storage was affected by water activity and sample time
(p<0.05). At aW of 0.75, lipid oxidation increased (p<0.05) at all time points. At a“, of
0.30. differences in hexanal concentration were observed between all time points except

at h 27. Hexanal concentration reached 7.5 ppm at aw of 0.75 and 3.9 ppm at aw of 0.30

59

10 5

4. .._ ._.#__ _.__

 

 

   
   
  
    
  

 

 

 

 

E?)
E 3 , + Aw 0.30
° -- Aw 0.75
8 1 E
A E 2 ~ 4
E ‘c’
o. 8 I
5 c:
C C.) 0 --_ . _._~.. -
g 6 ‘ o 10 20 30 40 so 60
g time (hr) _ 3
5 A
s s
o 4 - 3
i a.
a . 2 25
'° E
8 2 - E,
a o
8‘ o
g GC/MS SPME -o 30 m
g + ~ (a.- . ) . 1 if
m
I—

+ GC/MS-SPME (aw=0.75)
—e— TBARS (aw=0.30)
—B— TBARS (ah=0.75)

 

 

 

 

I I T I I I l 0
0 1 0 20 30 40 50 60
storage time at 50°C (h)

Figure 3.7. Comparison of the headspace hexanal concentration determined by
GC/MS and TBARS concentration of chicken myoﬁbrils stored for 50-h at
different water activities at 50°C. Chicken myoﬁbrils contained 0.6 umol methyl
linoleate/g protein. Inset: Effect of water activity and length of storage at 50°C
on the production of conjugated dienes (CD) during the same 50-h storage period.
Determinations were made in triplicate. Error bars represent standard deviations.

60

after 50 h of storage.

Correlations were established between hexanal, TBARS and conjugated diene
concentration during the 50 h storage trial (Table 3.3). Hexanal formation was correlated
to TBARS (r=0.938) and conjugated dienes (r=0.975) (p<0.05). The correlation (r) of
TBARS to conjugated diene formation wasO.952 (p<0.05). Correlations (r) between
hexanal and TBARS were found to be 0.60 and 0.894 for aw of 0.30 and 0.75,
respectively (Figure 3.9). The conjugated diene concentration increased as the secondary
product concentration increased, resulting in high correlations.

The increased sensitivity of the GC/MS-SPME at 12 h and 27 h compared to the
small differences in TBARS numbers at these same sample times resulted in a lower
correlation than reported in the 9-day study. Shahidi et al. (1987) used hexanal analysis
and sensory evaluation as indicators of oxidative stability and ﬂavor acceptability in
cooked ground pork Following two days of storage at 4°C. little difference between
TBARS values were reported whereas differences were detected using hexanal analysis
and sensory evaluation. The researchers went on to suggest that hexanal content is a
better measure of oxidative state than TBARS in meats during the early stages of storage.
3.4.5 SPME Applications in Lipid Oxidation Detection

In order to validate a novel lipid oxidation assessment method, it needs to l 3
compared to the classical TBARS method. Using this GC/MS-SPME method, a strong
correlation was achieved between hexanal and TBARS. Shahidi et al. (1987) reported a
strong correlation (r2 of 0.995) between TBARS and hexanal content quantiﬁed using

purge and trap GC in a slurry of stored cooked ground pork. In cooked beef. pork and

61

Table 3.3. Correlation coefﬁcient matrix (r) between methods used to
measure the extent of lipid oxidation in freeze-dried chicken myoﬁbrils
over 50-hr storage at 50°C. Tests used were GC/MS SPME to measure
hexanal, TBARS and CD to measure conjugated dienes.

 

 

 

 

Method GC/MS-SPME TBARS
TBARS 0.93 8" - -

CD 0.975a 0.952a
' Signiﬁcant at p<0.05

62

 

turkey burgers held at 4°C for 7 days, correlations between TBARS and hexanal
concentration, were 0.98, 0.99 and 0.97 respectively (Wen et al., 1997). The researchers
quantiﬁed hexanal using hexane extraction and separation by reversed-phase high
performance chomatography. Nielsen et al. (1997) published a correlation of 0.98
between TBARS and GC-SPME in minced pork over a 6-day storage time at 4°C .
Brunton et al. (2000) reported a strong correlation between SPME/GC and TBARS
(r3=0.994) when measuring lipid oxidation in cooked turkey during storage over 6 days at
4°C.

The use of SPME for analysis of volatile compounds, including those volatiles
generated during lipid oxidation, has been growing. Many different SPME procedures
have been published, varying in ﬁber use, equilibration time, and sampling temperature.
For detection of hexanal in chicken myoﬁbrils, the GC/MS-SPME method outlined in
this study employed a 22°C (room temperature) sampling temperature. Other SPME
procedures have been published describing the use increased temperatures in order to
maximize desorption. Ruiz et a1. (1998) studied the effect of extraction time and
temperature on volatile detection in ham observing that the chomatographic area of
hexanal increased with longer extraction times and higher extraction temperature (60°C).
The researchers suggested that the increase in chomatographic area may be attribuf :d not
only to the enhancement in extraction, but also to the formation of hexanal during
extraction due to the higher temperature. In their studies of lipid oxidation in minced
pork. Nielsen et al. (1997) employed a SPME method that involved sampling the pork for

15 min at 45°C. Brunton et al. (2000), in measuring lipid oxidation in cooked turkey

63

during storage, employed a 20 min elevated sampling temperature (40°C). While these
elevated temperatures enhance extraction of the hexanal from the meat sample, one must
exercise caution as hexanal can be thermally generated. Because of the high sensitivity of
the GC for hexanal detection, it is important to minimize hexanal production during
sampling, which would require the use of a low sampling temperature as was used in this
study.
3.5 CONCLUSIONS

The development of objectionable ﬂavors and odors generated by lipid oxidation
have detrimental consequences on food quality and consumer acceptability. Hexanal
content has been reported to be a sensitive and reliable indicator for the evaluation of the
oxidative status of meat products. Traditional methods of hexanal analysis by gas
chomatography, while sensitive, are labor intensive, often requiring dedicated equipment.
A precise and accurate GC/MS-SPME method for the rapid quantiﬁcation of hexanal was
described in this study. Over long-term storage of chicken myoﬁbrils at 50°C, this
hexanal analysis method was found to correlate well with TBARS. In a short term study
of chicken myoﬁbrils, the GC/MS-SPME method for hexanal appeared to be more
sensitive than TBARS during the initial stages of lipid oxidation. As early detection of
lipid oxidation is crucial in the food industry, this method shows promise as a fast and

accurate alternative to the traditional methods of hexanal analysis.

64

CHAPTER 4

VERIFICATION OF A MONOCLONAL ANTIBODY-BASED ELISA

FOR MONITORING LIPID OXIDATION IN CHICKEN

4.1. ABSTRACT

Hexanal, a secondary product of lipid oxidation, has often been used as in indicator of
lipid oxidation. Headspace-gas chromatography has typically been used to measure
hexanal; however, this method is expensive and involved. As a sensitive, rapid and
reproducible alternative, a monoclonal antibody was developed for the detection of
hexanal. The objective of this study was to compare the ability of a monoclonal
Competitive Indirect ELISA (CI-ELISA), thiobarbituric acid reactive substances assay
(TBARS) and a solid-phase microextraction- gas chromatography/mass spectrometry
(GC/MS-SPME) method for monitoring lipid oxidation in freeze-dried chicken protein.
Protein extraction and solubilization procedures were optimized to maximize the
formation of hexanal-protein adducts and percent extractable protein. Freeze-dried
myoﬁbrils (MF) with added methyl linoleate (0.6 mmol/g protein) were stored at 50°C at
water activities (aw ) of 0.30 and 0.75. Samples were removed each day over a 5 d .y
storage period. Hexanal was measured by GC/MS-SPME, TBARS and the CI-ELISA.
Lipid oxidation occurred more rapidly. and higher concentrations of hexanal and TBARS
were observed, when MF were stored at the higher water activity (p<0.05). Lipid

oxidation reached a maximum after 4 days of storage at both aw, then decreased. At an a“

65

of 0.30. 34.7 and 39.7 ppm hexanal were detected by GC/MS-SPME and CI-ELISA
respectively, after 4 days of storage. The CI-ELISA was well correlated with the GC/MS-
SPME (r=0.78) and TBARS (r=0.87) methods. The correlation of the hexanal-speciﬁc
Cl-ELISA to both GC/MS-SPME and TBARS veriﬁed the ability of the CI-ELISA to be
used as an index of lipid oxidation, offering the convenience for use in a kit to be utilized
within a food processing facility.
4.2. INTRODUCTION

Lipid oxidation is a major quality deterioration problem in all muscle foods.
affecting both the toxicity and organoleptic quality of the food. The oxidative
deterioration of muscle lipid involves the oxidation of the unsaturated fatty acid via a free
radical mechanism (Melton, 1983). In addition to the formation of hydroperoxides or
primary products of oxidation, other types of reactions may occur (Gray, 1978). The
peroxides break down to low molecular weight compounds, including alkanes, alkenes.
ketones. aldehydes and alcohols (Gray and Monahan, 1992). The carbonyl compounds
present have the greatest impact on ﬂavor and odor development in rancid foods owing to
their low odor threshold (Shahidi and Pegg, 1994). Based on sensory tests, the aldehyde
hexanal was found to have an odor threshold of 5.9 ppm in lean ground beef (Brewer and
Vega. 1995) and between 5 and 10 ppm in cereal (Fritsch and Gale. 1977). The
production of off-ﬂavors and odors were considered to be the most important
consequence of lipid oxidation in the past; however, increased attention is now being
given to the health risks that lipid oxidation may impose. Numerous studies have

reported the toxic effects of oxidized fats in animals (Yang et al.. 1998; Corpet et al..

66

1990). Primary and secondary lipid oxidation products have also been linked to diseases.
such as atherosclerosis (Khan et al., 2000) and Parkinsons’s disease (Uchida et al., 1995).

Lipid oxidation is a complex process and a number of methods exist by wh.ch
oxidative changes can be measured. Sensory methods provide the most useful
information related to consumer acceptance of the food; however, these methods are
time-consurning and display poor reproducibility (Frankel, 1993). While primary products
have been used as indicators of lipid oxidation in food products, such as cooked meat.
where oxidation occurs at an accelerated rate, the measurement of secondary lipid
oxidation products is considered a more appropriate index of lipid oxidation (Gray and
Monahan, 1992).

One of the most commonly used methods to assess lipid oxidation is the 2-
thiobarbituric acid reactive substances assay (TBARS) (Menton, 1983). In this assay. the
TBA reagent reacts primarily with malonaldehyde, a relatively minor product of
oxidation, to form a red-colored complex with an absorption maximum at 532 nm (Gray
and Monahan, 1992). The concentration of this complex is then used as an objective
measure for the evaluation of the oxidative state of cooked meats. While the TBARS
assay is a relatively rapid and simple procedure and has been highly correlated to sensory
scores (Igene et al., 1979), it has been criticized primarily for its lack of speciﬁcity
Aldehyde products other than malonaldehyde may react with thiobarbituric acid (TBA) to
produce yellow (455 nm), orange (495 nm) and red pigments (532 nm) (Sun et al., 2001).
Other compounds such as sucrose and compounds in woodsmoke have also been reported

to react with the TBA reagent (Dugan, 1961).

67

Hexanal is a dominant oxidation product of linoleic acid and has been
successfully used to follow lipid oxidation in a number of products. As hexanal arises
from both the 9- and l3-hydroperoxides of linoleate, it predominates among volatile
aldehydes formed during oxidation (Shahidi and Pegg, 1994). Strong linear relationships
were reported between hexanal content, sensory scores and TBARS in cooked pork
(Shahidi et al., 1987), chicken (Ahn et al., 2000) and beef (Spanier et al., 1992). Hexanal
analysis by conventional analysis techniques such as static headspace gas chromatography
(GC) lack sensitivity, while more sensitive dynamic methods are tedious and difﬁcult to
perform (Reineccius, 1996; Steffen and Pawliszyn, 1996). Recently, solid-phase
microextraction (SPME) has been introduced as an alternative to traditional methods
(Steffen and Pawliszyn, 1996). Direct extraction and sorption of hexanal onto a coated
silica ﬁber is followed by analyte desorption into the injection port of a GC. Because this
method is fast, inexpensive and simple to use, SPME has gained popularity in food
analysis. In the area of lipid oxidation, SPME has recently been employed for the .
detection of aldehydes in turkey (Brunton et al., 2000) and sunﬂower oil (Keszler et al..
1998). Headspace chromatographic methods, such as SPME and traditional GC
techniques, require the volatization of hexanal for quantiﬁcation. However. in food
systems, covalent binding between hexanal and proteins in the food may affect accurate
hexanal quantiﬁcation (O’Keefe, 1991).

The increased regulation of foodstuffs requires analytical methods which are easy
to perform, sensitive and speciﬁc (F ukal, 1991). Since the introduction of immunoassays,

extensive research efforts have focused on their development (Samaraj eewa et al.. 1991).

68

The development of immunoassay kits has allowed rapid and sensitive detection of
mycotoxins and microorganisms in the ﬁeld, with good agreement to standard
quantiﬁcation methods (Samarajeewa et al., 1991). The development of a commercially
available ELISA kit for hexanal detection would require several modiﬁcations, but does
show promise for development. The food industry would gain immense beneﬁts from a
hexanal ELISA kit because such kits are easy to operate, convenient, and require less
technical skill compared with conventional detection assays (Samarajeewa et al., 1991).
Recently, a Competitive Indirect-ELISA (CI-ELISA) using polyclonal (Smith et
al., 1999) and monoclonal antibodies (Zielinski et al., 2001) was developed for the
detection of hexanal. Hexanal is a hapten and non-irnmunogenic, and thus requires
conjugation with a large protein molecule, such as chicken serum albumin, in order to
become antigenic. Hexanal-conjugate production was accomplished through the
formation of Schiff bases via the lysine group on the protein, followed by stable
conjugate formation using the strong reducing agent, sodium cyanoborohydride (Jentoft
and Dearbom, 1979; Means and Feeney, 1968). Both the polyclonal and monoclonal
CI-ELISAs for hexanal were sensitive and showed potential for use in a meat system to
follow lipid oxidation (Zielinski et al., 2001). In an accelerated oxidation study in a
chicken thigh homogenate system, the polyclonal CI-ELISA for hexanal was founf to
correlate with dynamic headspace GC (r= 0.89) and with lipid oxidation determined by
TBARS (r= 0.85). While the polyclonal antibody was hexanal sensitive, it displayed
cross reactivity to pentanal (11.8%) and heptanal (86.3%) modiﬁed with chicken serum

albumin.

69

To improve the speciﬁcity of the CI-ELISA, monoclonal antibodies to hexanal
were developed (Zielinski et al., 2001). The monoclonal antibody displayed moderate
cross-reactivity with other aldehydes, reacting 76.6%, 45.0% and 37.9% to modiﬁed
heptanal, 2-t-hexenal and pentanal. respectively. In the detection of hexanal in an
oxidized chicken thigh homogenate accelerated oxidation study, hexanal contents
determined using the monoclonal CI-ELISA were not different from values determined
using dynamic headspace GC except at one time point r = 0.81). These initial studies
indicated that the use of hexanal speciﬁc monoclonal antibodies was feasible for accurate
assessment of lipid oxidation in a meat product.

The primary goal of this study was to verify the use of a monoclonal CI-ELISA
for monitoring lipid oxidation in chicken through the quantiﬁcation of hexanal. In order
to achieve this goal, speciﬁc objectives were to l) optimize the monoclonal CI-ELISA for
hexanal detection, 2) optimize protein extraction procedures for the formation of hexanal-
protein adducts, 3) conduct an accelerated oxidation study to correlate results from C1-
ELISA, the novel GC/MS-SPME method for hexanal, developed in the previous study.
and TBARS.

4.3. MATERIALS AND METHODS
4.3.1. Materials

Hybridoma media was composed of: Dulbecco’s Modiﬁed Eagle Medium
(DMEM; Sigma Chemicals, St. Louis MO), fetal bovine serum (FBS; Gibco BRL.
Rockville, Maryland), NCTC-109 (Gibco BRL), and penicillin-streptomycin solution

(100 U/mL) (Sigma Chemicals). Dimethyl sulfoxide was purchase from Sigma

70

Chemicals and tissue plastic ware were purchased from Corning Laboratory Science C 0.
(Corning, NY).

Microtiter wells (Immunolon-Z Removawells) were from Dynatech Laboratories
(Alexandria, VA). Goat anti-mouse immunoglobulin G conjugated to horseradish
peroxidase (GAM-IgG HRP) was purchased from Cappel Laboratories (West Chester.
PA), and 3,3',5,5'-tetramethyl-benzidine (TMB; T 8665) was purchased from Sigma
Chemicals Co. Fungal protease AS-L (Lot 9025-49-4), alkaline protease (Lot 3-0130-
000; Protease L-FG) and papain (Lot 3-6950-510; Liquipanol T-100) were all purchased
from the Enzyme Development Corporation (New York, NY). The fungal protease had
activity of 40,000 HUT/g, alkaline protease had activity of 560,000 Delft units/g, and
papain had an activity of 100 to 115 T.U.lmg.

Endoproteinase lys-C and trypsin were purchased from Sigma Chemicals. One
unit of endoproteinase lys-C activity was deﬁned as the amount of lys-C that would
hydrolyze 1.0 umol of N-p-tosyl-Gly-Pro-Lys-p-nitroanilide per minute at pH 7.7 at
25°C. Trypsin had activity of 1000 to 1500 BAEE units/mg solid. BioRad protein dye
was from BioRad Laboratories (Hercules, CA). Tenox 2 antioxidant was obtained from
East Chemical Company (Roebuck, SC). Hexanal, chicken serum albumin (C SA).
bovine serum albumin (BSA), picrylsulfonic acid and all other materials were purc .rased
from Sigma Chemicals.

4.3.2. Optimization of Extraction Procedures to Solubilize Myofibrils
Myoﬁbrils (15 mg/ mL) were mixed with 10 mL of 0.6 M KCl, 0.05 M potassium

phosphate (adjusted to desired pH), 200 ppm antioxidant and incubated for 2 h in a 50°C

71

shaking water bath. Following incubation, solutions were ﬁltered through 2 layers of
cheesecloth. An aliquot of the ﬁltrate was centrifuged 30 min at 20,000 x g (Morr et al..
1985), and protein content of the supernatant was determined using the Bradford protein
assay (1976) described in Section 4.3.7.
4.3.2.1 Effect of KCl on CI-ELISA

A 0.6 M KCl buffer is commonly used to solubilize myoﬁbrils (Smith et al.,
1990). BSA-hex standards containing between 3 and 69 ng hex/mL were prepared using
different dilutions of the extraction buffer (0.6 M KC1, 0.05 M potassium phosphate, pH
9.0). These standards were analyzed using the CI-ELISA to determine the effect of KC 1
on the assay.
4.3.2.2 Use of Proteases to Solubilize Myoﬁbrils

In the ﬁrst experiment, papain concentrations used were 0.05 and 0.15% (w/w) at
pH 5.5, alkaline protease concentrations were 0.05 and 0.15% (w/w) at pH 9.5, and
fungal protease concentrations were 0.05 and 0.5 % (w/w) at pH 7.4. Myoﬁbrils (‘5 mg/
mL) were mixed with 10 mL of pH-adjusted extraction buffer (0.6 M KCl, 0.05 M
potassium phosphate), and the procedure described in Section 4.3.2 was followed.

The ability of alkaline protease and a lysine-speciﬁc protease, endoproteinase lys-
C. to solubilize myoﬁbrils and maximize protein-hexanal conjugate formation was also
compared. For these studies, endoproteinase lys-C (3 units) was diluted 100, 250, 500.
1000 and 2000 times in extraction buffer adjusted to pH 7.7. The concentrations of
alkaline protease studied were 0.20%, 0.60% and 1 % (w/w), using 0.6 M KCl extraction

buffer at pH 9.0. After the addition of protease to the myoﬁbrils (0.15 mg/mL), the

72

solubilization procedure described in Section 4.3.2 was followed.

These samples were then subjected to the conjugation-reduction reaction. Brieﬂy.
8.75 mL of each sample was mixed with 246 uL hexanal and 1 mL of 800 mM sodium
cyanoborohydride dissolved in 0.1 N NaOH. Final reactant concentrations were 200 mM
hexanal and 80 mM of sodium cyanoborohydride . Samples were vortexed and held on
ice with periodic vortexing for l h. Samples were dialyzed overnight in 2 L of buffer
with 2 changes of PBS, after which the CI-ELISA was performed (Section 4.3.6.1).

Due to the expense of the endoproteinase lys—C. another protease with a lysine
cleavage site, trypsin, was studied. The endoproteinase lys-C concentration used was
1000x dilution of 3 units (based on the experiment above). The trypsin concentration
used was 0.25% (w/w) (manufacturer suggestion). Myoﬁbrils (15 mg/mL) or BSA were
mixed with 10 mL of 0.6 M KCl, 0.05 M potassium phosphate (pH 7.7) (Section 4.3.2).
The CI-ELISA was performed on each sample (Section 4.3.6.1). Percentage recovery of
hexanal was determined through comparison of spiked hexanal with recovered hexanal.
4.3.2.3 Trypsin Optimization and Inactivation

The parameters for trypsin optimization were pH (8.0 and 9.0). incubation time (2.
4. 5.5 h) and trypsin concentration (0.05, 0.1 and 0.25%). Speciﬁcally, myoﬁbrils or
BSA (30 mg/mL) were mixed with trypsin, and brought to 9 mL with extraction buffer
(0.6 M KCl, 0.05 M potassium phosphate buffer). Hexanal (246 uL) and 1 mL of 800
mM NaCNBH3 in 0.1 N NaOH were added and the protein-hexanal solution was
incubated at 50°C in a shaking water bath for the appropriate incubation time (Section

4.3.2). The protein-hexanal conjugates were analyzed using the TNBS assay to determine

73

the loss of reactive groups (Section 4.3.5).

Inactivation of trypsin by heat was examined. BSA (30 mg/mL) was mixed with
0.25 % (w/w) trypsin, and brought to 9 mL with buffer (0.6 M KCl, 0.05 M phosphate
buffer, pH 9.0). Hexanal (246 uL) and 1 mL of 800 mM NaCNBI‘I3 in 0.1 N NaOH were
added and the solution was incubated at 50°C in a shaking water bath for 2 h. Following
incubation and ﬁltering, the mixture was heated at 100, 90 and 80°C for 5 min. The
solution was centrifuged for 15 min at 20,000 x g and the supernatant used in the Cl-
ELISA. Standard curves ranging from 3 to 69 ng hex/mL were constructed using each
treatment (100, 90 and 80°C, and unheated) and compared to the BSA-hex standard curve
prepared with no trypsin.
4.3.3. Growth and Maintenance of Hybridomas

To generate a fresh supply of antibody, hybridomas were grown up each year.
The media used was Dulbecco’s Modiﬁed Eagle Medium (DMEM), containing 10% fetal
bovine serum (F BS), 1% NCTC-109 and penicillin-streptomycin solution (100 U/mL).
Hybridomas were previously frozen in 1 mL aliquots at 2.66 x 10° cells/mL in 15%
dimethyl sulfoxide (v/v) (DMSO) in F BS. These were resuspended in 10 mL of media
and centrifuged for 2 min at 200 x g. The pellet was resuspended in 10 mL of fresh
media and applied in 1 mL aliquots to a 24-well microtiter plate. The hybridomas were
incubated at 37°C with 5% C02, and when adequate growth was observed, the cultures
were split and grown in 75 mL tissue culture ﬂasks. The hybridomas were expanded
gradually and split. Upon reaching the desired volume of hybridomas, an aliquot was

removed, mixed with 15% DMSO, in F BS and frozen in liquid nitrogen. The remaining

74

hybridomas were grown until less than 20% viability was observed. Supernatant from the
hybridoma culture was collected and the antibody was puriﬁed by precipitation with 70%
ammonium sulfate and dialyzed for 2 days against phosphate buffered saline (PBS, 0.1 M
sodium chloride, 0.01 M sodium phosphate, pH 7.4) at 4°C (Zielinski, 1998). Puriﬁed
antibody was allocated and frozen at -80°C.
4.3.4. Protein-Hexanal Conjugation for Standard Preparation

The protein-hexanal conjugates were prepared using the procedure described by
Smith (1997). Brieﬂy, 0.15 g CSA and BSA was mixed with 8 mL of PBS. A 246 uL
aliquot of hexanal was added, and the mixture was diluted to 9 mL. Following vortexing.
1 mL of 800 mM sodium cyanoborohydride dissolved in 0.1 N NaOH was added to
reduce Schiff bases. Final reactant concentrations were 15 mg/mL protein, 200 mM
hexanal and 80 mM sodium cyanoborohydride. The mixture was held on ice for 60 min
with periodic vortexing, followed by overnight dialysis at 4°C with PBS. The degree of
modiﬁcation of the protein was assessed using the trinitrobenzenesulfonic acid assay
(TNBS) described in Section 4.3.5. Conjugates were diluted to a concentration of 1
mg/mL with PBS. Protein concentration was determined using the Bradford protein
assay (described in Section 4.3.7). The protein solution was allocated and frozen at
—20°C.
4.3.4.1 Use of Dialysis during Protein-Hexanal Conjugation

The necessity of a dialysis step was examined in order to minimize assay time.
For dialysis studies, 30 mg/mL BSA was added to 9 mL of PBS. Hexanal (246 uL) was

added; the reduction reaction was completed by adding 1 mL of 800 mM sodium

75

cyanoborohydride dissolved in 0.1 N NaOH. Final reactant concentrations were 200 mM
hexanal and 80 mM sodium cyanoborohydride. The solution was held on ice for 60 min
with periodic vortexing. Portions of the conjugate solutions were dialyzed at 4°C
ovemight with 2 changes of 2 L of PBS (pH 7.4), while the other half of the conjugate
solutions were not dialyzed. Standard solutions of 3 to 69 ng hexanal/mL were prepared
from the dialyzed and non—dialyzed samples and analyzed using the CI-ELISA. Standards
curves from both treatments were constructed and compared as described in Section

4.3.9.

4.3.5. Trinitrobenzenesulfonic Acid Assay (T NBS)

The degree of modiﬁcation of the protein was determined by measuring the loss
of reactive groups by the TNBS acid assay described by Habeeb (1966) and modiﬁed by
Smith et al. (1999). In this assay, a 0.25 mL aliquot of protein sample (0.15 mg/mL) was
added to 1 mL of 4% NaHCO3, 0.75 mL 0.5 M PBS buffer (pH 7.4) and 1 mL of 0.1 %
TNBS (picrylsulfonic acid). The solution was vortexed and incubated at 40°C in a,
circulating water bath for 2 h. Following incubation, 1 mL of 10% SDS was added to
each sample, followed by the addition of 0.5 mL 1N HCl. The solution was vortexed and
the absorbance was read at 335 nm. The concentration was determined from a standard
curve of leucine solutions ranging from 10 to 300 um. Results were expressed as moles
of reactive amino groups per mole protein.

4.3.6 CI-ELISA Optimization
4.3.6.1 Indirect Competitive ELISA (CI-ELISA)

The indirect competitive ELISA procedure for hexanal was adopted and modiﬁed

76

from Zielinski (1998). Microtiter plates were coated overnight at 4°C with 100 uL of
hexanal—chicken serum albumin (CSA-hex) conjugate (1 mg/mL). prepared in 0.1 M
carbonate buffer (pH 9.5). Following incubation, plates were washed 4 times with
PBS-Tween (0.05% Tween-20), pH 7.2. To minimize non-speciﬁc binding, plates were
blocked using 300 uL of PBS-casein (0.5% casein, pH 7.2) for 30 min at 37°C. Plates
were then washed 4 times with PBS-Tween. Puriﬁed monoclonal antibody (50 uL of 1.6
ug protein/mL in PBS, pH 7.2), together with 50 uL of CSA-hex or sample, were added
to each well and incubated for 60 min at 37°C. Plates were washed 6 times with
PBS-Tween. Secondary antibody, GAM-IgG HRP diluted 1:500 in 0.5% casein in PBS
was added to each well at 100 uL. The plates were incubated for 30 min at 37° and then
washed 8 times with PBS-Tween prior to substrate addition. Visualization was
accomplished using 100 uL of the substrate 3,3',5,5'-Tetramethyl-benzidine (TMB).
Following a 10 min incubation at room temperature, the absorbance of each well was
measured at 650 nm using a THERMOmax plate reader (Molecular Devices. Menlo Part.
CA).

On each plate, a standard curve from 14 to 1000 ng CSA-hex/mL (3 to 69 ng
hex/mL) was constructed. From the standard curve and using a 4-parameter curve ﬁt. the
concentration of CSA-hex in each unknown sample was determined using SoftMar.
software (version 2.34). Using the concentration of CSA-hex (ng/mL) determined from
the standard curve and the formula below, the concentration of hexanal (ng hex/mL) in
the sample was calculated.

4.3.6.2 Calculation of Hexanal Concentration from the CI-ELISA

77

Hexanal quantiﬁcation involved conjugation between the lysine groups on the
protein and hexanal. The standard curve was constructed using CSA, thus a calculation
was made to convert CSA concentration to hexanal concentration based on the moles of
lysine modiﬁed per mol of CSA. The following formula was used to convert protein

concentration to hexanal concentration:

gg protein x46 mol lysine x lmol protein x Lgm x 10° mmol = gmol lysine
mL 1 mol protein 66,300 g 10° mg 1 mol mL

 

The umol of lysine are multiplied by the percent modiﬁcation to account for differences
in lysine modiﬁcation during conjugation. Since one mole of lysine will bind one mole
of hexanal, the following formula describes the conversion from umol/mL of hexanal to

ng/mL of hexanal:

leOI hexanal x 100 g (MW of hexm x moi x ﬂig = ng hexanal
mL 1 mol 10° mmol l g ml.

The mol of lysine in each protein were determined using amino acid sequences from the
National Center for Biotechnology and Information (CBI; www.mcbinlmnihgov). The
mole percent of lysine was multiplied by the molecular weight of the protein to obtain the
molecular weight contributed by lysine. This was divided by the molecular weight of
lysine to obtain the moles of lysine per mole of protein. The molecular weight used for

C SA was 66,300 daltons and for BSA, 66,000 daltons. CSA contains 47 lysine groups.

while BSA contains 42 lysine groups.

78

4.3.6.3 Precision of the CI-ELISA

Intra- and inter-assay precision of the CI-ELISA was determined using the method
described by Deshpande (1996). Standard solutions of 0, 3.4, 6.7, 34.4 and 68.8, 171.9.
343.8, and 687.5 ng hex/mL were prepared from CSA-hex conjugates, corresponding to
100, 96, 87, 79, 66, 45, 27, 8, 3 and 1 percent of the maximum binding respectively. The
intra-assay variability was determined using 8 replicate wells and the inter-assay
variability was determined using 8 replicate wells from 13 plates. The average, standard
deviation, and coefﬁcient of variation, an indicator of variability, were determined for all
replicates and at each concentration.
4.3.7. Bradford Protein Assay

In this assay, 10 uL of diluted sample was added to a 96-well microtiter plate. A
200 uL aliquot of BioRad protein dye was added to each well and mixed. Following 5
min of incubation, wells were read at 590 nm. Protein concentration of the supernatant
was calculated using a BSA standard curve, prepared using concentrations from 100 to
1000 pg BSA/mL. Percent soluble protein was calculated by dividing the resulting
protein concentration by the starting protein concentration.
4.3.8. Accelerated Oxidation Study

Myoﬁbrils with 0.6 mmol methyl linoleate/g of protein were prepared as
described in Section 3.3.1. Freeze-dried myoﬁbrils were separated in 2 subsarnples and
incubated in the dark in desiccators containing saturated salt solutions. Samples were
incubated at 50°C, a temperature typical of military provision storage studies (Sun et al..

2001). Saturated salt solutions used for humidity control were sodium chloride (a\,=0.75)

79

and magnesium chloride (aw=0.30) (Labuza, 1984). Duplicate desiccators for each water
activity were used. Six beakers (80 mL) each containing 3 g of myoﬁbril were prepared
and placed in each desiccator.

Sample beakers were removed and myoﬁbrils analyzed at Day 0, l, 2, 3. 4, and 5
of storage. Hexanal was determined by GC/MS-SPME, using the procedure described in
Section 3.3.2. The same sample was then used for TBARS analysis as outlined in Section
3.3.6.2.

To prepare samples for CI-ELISA, 0.3 g myoﬁbrils were mixed with 0.25 %
trypsin (w/w), and 50 ppm Tenox—2 antioxidant in 10 mL of 0.6 M KCl in 0.05 M
potassium phosphate buffer, pH 9.0. Sodium cyanoborohydride was added to a ﬁnal
reactant concentration of 80 mM. The solution was incubated for 2 h at 50°C, after which
it was ﬁltered through 2 layers of cheesecloth. The trypsin was inactivated by heating the
ﬁltrate for 5 min at 90°C. Following cooling on ice, the solution was centrifuged at
20.000 x g for 15 min. Appropriate dilutions (10 or 100 times) of the supernatant were
made and the CI-ELISA was conducted.

4.3.9. Data Analysis

In order to assess the effect of dialysis and proteases on the Cl-ELISA,
comparisons were made to the CSA-hex standard curve prepared with PBS. Values
falling within two standard deviations of the mean of the control standard curve (the
measurement error) were not considered to be markedly different, indicating that the
treatment was not detrimental to the standard curve (Deshpande, 1996). Each treatment

level was analyzed in triplicate. The limit of detection of the CI-ELISA was determined

80

as the lowest concentration of hexanal which was greater than 2x the measurement error
of 0 ng hexanal/mL.

For the accelerated oxidation study, data were analyzed using SAS (Version 6.1 ;
SAS Institute, Inc., Cary, NC). A two-way ANOVA was performed, and signiﬁcance
was deﬁned as p<0.05. Data were log transformed in order to ensure homogeneity of
variance. Pairwise differences were tested using Tukey’s adjusted differences test to
control the pairwise error. Correlations were made between TBARS, GC/MS-SPME and
ELISA.
4.4 RESULTS AND DISCUSSION
4.4.1 CI-ELISA Optimization
4.4.1.1 Limit of Detection and Reproducibility

Determination of hexanal concentration from the CSA-hexanal standard curve
required the conversion formula described in the materials and methods section (Section
4.3.6.2). The limit of detection of the CI-ELISA was 1 ng hexanal/mL. supporting the
same value published by Zielinski et al. (2001). Compared to the polyclonal antibody
limit of detection of 7.4 ng/mL (Smith et al., 1999), the monoclonal antibody was shown
to be more sensitive. In general, the limit of detection of ELISAs is low, often in the
ng/mL range (Samarajeewa et al., 1991). In ELISAs for both lactate dehydrogenase
(Wang et al., 1995) and a staphylococcal enterotoxin (Morissette et al., 1990), the
minimum detection limit was 10 ng/mL. The sensitivity of immunoassays for
mycotoxins in food is higher, with limits of detection as low as 0.1 ng aﬂatoxin/mL

(Samarajeewa et al., 1991).

81

The working range of the CI-ELISA, as demonstrated by the percentage binding
response, ranged from 3 to 69 ng/mL corresponding to concentrations of 14 to 1000 ng
CSA-hexanal/mL (Figure 4.1). This working range was comparable to the working range
determined by Zielinski et al. (2001) of 1 to 50 ng hexanal/mL. The working range of the
hexanal polyclonal antibody was found to range to 740 ng hexanal/mL (Smith et al..
1999)

The reproducibility of the CI-ELISA within the assay (intra-assay) and between
assays (inter-assay) was calculated (Table 4.1). Intra-assay coefﬁcients of variation
ranged from 4.2 to 5.8%, while inter-assay coefﬁcients of variation were slightly higher,
ranging from 5.0 to 6.2%. The inter-assay precision had slightly higher coefﬁcients of
variation than the intra-assay precision, which is common as variables are introduced day
to day and run and run (Deshpande, 1996). These coefﬁcients of variation for the ..exanal
Cl-ELISA were similar to those published by other researchers. A previous study using
monoclonal-based CI-ELISA for hexanal reported coefﬁcients of variation ranging from
4.2 to 6.4% intra-assay, and 5.1 to 7.7% inter-assay (Zielinski, 1998). In an ELISA of
malonaldehyde-modiﬁed low density lipoproteins, Swets et al. (2001) reported intra- and
inter-assay coeffﬁcients of variation of 3.3 and 10.6% respectively. Similarly. Salomon
et al. (2000) reported intra-assay coefﬁcients of variation from 2 to 8%. and an inter-assay
coefﬁcient of variation of 5%.
4.4.1.2 Effect of Dialysis

The original protein-hexanal conjugation procedure employed overnight dialysis

of the conjugate to remove excess sodium cyanoborohydride (Zielinski. 1998). To

82

100

 

 

 

 

so «
°\°

c 60 «
.2
I—§
.:
.E

on 40 -
.E
.0
.E
.0

20 «

O -1

I I

1 10 100 1000
log hexanal concentration (ng/mL)

Figure 4.1. The percent binding inhibition produced in the CI-ELISA from
concentrations of hexanal. A stock solution of 1mg/mL CSA was prepared,
reduced, dialyzed and diluted. Standards between 0 and 690 ng hex/mL were
prepared. Each point represents 6 replicates (13:0.975).

83

Table 4.1. Reproducibility of the CI-ELISA as described by the coefﬁcient
of variation for intra-assay and inter-assay reproducibility.

 

Hexanal concentration Intra-assay' Inter-assayb
(ng/mL) coefﬁcient of variation coefficient of variation

(‘70) (”/01

0 4.2 5.0

50 5.2 6.2

100 4.9 5 .1

250 5.8 5.8

500 5.1 5.2

1000 5.5 5.7

 

 

 

a - lntra-assay precision was calculated using 8 replicate wells
° - Inter-assay precision was calculated using 8 replicate wells from 13 different plates

84

minimize assay time, the necessity of the dialysis step was evaluated through comparison
of standard curves prepared with dialyzed and non-dialyzed BSA (Figure 4.2). No effect
of dialysis was seen on the standard curve. indicating that the sodium cyanoborohydride
did not interfere with the antibody reaction during the CI-ELISA, either due to the nature
of the chemical or the sample dilution prior to analysis. Dialysis was eliminated from the
conjugation procedure, considerably decreasing assay time.

4.4.2 Optimization of Extraction Procedure

4.4.2.1 Effect of Extraction Buffer

A major problem facing the application of monoclonal antibodies in food research
is the solubilization of the sample (Paraf, 1992). Myoﬁbril solubility is also decreased
during storage, as observed by Smith et al.(l990) and Xiong and Brekke (1989), thus the
low solubility of the freeze-dried myoﬁbrils was a signiﬁcant hurdle.

The ionic strength and the pH of the solubilization buffer are important factors in
maximizing solubility. Several studies have been conducted varying the characteristics of
the solubilization buffer, including employing phosphates (Xiong et al., 2000), urea (Lin
and Park, 1998), and low ionic strengths (Stanley et al.. 1994). In preliminary
experiments, none of these buffers were found to be effective on myoﬁbril solubilization.
The traditional method to solubilize myoﬁbrillar proteins is through the use of a buffered
0.6 M KCl solution (Smith et al., 1990; Xiong and Brekke, 1989). The effect of KCl
concentration on the antibody was assessed. Standards in different dilutions of 0.6 M
KC1, 0.05 M phosphate buffer were prepared (Figure 4.3). Only the standard curve

prepared using undiluted 0.6 M KC1 in 0.05 M phosphate buffer was found to fall outside

85

 

 

 

 

 

 

 

 

 

—I- - No dialysis
+ Overnight dialysis

e: 2 «

I:

O

W

8

Q)

U

C

(U

.D

L—

o

3

cc 1 ‘

0 I V T 1

 

0.0 0.5 1.0 1.5 2.0

log hexanal concentration (ng/mL)

Figure 4.2. Effect of overnight dialysis following conjugation/reduction on the
Cl-ELISA standard curve. The CSA-hex conjugate (1 mg/mL) was prepared,
reduced and dialyzed (control was not dialyzed) and standards from 3 to 69 ng
hex/mL were prepared. Each point represents three replicates.

86

 

 

 

 

 

 

 

 

 

2 -<
E
r:
o
E f
cc
0
s:
cc
2 1 ~
(I)
.0
as
+, 0.05 M phosphate buffer
—a— 0.6 M KCl in 0.05 M phosphate buffer
+ 0.06 M KCl in 0.05 M phosphate buffer
—e— 0.006 M KCl in 0.05 M phosphate buffer
0 I f F I

 

0.0 0.5 1.0 1.5 2.0

log hexanal concentration (ng/mL)

Figure 4.3. Effect of the addition of different concentrations of 0.6 M KC1 in 0.05
M potassium phosphate on the CI-ELISA standard curve. From a stock of 1
mg/mL CSA-hex, standards from 3 to 69 ng hex /mL were prepared in different
dilutions 0.6 M KC1 in 0.05 M potassium phosphate. The control standard curve

was prepared in 0.05 M phosphate buffer. Determinations were made in triplicate
using the CI-ELISA.

87

the measurement error of the control standard curve, indicating that the undiluted buffer
did have an effect on the standard curve. Dilutions of the buffer above 10 x did not have
an effect on the CI-ELISA standard curve. Therefore, the minimum sample dilution for
CI-ELISA was considered to be 10x.
4.4.2.2 Effect of Proteases

The use of proteases was evaluated to further increase myoﬁbril solubility. In the
first study, papain, fungal and alkaline protease were assessed under their optimum
conditions as described by the manufacturer (Table 4.2). Alkaline protease at 0.2 and
1.0% were the most effective concentration, increasing protein solubility by about 4-fold.
Alkaline protease has been used in previous studies to increase the solubility of spray-
dried lactalbumin (Ennis and Harper, 1986) and cooked chicken thigh (Zielinski, 1998).

Because of its effectiveness at increasing protein solubility, alkaline protease was
compared to endoproteinase lys-C in the next study (Table 4.3). Available lysine groups
are important during the conjugation reaction as lysine is the amino acid responsible for
the formation of Schiff bases (Jentoft and Dearbom, 1979). Endoproteinase lys-C is a
lysine speciﬁc enzyme. It hydrolyzes peptide bonds on the carboxyl side of lysine, and
may leave more available lysine for conjugation (Sorensen et al., 1995). In hexanal
spiked chicken myoﬁbrils, endoproteinase lys-C was found to be approximately twice as
effective at increasing protein solubility and hexanal recovery than alkaline protease.
Both greater protein solubility and hexanal recovery indicated the effectiveness of the
protease at protein digestion.

Due to the expense of the endoproteinase lys-C, trypsin was evaluated as an

88

Table 4.2. The effect of papain, fungal protease and alkaline protease on
the solubility of chicken myoﬁbrils. Myoﬁbrils (0.15 g/mL) were mixed
with 10 mL of extraction buffer and protease and incubated at 50°C for 2
hours in a shaking water bath. Following straining through cheesecloth, %
soluble protein was determined. Determinations were made in triplicate.

 

Protease pH Concentration' °/o Soluble Protein”

Papain 5.5 0 5.8 i 1.7

5.5 0.05 5.5 i 0.39

5.5 0.15 5.1i 0.01
Fungal Protease 7.4 0 5.3 i 0.89

7.4 0.05 5.2 i 1.2

7.4 0.5 4.9 i 1.1
Alkaline Protease 9.5 0 3.9 i 0.9

9.5 0.2 16.0 i 1.8

9.5 1 18.7 i 2.5

 

 

 

a - Enzyme concentration expressed as % of total protein in chicken myoﬁbrils
b - Theoretical maximum protein recovery was 0.03 g/mL

89

Table 4.3. Effect of alkaline protease and endoproteinase lys-C on the
soluble protein and hexanal detection in chicken myoﬁbrils. Myoﬁbrils (15
mg/mL) were mixed with 10 mL of extraction buffer and protease and incubated
in a 50°C circulating water bath for 2 hours. Following straining through
cheesecloth, soluble protein was determined. Myoﬁbrils were conjugated to
200 mM hexanal, Schiff bases reduced, and hexanal concentration was
determined using the CI-ELISA; the amount of spiked hexanal was compared
to the amount of recovered hexanal. Determinations were made in triplicate.

 

Protease

alkaline protease 9.5
9.5
9.5

endoproteinase Lys-C 7.7
7.7
7.7
7.7
7.7
7.7

 

pH Concentration'

0.05
0.15

003
omz
0mm
own

cams

% Soluble
Protein”

35:10
265:58
148:2]
45:071
388:44
356i22
324i56
394i36
306:0]

% Hexanal
Recoveryc

11.3 i 0.05
17.7 i 0.12
24.0 i 0.11
13.3 i 0.52
58.6 i 0.74
54.5 i 0.66
63.9 i 5.6
59.9 i 0.58
53.3 i- 0.89

 

a — Alkaline protease concentration expressed as weight protease/ g protein in chicken
myoﬁbrils. Endoproteinase-Lys-C concentration expressed as dilution of 3 units.
b — Theoretical maximum protein recovery was 0.03 g/mL

c - Theoretical maximum hexanal recovery was 7.51 mg hexanal/mL

9O

 

alternative protease as it also cleaves at the lysine group (Kunitz and Northrop, 1948). In
BSA, the addition of either trypsin or endoproteinase lys-C had little impact on the
percentage soluble protein and the percentage recovery of hexanal (Table 4.4). However.
in chicken myoﬁbrils, the addition of either trypsin or endoproteinase lys-C increased
protein solubility by 9.1-fold and 7.9-fold, respectively. Similarly, hexanal recovery was
markedly increased in chicken myoﬁbrils when either enzyme was added during the
extraction process. Therefore, trypsin was selected as the protease for protein
solubilization.
4.4.3 Optimization of Trypsin Digestion

The effectiveness of trypsin on improving protein solubility and protein-hexanal
conjugation was optimized by varying incubation times, pH of extraction buffer (0.6 M
KCl in 0.05 M phosphate buffer) and trypsin concentration (Table 4.5). For BSA (data
not shown), the degree of protein modiﬁcation determined using the TNBS assay and
percent soluble protein were all greater than 95% for each trypsin optimization parameter
tested. Greater percent modiﬁcation indicated the loss of reactive groups, consequently.
the increase in the formation of conjugates. The percent modiﬁcation of chicken
myoﬁbrils at pH 8.0 at varying trypsin concentrations and incubation times was similar to
modiﬁcation at pH 9.0 under the same conditions; at pH 8.0, modiﬁcation increased from
69.2% to 91.7%, while at pH 9.0, modiﬁcation increased from 67.5% to 89.3%. S luble
protein showed a greater increase under pH 9.0, increasing from 56% to 96.9% over the
range of conditions. At pH 8.0, solubility increased from 55.2% to 85.4% over the

parameters tested. The ﬁnal extraction conditions selected were a 2 h incubation time at

91

Table 4.4. Effect of trypsin and endoproteinase lys-C on protein solubility and
hexanal recovery in spiked chicken myoﬁbrils. Myoﬁbrils (0.6 g) or BSA
were mixed with 20 mL of buffer (pH 7.7). Treatments included the addition
0.25% (w/w) trypsin or 0.005% (v/v) Endoproteinase Lys-C. Samples were
incubated at 50°C for 2 hours in a shaking water bath, strained through
cheesecloth, and soluble protein determined. The conjugation and reduction
reactions were performed, and hexanal concentration was determined using the
CI-ELISA. Recovery of hexanal was determined by comparing the amount of
hexanal used for spiked and the amount determined using the CI-ELISA.
Determinations were made in triplicate.

 

Sample % Soluble Protein' % Recovery of Hexanal”
BSA 91.2i7.2 91.3 i5.5
BSA '1' Lys-C 95.3 i 2.2 77.5 i 7.1
BSA + trypsin 94.7 :t 0.6 80.5 i 6.2
Myoﬁbrils 4.8 i- 1.1 5.7 i 2.8
Myoﬁbrils + Lys-C 38.7: 0.51 38.4 i 9.9
Myoﬁbrils + trypsin 43.9 i 1.3 48.7 :1: 10.6

 

 

 

a — Theoretical maximum protein recovery of 0.03 g/mL
b - Theoretical maximum hexanal recovery was 7.51 mg hexanal/mL

92

Table 4.5. Protein solubility as affected by trypsin concentration (0.05. 0.1

0.25 % w/w), pH level (8.0 and 9.0), and incubation time at 50°C

(2, 4, 5.5 hours). Myoﬁbrils (0.15g/mL) were mixed with 10 mL of 0.6 M
KC1 in 0.05 M phosphate buffer, trypsin, and hexanal, making a ﬁnal
concentration of 200 mM hexanal. Solutions were conjugated, reduced and
incubated at 50°C for the appropriate time. Solutions were analyzed for % soluble
protein using Bradford Protein Assay, and % modiﬁcation using the TNBS Assay.
Determinations were made in triplicate.

 

trypsin
conc %

w/w

0
0
0
0.05
0.05
0.05
0.]
0.1
0.]
0.25
0.25
0.25

 

incubation
time (hrs)

5.5

°/o modiﬁcation of protein

pHSD

682:35
692:27
702:3]
7L2iL1
7L4i14
7L2i24
702:32
722:29
732i14
879i09
905:08
968:10

pH 9.0

66.8 i 1.4
66.8 i 1.7
68.9 i 1.3
68.6 i 0.9
69.2 i 1.2
70.5 i 1.3
71.5 i 1.2
71.8 i 1.8
83.7 i 2.1
88.5 i 1.7
87.2 i 1.3
92.3 i 1.1

% soluble protein'

pHSD
552:2]
552iL9
552:20
552:23
735i2l
759:24
780iL9
780iL5
802:12
900i30
850i28
8L2i26

pH 9.0
55.2 i 1.0
55.4 i 1.8
57.4 i 2.1
73.5 i 1.9
79.7 i 2.0
79.5 :19
89.6 : 2.0
89.2 i 1.4
91.9 i 1.2
99.1 i 3.1
95.6 i 1.9
96.1 i 2.2

 

a - Theoretical maximum protein recovery of 0.03 g/mL

93

 

pH 9.0 buffer (0.6 M KCl in 0.05 M phosphate buffer), with a trypsin concentration of
0.25% (w/w). In the enzymatic hydrolysis of shrimp using trypsin, Simpson et al. (1998)
listed similar optimal condition; 0.25% trypsin at pH 8.2 for 2.5 to 3 h at 40°C.

Heat was used to inactivate the trypsin prior to the CI-ELISA in order to prevent
proteolysis of the antibody. Depending on the purity, trypsin can be either irreversibly or
reversibly denatured by temperature or pH (Kunitz and Northrop, 1948). Crude trypsin
preparations are completely and permanently inactivated at temperatures above 70°C
while solutions of puriﬁed trypsin may be heated to boiling for a short time without a
permanent loss in activity (Kunitz and Northrop, 1948). In these studies, pure trypsin was
used. necessitating a higher temperature of inactivation.

BSA was used for these studies, as BSA is soluble without the aid of trypsin and it
simpliﬁed the protein-conjugation procedure. The BSA showed that the CI-ELISA was
still valid with the addition of trypsin and still recognized the hydrolyzed protein.
Heating at 90 and 100° both had a similar effect on the standard curve and were not
markedly different from standard curve prepared from the non-trypsin control (Figure
4.4). Low absorbances of the CI-ELISA standard curve indicated the proteolysis «,f the
antibody by the trypsin. A heating time of 90°C for 5 min was selected for the remaining
studies in order to minimize the harsh effect of heat on the protein. The optimized
procedure for the CI-ELISA is summarized in Figure 4.5.

With Western Blot analysis, Zielinski (1998) showed that the antibody displayed
recognition of the hexanal modiﬁed salt soluble protein but showed no recognition of the

native salt soluble protein. Also, the hexanal modiﬁed protein did not migrate as far

94

Table 4.5. Protein solubility as affected by trypsin concentration (0.05. 0.1

0.25 % w/w), pH level (8.0 and 9.0), and incubation time at 50°C

(2, 4, 5.5 hours). Myoﬁbrils (0.15g/mL) were mixed with 10 mL of 0.6 M
KC1 in 0.05 M phosphate buffer, trypsin, and hexanal, making a ﬁnal
concentration of 200 mM hexanal. Solutions were conjugated, reduced and
incubated at 50°C for the appropriate time. Solutions were analyzed for % soluble
protein using Bradford Protein Assay, and % modiﬁcation using the TNBS Assay.
Determinations were made in triplicate.

 

trypsin
conc °/o

w/w

O
O
O
0.05
0.05
0.05
0.1
0.1
0.1
0.25
0.25
0.25

 

incubation
time (hrs)

5.5

% modiﬁcation of protein

pH8£
682:35
692:27
702:1]
7L2:L1
7L4:L4
7L2:24
702:12
722:29
732:L4
879:09
905:08
968:LO

pH9o

668:L4
668:L7
689:L3
686:09
692:L2
705:L3
7L5:L2
7L8:L8
837:2]
885:L7
822:L3
913:11

% soluble protein'

pHSﬂ
552:2]
552:L9
552:20
552:23
735:2]
759:24
780:L9
780:15
802:L2
900:30
850:28
8L2:26

pH 9.0
55.2 : 1.0
55.4 : 1.8
57.4 : 2.1
73.5 : 1.9
79.7 : 2.0
79.5 : 1.9
89.6 : 2.0
89.2 : 1.4
91.9 : 1.2
99.1 g 3.1
95.6 : 1.9
96.] : 2.2

 

a - Theoretical maximum protein recovery of 0.03 g/mL

95

 

Weigh myoﬁbrils (0.3 g)

l

Mix myoﬁbrils with:
0.25% (w/w) trypsin,
9 mL 0.6 M KCl, 0.05 M potassium phosphate buffer
200 ppm antioxidant
1 mL of 800 mM NaCNBH3 in 0.1 N NaOH

i

Incubate 2 h at 50°C

Filter through 2 layers of cheesecloth

1

Heat at 90°C for 5 min. Centrifuge at 20.000 x g for 20 min

i

Dilute the supernatant. Assay using the CI-ELISA

Figure 4.5. Flow chart of the optimized CI-ELISA method

96

through the gel as the unmodiﬁed protein, consistent with the higher molecular weight of
the modiﬁed protein due to the addition of hexanal.
4.4.4 Accelerated Oxidation Study

Hexanal, measured by GC/MS-SPME and CI-ELISA, and TBARS increased in
the chicken myoﬁbrils during the 5 days of storage at 50°C. Over the 5 day study. lipid
oxidation occurred in myoﬁbrils stored at aw 0.30 (Figure 4.6) and myoﬁbrils stored at a“.
0.75 (Figure 4.7). For both aw, changes in hexanal production between each preceding
sample day were detected by the CI-ELISA. The GC/MS-SPME method detected
signiﬁcant differences in hexanal production between all sample days at both a“. except at
day 3. TBARS did not differ from the preceding day at day 1 and 5 for either a“, (p<0.05).

The aw of storage was also a factor in the progression of oxidation. For both water
activities and all methods, lipid oxidation proceeded slowly initially, reaching a
maximum at day 4 of storage. Both CI-ELISA and TBARS showed differences in
hexanal and malonaldehyde production, respectively, between the two water activities
from day 2 to 5 (p<0.05). GC/MS-SPME data demonstrated differences in hexanal
production between water activities at day 2, 3 and 4 (p<0.05).

The three methods were highly correlated (Table 4.6). CI-ELISA and GC/MS-
SPME had an overall correlation (r) of 0.78, while CI-ELISA and TBARS had an overall
correlation of 0.87. The CI-ELISA and GC/MS-SPME were more strongly correlated in
myoﬁbrils stored at 0.30 21,, than those stored at 0.75 aw. The decreased correlation may
be attributed to lower solubility of the myoﬁbrils stored at 0.75 aw. At a“. 0.30, the C I-

ELISA and GC/MS-SPME had a correlation of 0.95, while at aw 0.75, the correlation was

97

 

 

 

 

 

 

 

 

<
a
{j 60 14
1:
g 50 _ + TBARS * 12
z; + GC/MS-SPME
~_ + CI-ELISA . 10 E
w 40 ~ g;
w v
.23: 8 .5.
J ...
’9 30 ‘ .‘é’
3‘ a
E E
a 20 3 :3
Z 4 a:
o <
a; 10 , 53
3; -]
5 // ~ 2
E) I
8 0 q ‘
To 0
C
:6
x
U
.2: I T r I I I

o 1 2 3 4 5 6

storage time at 50°C (days)

Figure 4.6. Comparison of the hexanal concentration determined by
GC/MS-SPME and CI-ELISA, and TBARS of chicken myoﬁbrils stored
for 5 days at 50°C. GC/MS-SPME quantiﬁed headspace hexanal
concentration and CI-ELISA quantiﬁed hexanal content within the
myoﬁbrils. Chicken myoﬁbrils contained 0.6 umol methyl linoleate/g
protein and were stored at aw of 0.30. Determinations were made in
triplicate.

98

 

 

 

 

 

 

 

 

<

Q

..1

”j 100 18
U

'3 - 16
L: 80 _ + TBARS

E + GC/MS-SPME - 14 A
m + CI-ELISA g
c}: » 12 3-
E 60 - ' g
3 ~ 105;
3‘ E
A -< " 8 Q
E 40 S
3 o
5 ~ 6 2
= 20‘ 2
3:. h 4 (—
C

8 l

3 I'd _ 2

o 0 a

72 » o

.2 I I I I I I

O 1 2 3 4 5 6
storage time at 50°C (days)

Figure 4.7. Comparison of the hexanal concentration determined by
GC/MS-SPME and CI-ELISA, and TBARS of chicken myoﬁbrils stored for 5
days at 50°C. GC/MS-SPME quantiﬁed headspace hexanal concentration
and CI-ELISA quantiﬁed hexanal content within the myoﬁbrils. Chicken
myoﬁbrils contained 0.6 umol methyl linoleate/ g proteinand were stored at a“
of 0.75. Determinations were made in triplicate.

99

Table 4.6. Correlation coefﬁcient (r) matrix between methods used

to measure the extent of lipid oxidation in freeze-dried chicken myoﬁbrils
over 5 days of storage at 50°C at 2 water activities (aW 0.30 and 0.75).
Methods used were GC/MS-SPME and CI-ELISA to measure hexanal and
TBARS.

 

Method GC/MS-SPME CI-ELISA

overall aw=0.30 a,=0.75 overall a,,=0.30 aw=0.75
CI-ELISA 0.7758 0.947‘1 0.7303 - - -

 

TBARS 0.838“ 0.937“ 0.8363 0.8653 0.848“ 0.8618

 

 

aSigniﬁcant at p<0.05

100

0.73. These correlations were similar to those reported by Zielinski et al. (2001) in the
quantiﬁcation of hexanal in a chicken thigh homogenate system using a hexanal CI-
ELISA. These researchers reported a correlation (r) of 0.81 between a dynamic
headspace GC method and the monoclonal CI-ELISA, 0.97 between hexanal determined
using a dynamic headspace GC method and TBARS, and 0.77 between TBARS and the
CI-ELISA.

While several other researchers made reference to the TBARS method in the
development of their aldehyde ELISA (Khan et al., 2001; Kondo et al., 200] ). no other
studies were found in which correlations between the two methods were determined. For
other substances, studies have reported correlations between GC and ELISA data. To
verify an ELISA developed for the detection of l3-hydroxyoctadecadienoic acid, Spindler
et al. (1997) compared ELISA and GC results, concluding that the two methods were
comparable. One of the few studies to report correlations between GC-SPME and
immunoassays identiﬁed metolachlor in water samples (Gaynor et al., 1996). A
correlation (r2) of 0.80 was reported between GC-SPME and an ELISA technique.

The CI-ELISA and the GC/MS-SPME displayed a high correlation quantifying
different forms of hexanal. While both methods quantify hexanal, the GC/MS-SPME
method quantiﬁes only volatile hexanal found in the headspace, while the Cl-ELISA
detects chemically and physically bound hexanal (Jennings, 1977). Hexanal binding has
been studied in a number of protein systems, including soy protein in which 37 to 44% of
hexanal was found to be reversibly bound, while less than 5 % was irreversibly bound

(Gremli, 1974). Myosin was reported to bind less than 10% of the hexanal while in actin,

101

an average of 25 to 30% of the added hexanal bound to the protein (Gutheil and Bailey.
1992). This hexanal-protein binding may complicate quantiﬁcation by the GC/MS-
SPME, which measures only volatile hexanal, neglecting the hexanal physically or
chemically bound in the sample.

Preliminary studies (data not shown) showed that the addition of excess C SA to
the hexanal-myoﬁbril system did not result in additional hexanal detection by the CI-
ELISA, indicating enough binding sites on the myoﬁbrils were available for free hexanal.
The number of lysine groups in myosin and actin, both components of chicken
myoﬁbrils, were 209 and 285, respectively (NCBI; www.ncbi.nlm.nih.gov). Due to the
large number of available lysine groups in the myoﬁbrils, an adequate number of binding
sites was available for conjugate formation.

4.5 CONCLUSION

Lipid oxidation is a major determinant in the consumer acceptance of food due to
the off-odor and ﬂavors that are generated during the oxidative process. The current
methods used to quantify lipid oxidation, including those used to quantify hexanal, have
many limitations including low sensitivity and labor intensiveness. The use of a C I-
ELISA for the quantiﬁcation of hexanal as an indicator of lipid oxidation would provide a
fast and easy-to-use alternative to the traditional methods. This study described the use of
a CI-ELISA for the detection of hexanal as an indicator of lipid oxidation in freeze-dried
chicken myoﬁbrils. The CI-ELISA assay was optimized, as was the trypsin solubilization
of the myoﬁbrils. Over long-terrn 50°C storage of chicken myoﬁbrils at two water

activities, the CI-ELISA was found to correlate well with the GC/MS-SPME method, and

102

appeared to be more sensitive than the TBARS method. In addition the advantages
outlined above, the CI—ELISA also offers convenience for use in a kit. which can easily be

utilized within a food processing industry.

103

CHAPTER 5

CONCLUSIONS

We successfully developed a rapid and sensitive GC/MS-SPME method
for the detection of hexanal as an indicator of lipid oxidation. By using a
short sampling time and a low sampling temperature, we were able to
minimize hexanal production dming sampling. We also achieved thh
recovery of hexanal from spiked myoﬁbrils. In an accelerated oxidation
study, the GC/MS-SPME detected differences in hexanal between
myoﬁbrils stored at two water activities at 50°C over 50 hours; this

method also had a strong correlation with TBARS.

We optimized protein extraction and hexanal conjugate formation in
chicken myoﬁbrils. Through the use of trypsin, a lysine speciﬁc enzyme,
we were able to solubilize the protein and expose the lysine for hexanal
conjugation. As trypsin remained active during the CI-ELISA. trypsin

inactivation via heating was required.

We validated the use of the monoclonal CI-ELISA for the detection of
hexanal during lipid oxidation. The CI-ELISA was able to detect

differences in hexanal concentration in chicken myoﬁbrils stored at two

104

water activities over a 5-day storage period at 50°. The CI-ELISA also
correlated well with both GCfMS-SPME and TBARS. The CI-ELISA
proved to be a viable alternative for industry to detect lipid oxidation. and

has potential for use in a commercial kit.

105

CHAPTER 6

FUTURE RESEARCH

The use of the CI-ELISA showed promise for hexanal detection in chicken
myoﬁbrils. Because the hexanal CI-ELISA is a novel method, no sensory studies have
yet been conducted relating CI-ELISA to sensory data. To support the use of the C1-
ELISA as a viable indicator of lipid oxidation, sensory evaluation needs to be included in
future studies (Gray and Monahan, 1992).

To continue pursuing application of the CI-ELISA in the meat industry, the next
stage of the project is to study a commercial meat product. Appropriate extraction and
conjugation procedures need to be developed and optimized for different products,
including whole muscle and comminuted meat, fresh and frozen products. In order to
provide a viable alternative for industry, hexanal detection using the CI-ELISA needs to
be completed in one working day (Fukal, 1991). To meet this criteria, various aspects of
the assay could be examined, including the length of incubation time of protein with
trypsin, and CI-ELISA incubation times. Alternate methods of trypsin inactivation could
also be examined, including the use of a water-soluble trypsin inhibitor such as

In order to maximize its usefulness to the industry, the CI-ELISA would need to
be adopted for commercial ELISA use in a kit. Before the CI-ELI SA can be modiﬁed for
kit development, a number of questions need to be addressed (Pestka, 1988):

1. What are the limits of detection and sensitivity range of the ELISA? Are

106

these values consistent with the needs of the user?

2. Are practical sampling protocols described in the assay kit?

3. Are the protein extraction and protein-hexanal conjugate procedures
maximized?

4. How may samples can be processed in a given period of time?

5. How stable are the assay components, speciﬁcally, the primary and
secondary antibody?

Quality control guidelines would also need to be established. The odor threshold of
hexanal in lean ground beef is reported to range between 2.4 and 36.8 ppm (Brewer and
Vega, 1995). Using these values as a reference, the acceptable limit of hexanal in a given
product would need to be determined by the company. _

With increased numbers of test kits becoming available on the market, the
establishment of guidelines for ELISA kits has become necessary. The Ofﬁcial Methods
Board Task Force on Test Kits and Proprietary Methods has described certain
requirements pertaining to test kits: the need to include appropriate controls to the users.
need to consider false positive/negative results, sensitivity and speciﬁcity in relation to
intended use and the necessity to have a speciﬁc approach (Samarajeewa et al.. 199] ).

The number of commercially available ELISA kits available on the market
continues to increase (Swaminathan and Feng, 1994). Currently. commercial ELISA kits
are available for the detection of malonaldehyde and nonenal, although the use of these
kits is targeted to the medical profession. An ELISA kit for the detection of lipid

oxidation would be beneﬁcial to the food industry.

107

REFERENCES

Ahn, D.U., Jo, C. and Olson, D.G. 2000. Analysis of volatile components and the
sensory characteristics of irradiated raw pork. Meat Sci. 54: 209-215.

Ahn, D.U., Nam, K.C., Du, M. and Jo, C. 2001. Volatile production in irradiated
normal, pale soft exudative (PSE) and dark ﬁrm dry (DF D) pork under different
packaging and storage conditions. Meat Sci. 57: 419-426.

Ahn, D.U., Sell, J .L., Chen, X., Wu, C., and Lee, J .L. 1998. Effect of dietary vitamin E

of lipid oxidation and volatiles content of irradiated cooked turkey meat patties with
different packaging. Poultry. Sci. 77: 912-920.

Ajuyah, A.O., Fenton, T.W., Hardin, RT and Sim, J .S. 1993. Measuring lipid oxrdation
volatiles in meats. J. Food Sci. 58: 270-273.

Alasnier, C ., Meynier, A., Viau, M. and Gandemer, G. 2000. Hydrolytic and oxidative

changes in the lipid of chicken breast and thigh muscles during reﬁigerated storage. J.
Food Sci. 65: 9-14.

Alonzo, G., Saiano, F., Tusa, N and Del Bosco, SF. 2001. Analysis of volatile
compounds released by embryogenic cultures and somatic embryos of sweet organges by
headspace SPME. Plant Cell Tissue Organ Culture. 66: 31-34.

Ang, C.Y.W. and Young, LL. 1989. Rapid headspace gas chromatographic method for
assessment of oxidative stability of cooked chicken meat. J. Assoc. Off. Anal. Chem.
72: 277-281.

Ang, C.Y.W., Liu, F. and Sun, T. 1994. Development of a dynamic headspace GC
method for assessing the inﬂuence of heating end-point temperature on volatiles of
chicken breast meat. J. Agric. Food Chem. 42: 2493-2498.

Arnold. J .W. and Senter, SD. 1998. Use of digital aroma technology and SPME GC-MS
to compare volatile compounds produced by bacteria isolated from processed poultry. J.

Sci. Food Agric. 78: 343-348.

Arthur, CL. and Pawliszyn. J. 1990. Solid phase microextraction with thermal
desorption using fused silica optical ﬁbers. Anal. Chem. 62: 2145-2148.

Asghar. A., Gray, J .I., Buckley, D.J., Pearson, A.M. and Booren, A.M. 1988.
Perspectives on warmed over ﬂavor. Food Technol. 42 (6): 102-108.

108

Aspelund, T.G. and Wilson, LA. 1983. Adsorption of off-ﬂavor compounds onto soy
protein: A thermodrynamic study. J. Agric. Food Chem. 31: 539-545.

Aubourg, SP. 1999. Lipid damage detection during the frozen storage of an
underutilized ﬁsh species. Food Res. Inter. 32: 497-502.

Beaulieu, J .C. and Grimm, CC. 2001. Identiﬁcation of volatile compounds in
cantaloupe at various developmental stages using solid phase microextraction. J. Agric.
Food Chem. 49: 1345-1352.

Beltran, J ., Lopez, F.J., Hernandez, F. 2000. Solid-phase microextraction in pesticide
residue analysis. J. Chromatography A. 885: 389-404.

Bird, R.P., Draper, H.H. and Valli, V.E.O. 1982. Toxicological evalulation of
malonaldehyde: A 12-month study of mice. J. Toxic. Environ. Health. 10: 897-905.

Biswas, S. and Staff, C. 2001. Analysis of headspace compounds of distillers grains
using SPME in conjugation with GC/MS and TGA. J. Cereal Sci. 33: 223-229.

Bradford, M.M. 1976. A rapid and sensitive method for the quantitation of microgram
quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72:
248-254.

Brewer, MS. and Vega, J .D. 1995. Detectable odor thersholds of selected lipid
oxidation compounds in a model meat system. J. Food Sci. 60: 592-595.

Bruce, W.B., Archer, M.C., Corpet, D.E., Medline. Al., Minkin, 8., Stamp, D., Yin, Y.
and Zhang, X. 1993. Diet, aberrant crypt foci and colorectal cancer. Mutation Res. 290:
111-118.

Brunton, N.P., Cronin, D.A., Monahan, F.J. and Durcan, R. 2000. A comparison of
solid-phase microextraction (SPME) ﬁbres for measurement of hexanal and pentanal in
cooked turkey. Food Chem. 68: 339-345.

Chen, W.. Poon, K., and Lam, M.H. 1998. The application of solid phase
microextraction in the analysis of organophosphorus pesticides in a food plant. Environ.
Sci. Technol. 32: 3816-3820.

Chin. H.W., Bernhard, RA. and Rosenberg, M. 1996. Solid phase microextraction for
cheese volatile compound analysis. J. Food Sci. 61(6): 1118-1122, 1128.

Chinachoti, P. 2000. Water activity. In Food Chemistry: Principles and Applications.
G.L. Christen and J .S. Smith, (Ed), p. 21-34. Science Technology System, West
Sacramento, CA.

109

Chung, S., and Champagne, ET. 2001. Association of end-product adducts with
increased IgE binding of roasted peanuts. J. Agric. Food Chem. 49: 3911-3916.

Clark, T.J. and Bunch, J .E. 1997. Qualitative and quantiative analysis of ﬂavor additives

on tobacco products using SPME-GC-Mass spectroscopy. J. Agric. Food Chem. 45:
844-849.

Corpet, D.E., Stamp, D., Medline, A., Minkin, S., Archer, M.C., and Bruce, W.R. 1990.
Promotion of colonic microadenoma growth in mice and rats fed cooked sugar or cooked
casein and fat. Cancer Res. 50: 6955—6958.

Correia, M., Delerue-Matos, C. and Alves, A. 2001. Development of a SPME-GC-ECD
methodology for selected pesticides in must and wine samples. F resenius J. Anal. Chem.
369: 647-651.

Damodaran, S., and Kinsella, J .E. 1980. Flavor protein interactions: Binding of
carbonyls to bovine serum albumin: thermodynamic and conformational effects. J. Agric.
Food Chem. 28: 567-571.

Deshpande, SS. 1996. Assay development, evalulation and validation. In Enzyme
Immunoassay: From Concept to Product Development, p. 275-359. Chapman and Hall,
New York, NY.

Du. M., Ahn, D.U., Nam, K.C. and Sell, J .L. 2001. Volatile proﬁles and lipid oxidation
of irradiated cooked chicken meat from laying hens fed diets containing conjugated
linoleic acid. Poultry Sci. 80: 235-24].

Dugan, LR. 1961. Development and inhibition of oxidative rancidity in foods. Food
Tech. 15:10-18.

Dupuy, H.P., Bailey, M.E., St. Angelo, A.J., Vercellotti, J .R. and Legendre, MG. 1987.
Instrumental analysis of volatiles related to warmed-over ﬂavor of cooked meats. In
Flavor Quality of Fresh Meats: Warmed Over Flavor, A.J. St. Angelo and ME. Bailey.
(Ed). p. 165-190. Academic Press, Orlando, FL.

Elmore, J .S., Erbahadir, M.A., and Mottram, D.S. 1997. Comparison of dynamic
headspace concentration of tenax with solid-phase microextraction for the analysis of
aroma volatiles. J. Agric. Food Chem. 45: 2638-2641.

Elmore, J .S., Mottram, D.S., and Hierro, E. 2000. Two-ﬁbre solid-phase microextraction
combined with gas chromatography-mass spectrometry for the analysis of volatile aroma
compounds in cooked pork. J. Chromatography A. 905: 233-240.

110

Ennis, B.M. and Harper, W.J. 1986. Properties of spray-dried lactalbumin treated with a
protease. New Zealand J. Dairy Sci. Tech. 21: 205-216.

Esterbauer, H. 1993. Cytotoxicity and genotoxicity of lipid oxidation products. Am. J.
Clin. Nutr. 57: 7798-7858.

F ouad, F .M., Shahidi, F., and Mamer, CA. 1995. Comparison of thermally oxidized
lipids and acetaminophen with concurrent consumption of ethanol as inducers of li «er
cirrhosis. J. of Toxic. Environ. Health. 46, 217-232.

Frankel, EN. 199]. Review- Recent advancements in lipid oxidation. J. Sci. Food
Agric. 54: 495-511.

Frankel, EN. 1993. In search for better methods to evaluate natural antioxidants and
oxidative stability in food lipids. Trends Food Sci. Technol. 4: 220-225.

Franzen, KL. and Kinsella, J .E. 1974. Parameters affecting the binding of volatile ﬂavor
compounds in model food systems. 1. Proteins. J. Agric. Food Chem. 22: 675-67 8.

Fritsch, CW. and Gale, J .A. 1977. Hexanal as a measure of rancidity in low fat foods.
J. Amer. Oil Chem. Soc. 54: 225-228.

Fukal, L. 1991. Modern immunoassays in meat-product analysis. Die. Nahrung. 35:
431-448.

Garcia, D.D., Reichenbacher, M., Danzer, K., Hurlbeck, C., Bartzsch, C., and Feller, K.
1998. Use of solid-phase microextraction-capillary-gas chromatography (SPME-CGC)

for the varietal characterization of wines by means of chemometrical methods. Fresenius
J. Anal. Chem. 360: 784-787.

Gaynor, J .D., Cancilla, D.A., Webster, G.R., Sarna, L.P., Graham, K.N., Ng, H., Tan,
C.S., Drury, CF. and Welacky, T. 1996. Comparative solid phase extraction, solid phase

microextraction, and immunoassay analyses of metolachlor in surface runoff and tile
drainage. J. Agric. Food Chem. 44: 2736-2741.

Gillespie, R.J., Humphreys, D.A., Baird, NC. and Robinson, EA. 1989. Chemistry.
Allyn and Bacon, Toronto, ON.

Gorecki, T., Yu, X., and Pawliszyn, J. 1999. Theory of analyte extraction by selected
porous polymer SPME ﬁbres. Analyst. 124: 643-649.

Grau, A., Codony, R., Rafecas, M., Barroeta, A.C., and Guardiola, F. 2000. Lipid
hydroperoxide determination in dark chicken meat through a ferrous oxidation-xylenol
orange method. J. Agric. Food Chem. 48: 4136-4143.

111

Gray. J .l. 1978. Measurement of lipid oxidation: A review. J. Amer. Oil Chem. Soc.
55: 539-546.

Gray, J .I. and Monahan, F .J . 1992. Measurement of lipid oxidation in meat and meat
products. Trends Food Sci. Technol. 3: 315-319.

Gray, J .I., Gomaa, EA. and Buckley, DJ. 1996. Oxidative quality and shelf life of
meats. Meat Sci. 43: 8111-8123.

Gremli. HA. 1974. Interaction of ﬂavor compounds with soy protein. J. Am. Oil Chem.
Soc. 51: 95A-97A.

Grimm, C.C., Bergman, C., Delgado, J., and Bryant, R. 2001. Screening for 2-acetyl-1-
pyrroline in the headspace of rice using SPME/GC-MS. J. Agric. Food Chem. 49: 245-
249.

Gutheil, R.A. and Bailey, ME. 1992. A method for determining binding of hexanal by
myosin and actin using equilibrium headspace sampling gas chromatography. In Food
Science and Human Nutrition: Developments in Food Science 29, G. Charalambous
(Ed.), p. 783-815. Elsevier Science Publishers, Amsterdam, Netherlands.

Gutterridge, J .M.C . and Halliwell, B. 1990. The measurement and mechanism of lipid
peroxidation in biological systems. Trends. Biol. Sci. 15: 129-135.

Habeeb. A.F.S.A. 1966. Determination of free amino groups in proteins by
trinitrobenzenesulfonic acid. Anal. Biochem. 14: 328-336.

Hernandez-Munoz, P., Catala, R., and Gavara, R. 1999. Effects of sorbed oil on food
aroma loss through packaging materials. J. Agric. Food Chem. 47: 4370-4374.

Hernandez-Munoz, P., Catala, R., Hernandez, R]. and Gavara, R. 1998. Food arc .na
mass transport in metallocene ethylene-based copolymers for packaging applications. J.
Agric. Food Chem. 46: 5238-5243.

Hinz, D.C., Kwarteng-Achearnpong, W. 1999. Analysis of volatile varnishes of coated
wires by SPME. Fresenius J. Anal. Chem. 364: 641-642.

Hsieh, Y. P., Sheu, S. and Bridgman, R. 1998. Development of a monoclonal antibody
speciﬁc to cooked mammalian meats. J. Food Prot. 61: 476-481.

Igene. J .0, Pearson, A.M., Merkel, R.A. and Coleman, TH. 1979. Effect of frozen
storage time, cooking and holding temperature upon extractable lipids and TBA values of
beef and chicken. J. Animal Sci. 49 (3): 701-707.

112

Igene, J .O., Yarnauchi, K., Pearson, A.M., Gray, J .I. and Aust, SD. 1985. Evaluation of
2-thiobarbituric acid reactive substances (TBARS) in relation to warrned-over ﬂavor
(WOF) development in cooked chicken. J. Agric. Food Chem. 33: 364-367

Israel, Y., Hurwitz, E., Niemela, O., and Amon, R. 1986. Monoclonal and polyclonal
antibodies against acetaldehyde-containing epitopes in acetaldehyde-protein adducts.
Proc. Natl. Acad. Sci. USA. 83: 7923-7927.

Jakobsen, M. and Bertelsen, G. 2000. Colour stability and lipid oxidation of fresh beef.
Development of a response surface model for predicting the effects of temperature,
stroage time and modiﬁed atmosphere composition. Meat Sci. 54: 49-57.

Jennings, W.G. 1977. Objective measurements of ﬂavor quality: General approaches.
problems, pitfalls and accomplishments. In Flavor Quality: Objective Measurements.
R.A.Scanlan (Ed.). American Chemical Society, Washington, DC.

Jentoft, N. and Dearbom, D.G. 1979. Labeling of proteins by reductive methylation
using sodium cyanoborohydride. J. Biol. Chem. 245(11): 4359-4365.

Kanner, J. 1994. Oxidative processes in meat and meat products: Quality implications.
Meat Sci. 36: 169-189.

Karel, M. 1980. Lipid oxidation, secondary reactions and water activity of foods. In
Autoxidation in Food and Biological Systems. M.S. Simic and M. Karel (Ed.), p. :9]-
206. Plenum Press, New York, NY.

Karel, M., Schaich, K., and Roy, RB. 1975. Interaction of peroxidizing methyl linoleate
with some proteins and amino acids. J. Agric Food Chem. 23: 159-163.

Kaur, H. and Perkins, M.J. 1991. The free radical chemistry of food additives. In Free
Radicals and Food Additives. O.I. Aruoma and B. Halliwell (Ed.). p. 17-35. Taylor and
Francis. Ltd., Bristol, PA.

Kerler, J. and Grosch, W. 1996. Odorants contributing to warrned-over ﬂavor (WOF) of
refrigerated cooked beef. J. Food Sci. 61: 1271-1274.

Keszler, A., Heberger, K. and Gude, M. 1998. Quantitative analysis of aliphatic
aldehydes by headspace SPME sampling and ion-trap GC/MS. Chromatographia. 48:
127-132.

Khan, M.F., Wu, X., and Ansari, AS. 2000. Contribution of nitrosobenzene to splenic
toxicity of aniline. J. Toxic. Environ. Health. 60: 263-273.

113

Khan, M.F., Wu, X., and Ansari, GA. 2001. Anti-malondialdehyde antibodies in MRL
+/+ mice treated with trichloroethene and dichloroacetyl chloride: Possible role of lipid
peroxidation in autoimmunity. Toxic. Appl. Pharm. 170: 88-92.

Kinsella, J.E., and Fox, PF. 1986. Water sorption by proteins: Milk and whey proteins.
Crit. Rev. Food Sci. Nutr. 24: 91-139.

Klassen, L.W., Tuma, D.J., Sorrel], M.F., McDonald, T.L., DeVasure, J .M., and Thiele.
GM. 1994. Detection of reduced acetaldehyde protein adducts using a unique
monoclonal antibody. Alcohol Clin. Exp. Res. 18: 164-171.

Kondo, A., Muranaka, Y., Ohta, I., Notsu, K., Manabe, M., Kotani, K., Saito, K.,
Maekawa, M., and Kanno, T. 200]. Relationship between triglyceride concentrations
and LDL size evaluated by malondialdehyde-modiﬁed LDL. Clin. Chem. 47: 893-907.

Kubow. S. 1993. Lipid oxidation products in food and atherogenesis. Nutr. Rev. 51:
33-40.

Kuby, J. 1997. Immunology (3rd ed). W.H. Freeman and Company, New York, NY.

Kunitz, M., and Northrop, J. 1948. Trypsinogen, trypsin, and trypsin inhibitor. In
Crystalline Enzymes. J. Northrop, M. Kunitz and R. Herriott (Ed.), p. 125-167.
Columbia University Press, New York, NY.

Kvanli, AH. 1988. Statistics: A computer integrated approach. West Publishing
Company. New York, NY.

Labuza, TR 1984. Determination of moisture sorption isotherms. In Moisture sorption:
Practical aspects of isotherm measurement and use, p. 64-66. American Association of
Cereal Chemists, St. Paul, MN.

Ladikos. D. and Lougovois. V. 1990. Lipid oxidation in muscle foods: A review. Food
Chem. 35: 295-314.

Lai. 8., Gray, J.I., Booren, A.M, Crackel, R.L and Gill, J .L. 1995. Assessment of off-
ﬂavor development in restructured chicken nuggets using hexanal and TBARS

measurements and sensory evaluation. J. Sci. Food Agric. 67: 447-452.

Li. S. and King, A.J. 1996. Lipid oxidation and myosin denaturation in dark chicken
meat. J. Agric. Food Chem. 44: 3080-3084.

Lin, T. and Park, J.W. 1998. Solubility of salmon myosin as affected by conformational
changes at various ionic strengths and pH. J. Food Sci. 63(2): 215-218.

114

Love, JD. and Pearson, A.M. 1971. Lipid oxidation in meat and meat products: A
review. J. Am. Oil Chem. Soc. 4: 547-549.

Lynch, MP. and Faustrnan, C. 2000. Effect of aldehyde lipid oxidation products on
myoglobin. J. Agric. Food Chem. 48: 600-604.

Marin, M., Baek, I, and Taylor, A.J. 1999. Volatile release from aqueous solutions under
dynamic headspace dilution conditions. J. Agric. Food Chem. 47: 4750-4755.

Marsili, RT. 1999. SPME-MS-MV A as an electronic nose for the study of off-ﬂavors in
milk. J. Agric. Food Chem. 47: 648-654.

Marsili, RT. 2000. Shelf-life prediction of processed milk by solid-phase
microextraction, mass spectrometry, and multivariate analysis. J. Agric. Food Chem. 48:
3470-3475.

Means, GE. and F eeney, RE. 1968. Reductive alkylation of amino groups in proteins.
Biochem. 7: 2192-2201.

Medina, M. 1988. Extraction and quantiﬁcation of soy protein in sausages by ELISA. J.
Agric. Food Chem. 36: 766-771.

Melton, S.L. 1983. Methodology for following lipid oxidation in muscle foods. Food
Tech. 42(7): 105-1]].

Miller, ME. and Stuart, J .D. 1999. Comparison of gas-sampled and SPME-sampled

static headspace for the determination of volatile ﬂavor components. Anal. Chem. 7]:
23-27.

Monahan, F .J ., Crackel, R.L., Gray, J.I., Buckley, DJ. and Morrissey, RA. 1993.
Catalysis of lipid oxidation in muscle model systems by haem and inorganic iron. Meat
Sci. 34: 95-106.

Morissette, C., Goulet, J ., and Lamoureux, G. 1990. Simple and rapid inhibition enzyme
immunoassay for the detection of staphylococcal enterotoxin B in foods. J. Food Prot.
53: 834-840.

Morr, C.V., German, B., Kinsella, J .E., Regenstein, J.M., Van Buren, J.P., Kilara, A.,
Lewis, BA. and Mangino, ME. 1985. A collaborative study to develop a standar iized
food protein solubility procedure. J. Food Sci. 50: 1715-1718.

Nawar, W.W 1985. Lipids. In Food Chemistry, O.R. Fennema, (Ed), p. 139-244.
Marcel Dekker, New York, NY.

115

Nielsen, H.K., F inot, PA. and Hurrell, RF. 1985. Reactions of proteins with oxidizing
lipids: Inﬂuence on protein quality and on the bioavailability of lysine, methionine.
cysteine and tryptophan as measured in rat assays. Brit. J. Nutr. 53: 75-86.

Nielsen, J .H., Sorensen, B., Skibsted, L.H., and Bertelsen, G. 1997. Oxidation in pre-
cooked minced pork as inﬂuenced by chill storage of raw muscle. Meat Sci. 46: 19] -
197.

O’Keefe, S.F., Resurreccion, A.P., Wilson, LA. and Murphy, P.A. 1991a. Temperature
effect on binding of volatile ﬂavor compounds to soy protein in aqueous model systems.
J. Food Sci. 56: 802-806.

O’Keefe, S.F., Wilson, L.A., Resurreccion, AP. and Murphy, P.A. 1991b.
Determination of the binding of hexanal to soy glycinin and B-conglycinin in an aqueous
model system using a headspace technique. J. Agric. Food Chem. 39: 1022-1028.

Orlien, V., Andersen, A.B., Sinkko, T., and Skibsted, L.H. 2000. Hydroperoxide
formation in rapeseed oil encapsulated in a glassy food model as inﬂuenced by
hydrophilic and lipophilic radicals. Food Chem. 68: 191-199.

Paraf, A. 1992. A role of monoclonal antibodies in the analysis of food proteins. Trends
in Food Sci Tech. 3: 263-267.

Park, PS. and Goins, RE. 1992. Determination of volatile lipid oxidation products by
dynamic headspace-capillary gas chromatographic analysis with application to milk-based
nutritional products. J. Agric. Food Chem. 40: 1581-1585.

Peres, C., Viallon, C., and Berdague, J. 2001. Solid-phase microextraction -mass
spectrometry: A new approach to the rapid characterization of cheese. Anal. Chem. 73:
1030-1036.

Pestka, J .J . 1988. Enhanced surveillance of foodbome mycotoxins by immunochcmical
assay. J. Assoc. Off. Anal. Chem. 71: 1075-1081.

Raharjo, S. and Sofos, J .N. 1993. Methodology for measuring malonaldehyde as a
product of lipid peroxidation in muscle tissues: A review. Meat Sci. 35: 145-169.

Ramarathnarn, N., Rubin, L.J., and Diosady, LL. 1991. Studies on meat ﬂavor. 2. A
quantitative investigation of the volatile carbonyls and hydrocarbons in uncured and
cured beef and chicken. J. Agric. Food Chem. 39: 1839-1847.

Reineccius, GA. 1996. Instrumental means of monitoring the ﬂavor quality of foods.
Cp. 22 in Chemical Markers for Processed and Stored Foods. , T.C. Lee and HI. Kim
(ed), p. 241-25 1. American Chemical Society, Washington, DC.

116

Rhee, K.S., Anderson, L.M. and Sams, AR. 1996. Lipid oxidation potential of beef.
chicken and pork. J. Food Sci. 61: 8-12.

Ruiz, J ., Cava, R., Ventanas, J. and Jensen, M.T. 1998. Headspace solid phase
microextraction for the analysis of volatiles in a meat product: Dry-cured Iberian ham. J.
Agric. Food Chem. 46: 4688-4694.

St. Angelo, A.J., Vercellotti, J.R., Legendre, M.G., Cinnett, C.H., Kuan, J .W., James. C.
and Dupuy, HP. 1987. Chemical and instrumental analysis of warmed-over ﬂavc; in
beef. J. Food Sci. 52: 1163-1168.

Saleeb, F.Z. and Schenz, T.W. 1978. A technique for the determination of volatile
organic compounds under equilibrium and non-equilibrium conditions. In Analysis of
Foods and Beverages, G. Charalambous (Ed), p. 37-56. Academic Press, New York, NY.

Salomon, R.G., Laur, K., Podrez, E., Hoff, H.F., Krushinsky, A.V., and Sayre, L.M.
2000. HNE-derived 2-penty1pyrroles are generated during oxidation of LDL, are most
prevalent in blood plasma from patients with renal disease or atherosclerosis, and are
present in atherosclerotic plaques. Chem. Res. Toxicol. 13: 557-564.

Samarajeewa, U., Wei, C.J., Huang, TS. and Marshall, MR. 1991. Application of
immunoassay in the food industry. Crit. Rev. Food Sci. Nutr. 29: 403-434.

SchirIe-Keller, J .P., Reineccius, G.A., and Hatchwell, LC. 1994. Flavor interacticns
with fat replacers: Effect of oil level. J. Food Sci. 59: 813-815.

Sen, N.P., Seaman, SW. and Page, B.D. 1997. Rapid semi-quantitative estimation of N-
nitrosodibutylamine and N-nitrosodibenzylamine in smoked hams by solid-phase

microextraction followed by gas chromatography-thermal energy analysis. J.
Chromatography A. 788: 131-140.

Shahidi, F. and Pegg, RB. 1994. Hexanal as an indicator of the ﬂavor deterioration of
meat and meat products. In Lipid Oxidation in Food, A.J.St. Angelo (ed.), p. 105-119.
ACS Symposium Series 500. American Chemical Society, Washington, DC.

Shahidi, F., Yun, J. and Rubin, L.J. 1987. The hexanal content as an indicator of
oxidative stability and ﬂavor acceptability in cooked ground pork. Can. Inst. Food Sci.
Technol. J. 20(2): 104-106.

Sides, A., Robards, K., Helliwell, S., and An, M. 2001. Changes in the volatile proﬁle of
oats induced by processing. J. Agric. Food Chem. 49: 2125-2130.

Siegmund, B., Derler, K., and Pfannhauser, W. 2001. Changes in the aroma of a
strawberry drink during storage. J. Agric. Food Chem. 49: 3244-3252.

117

Simpson, B.K., Nayeri, G., Yaylayan, V. and Ashie, I.N.A. 1998. Enzymatic hydrolysis
ofshrimp meat. Food Chem. 61: 131-138.

Sinnhuber, RD. and Yu, TC. 1958. 2-Thiobarbituric acid method for the measurement
of rancidity in ﬁshery products. Food Tech. 12: 9-11.

Sista, R.V., Erickson, M.C., and Shewfelt, R.L. 2000. Lipid oxidation in a chicken
muscle model system: Oxidative response of lipid classes to iron ascorbate or
methemoglobin catalysis. J. Agric. Food Chem. 48: 1421-1425.

Smith, D.M. 1987. Functional and biochemical changes in deboned turkey due to frozen
storage and lipid oxidation. J. Food Sci. 52: 22-27.

Smith, D.M., Noormarji, S.H., Price, J.F., Bennink, MR. and Herald, T.J. 1990. Effect
of lipid oxidation on the ftmctional and nutritional properties of washed chicken
myoﬁbrils stored at different water activities. J. Agric. Food Chem. 38: 807-131..

Smith, SA. 1997. Production and characterization of polyclonal antibodies to hexanal-
lysine adducts for use in an ELISA to monitor lipid oxidation in meat model system.
Ph.D. Dissertation. Michigan State University, East Lansing, MI.

Smith, S.A., Pestka, J.J., Gray, J.I. and Smith, D.M. 1999. Production and speciﬁcity of
polyclonal antibodies to hexanal-lysine adducts. J. Agric. Food Chem. 47: 1389-1395.

Snyder, J .M., King, J .W., and Zhang, Z. 1998. Comparison of volatile analysis of lipid-
containing and meat matrices by solid phase micro- and supercritical ﬂuid-extraction. In.
Flavor Analysis: Developments in Isolation and Characterization. C.J. Mussinan and M.J.
Morello (Ed), p. 107-115. ACS Symposium Series 705. American Chemical Society,
Washington, DC.

Song, J ., Fan, L., and Beaudry, RM. 1998. Application of solid phase microextra.tion
and gas chromatography/time-of-ﬂight mass spectrometry for rapid analysis of ﬂavor
volatiles in tomato and strawberry fruits. J. Agric. Food Chem. 46: 3721-3726.

Song, J ., Gardner, B.D., Holland, J .F . and Beaudry, RM. 1997. Rapid analysis of
volatile ﬂavor compounds in apple fruit using SPME and GCfI‘ime-of-ﬂight mass
spectrometry. J. Agric. Food Chem. 45: 1801-1807.

Sorensen, E.S., Hojrup, P., and Petersen, TE. 1995. Posttranslational modiﬁcations of

bovine osteopontin: Identiﬁcation of twenty-eight phophorylation and three 0-
glycosylation sites. Protein Sci. 4: 2040-2049.

118

Sostaric , T., Boyce, M.C., and Spickett, BE. 2000. Analysis of the volatile compounds
in vanilla extracts and ﬂavorings by solid-phase microextraction and gas chromatography.
J. Agric. Food Chem. 48: 5802-5807.

Spanier, A.M., Miller, J .A. and Bland, J .M. 1992. Lipid oxidation: Effect on meat
proteins. In Lipid Oxidation in Food, A.J. St. Angelo (Ed.), p. 105-110. ACS
Symposium Series 500. American Chemical Society, Washington, DC.

Spindler, S.A., Clark, K.S., Blackburn, M.L., Bull, A.W., and Reddy, R.G. 1997.
Occurrence of I3(S)-hydroxyoctadecadienoic acid in biological samples. Prostaglandins.
54: 875-880.

Stanley, D.W., Stone, A.P., Hultin, HQ. 1994. Solubility of beef and chicken
myoﬁbrillar proteins in low ionic strength media. J. Agric. Food Chem. 42: 863-867.

Steffen. A. and Pawliszyn, J. 1996. Analysis of ﬂavor volatiles using headspace solid-
phase microextraction. J. Agric. Food Chem. 44: 2187-2193.

Sun, Q., Faustman, C., Seneca], A., Wilkinson, AL. and Furr, H. 200]. Aldehyde
reactivity with 2-thiobarbituric acid and TBARS in freeze-dried beef during accelerated
storage. Meat Sci. 57: 55-60.

Su, Y., Ang, C.Y.W. and Lillard, DA. 199]. Precooking method affects warmed-over
ﬂavor of broiler breast patties. J. Food Sci. 56(4): 881-883, 898.

Swaminathan, B. and Feng, P. 1994. Rapid detection of food-home pathogenic bacteria.
Annu. Rev. Microbiol. 48: 401-426.

Swets, B.P., Brouwer, D.A., and Tervaert, J .W.C. 200]. Patients with systemic
vasculitis have increased levels of autoantibodies against oxidized LDL. Clin. Exp.
Immun. 124: 163-167.

Tamura, H.. Kitta, K., and Shibamoto, T. 1991. Formation of the reactive aldehydes
from fatty acids in a Fe‘Z/HZO2 oxidation system. Am. Chem. Soc. 39: 439-442.

Tang. S.Z., Kerry, J .P., Sheehan, D., Buckley, DJ. and Morrissey, P.A. 2001a.
Antioxidative effect of dietary tea catechins on lipid oxidation of long-term frozen stored
chicken meat. Meat Sci. 57: 331-336.

Tang, 8., Kerry, J.P., Sheehan, D., Buckley, D.J., and Morrissey, P.A. 2001b.

Antioxidative effect of added tea catechins on susceptibility of cooked red meat, poultry
and ﬁsh patties to lipid oxidation. Food Res. Inter. 34: 651-657.

119

Tarladgis, B.G., Watts, B.M., Younathan, M.T. and Dugan, LR. 1960. A distillation
method for the quantitative determination of malonaldehyde in rancid foods. J. Am. Oil
Chem. Soc. 37: 44-48.

Taylor, A.J. 1998. Physical chemistry of ﬂavor. Inter. J. Food Sci. Tech. 33: 53-62.

Thiele, G.M., Wegter, K.M., Sorrell, M.F., Tuma, D.J., McDonald, TL, and Klassen,
L.W. 1994. Speciﬁcity to N-ethyl lysine of a monoclonal antibody to acetaldehyde-
modiﬁed proteins prepared under reducing conditions. Biochem. Pharm. 48: 183-189.

Tims, M.J. and Watts, BM. 1958. Protection of cooked meats with phosphates. Food
Technol. 12: 240-243.

Uchida, K., Itakura, K., Kawakishi, S., Hiai, H., Toyokuni, S. and Stadtrnan, ER. 1995.
Characterization of epitopes recognized by 4-hydroxy-2-nonenal speciﬁc antibodies.
Arch. Biochem. Biophysics. 324: 241-248.

van Aardt, M., Duncan, S.E., Boume, D., Marcy, J.E., Long, T.E., Hackney, CR. and
Heisey, C. 200]. Flavor threshold for acetaldehyde in milk, chocolate milk, and spring

water using solid phase microextraction gas chromatography for quantiﬁcation. J. Agric.
Food Chem. 49: 1377-1381.

van Ruth, S.M., Roozen, JP, and Jansen, F .J .H.M. 2000. Aroma proﬁles of vegetable
oils varying in fatty acid composition vs. concentrations of primary and secondary lipid
oxidation products. Nahrung. 44: 318-322.

Vas. G., Koteleky, K., Farkas, M., Dobo, A., and Vekey, K. 1998. Fast screening
method for wine headspace compounds using solid-phase microextraction (SPME) and
capillary GC technique. Am. J. Enol. Vitic. 49(1): 100-103.

Vitali, M., Guidotti, M., Giovinazzo, R., and Cedrone, O. 1998. Determination of
pesticide residues in wine by SPME and GCfMS for consumer risk assessment. Fc od
Additives Contam. 15(3): 280-287.

Volante. M., Pontello, M., Valoti, L., Cattaneo, M., Bianchi. M., and Colzani, L. 2000.
Application of solid phase micro-extraction (SPME) to the analysis of pesticide residues
in vegetables. Pest. Mgt. Sci. 56: 618-636.

Wang, C.H., Abouzied, M.M., Pestka, J .J . and Smith, D.M. 1995. Lactate

dehydrogenase polyclonal antibody sandwich ELISA for determination of endpoint
heating temperatures of ground beef. J. Food Prot. 59: 51-55..

120

Wang, C.H., Booren, A.M., Abouzied, M.M., Pestka, U. and Smith, D.M. 1993. ELISA
determination of turkey roll endpoint temperature: Effects of formulation, storage and
processing. J. Food Sci. 58: 1258-1261.

Wen, J ., Morrissey, P.A., Walton, J ., and Sheehy, P.J.A. 1997. Rapid and quantiative
determination of hexanal in cooked muscle foods. Irish J. Agric. Food Res. 36: 75-84.

Whang, K. and Peng, 1.1. 1987. Lipid oxidation in ground turkey skin and muscle during
storage. Poultry Sci. 66: 458-466.

Wijesundera, C., Drury, L. and Walsh, T. 1998. Determination of free fatty acids and
lactones in cheese by solid phase microextraction. Aust. J. Dairy Tech. 53: 140.

Wilkinson, A.L., Sun, Q., Seneca], A., and Faustman,C. 2001. Antioxidant effects on
TBARS and ﬂuorescence measurements in freeze-dried meats. J. Food Sci. 66: 20-24.

Wu, Y., Hamouz, F ., and Schnepf, M. 1998. Factors affecting static headspace-gas

chromatographic analysis of lipid oxidation in precooked meat. J. Agric. Food Chem.
46: 3677-3682.

Xiong, Y.L. and Brekke, C.J. 1989. Changes in protein solubility and gelation properties
of chicken myoﬁbrils during storage. J. Food Sci. 54: 1141-1146.

Xiong, Y.L., Lou, X., Wnag, C., Moody, W.G., and Harmon, R.J. 2000. Protein
extraction from chicken myoﬁbrils irrigated with various polyphosphate and NaCl
solutions. J. Food Sci. 65(1): 96-100.

Yang, C.M., Kendall, C.W., Stamp, D., Medline, A., Archer, MC. and Bruce, W.R.
1998. Thermally oxidized dietary fat and colon carcinogenesis in rodents. Nutrition
Cancer. 30: 69-73.

Yang, M.J., Orton, M.L., and Pawliszyn. 1997. Quantiative determination of caffeine in
beverages using a combined SPME-GC/MS method. J. Chem. Ed. 74(9): 1130-1132.

Yang, X. and Peppard, T. 1994. Solid-phase microextraction for ﬂavor analysis. J.
Agric. Food Chem. 42: 1925-1930.

Yau, TM. 1979. Mutagenicity and cytotoxicity of malonaldehyde in mammalian cells.
Mech. Ageing Dev. 1]: 137-144.

Yoritaka, A., Hattori, N., Uchida, K., Tanaka, M., Stadtrnan, ER. and Mizuno, Y. 1996.

Immunohistochernical detection of 4-hydroxynonena1 protein adducts in Parkinson
disease. Proc. Natl. Acad. Sci. USA. 93: 2696-2701.

121

Zielinski, TL. 1998. Development and characterization of a monoclonal antibody. based
ELISA for monitoring lipid oxidation. Msc. Thesis. Michigan State University, East
Lansing, MI.

Zielinski, T.L., Smith, S.A., Pestka, J.J., Gray, J .I. and Smith, D. 2001. ELISA to

quantify hexanal-protein adducts in a meat model system. J. Agric. Food Chem. 49:
3017-3023.

Zhang, Z. and Pawliszyn, J. 1993. Headspace solid-phase microextraction. Anal. Chem.
65: 1843-1846.

122