THESIS
2.

Illlllllllll lllllllll‘lllllllillllllll

3 1293 01688 7501

This is to certify that the

thesis entitled

Development and Characterization of a
Monoclonal Antibody Based ELISA for Monitoring
Lipid Oxidation

presented by

Tamara L. Zielinski

has been accepted towards fulfillment
of the requirements for

M. S 0 degree in FOOd Science

 

 

jam ”7. W

Major professor

Date W8—

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

 

 

 

 

LIBRARY

Mlchlgan State
1 Unlveralty

r

 

 

PLACE IN RETURN BOX

 

 

 

 

 

 

 

 

 

 

 

to remove this checkout from your record.
TO AVOID FINE-3 return on or before date due.
DATE DUE MTE DUE DATE DUE
1‘: t-t‘ .1. \ I
\ \ _ _
l
\ _ _ \
l fl i *
\ \ _ _
l

 

 

 

 

 

11” MM“

DEVELOPMENT AND CHARACTERIZATION OF A MONOCLONAL
ANTIBODY BASED ELISA FOR MONITORING LIPID OXIDATION

By

Tamara L. Zielinski

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Food Science and Human Nutrition

1998

ABSTRACT

DEVELOPMENT AND CHARACTERIZATION OF A MONOCLONAL BASED
ELISA FOR MONITORING LIPID OXIDATION

By

Tamara L. Zielinski

There is interest in developing rapid methods to monitor lipid oxidation as
chromatographic methods are tedious and require sophisticated equipment. Objectives of
this study were to produce and characterize monoclonal antibodies to hexanal, a common
index of lipid oxidation in food. then devise and verify an indirect competitive enzyme
linked immunosorbent assay (IC-ELISA) to quantify hexanal. Monoclonal antibodies
against hexanal modified protein were produced and an IC-ELISA with a working range
of 1-50 ng hexanal/mL was devised. Antibodies cross-reacted 37.9%. 76.6% and 45.0%,
respectively. with pentanal. heptanal and 2-t-hexenal protein conjugates. Antibodies
reacted strongly with hexanal-aminocaproic acid conjugates. but did not recognize free
hexanal, native proteins, free amino acids, other hexanal-amino acid conjugates or other
aldehyde, alcohol or ketone protein conjugates. Hexanal concentration of conjugates
measured by GC and lC-ELISA were highly correlated (r = 0.97). indicating potential use

of IC-ELISA as a rapid method to monitor lipid oxidation.

To my husband Matt.
Thank you for all your love and support.

ACKNOWLEDGMENTS

For their support and guidance, I would like to thank my committee members, Drs. Ian

Gray, James Pestka and my advisor, Denise Smith.

I would like to recognize the Michigan State University Crop and Food Bioprocessing

Center and the Michigan State University Agricultural Experiment Station for providing

financial support.

For their technical expertise, I would like to extend a special thank you to Drs. James

Clarke, Enayat Gomaa and Virginia Vega-Wamer.

TABLE OF CONTENTS

List of Tables ......................................................................................................... ix
List of Figures ........................................................................................................ x
Chapter 1. Introduction .................................................................................... 1
Chapter 2. Literature Review ........................................................................... 5
2.1 Mechanism of Lipid Oxidation ........................................................ 5
2.1.1 Initiation of Lipid Oxidation ................................................ 6
2.2 Current Methods For Monitoring Lipid Oxidation .......................... 7
2.2.1 TBA Test ............................................................................. 8
2.2.2 Hexanal ............................................................................... 9

2.3 Oxidation of Chicken ...................................................................... 1 1

2.3.1 Fatty Acid Composition of Chicken .................................... 11

2.3.2 TBA Values for Oxidized Chicken ...................................... 11

2.3.3 Hexanal Concentration in Oxidized Chicken ....................... 14

2.4 ELISA Development ....................................................................... 16

2.5

2.4.1 Conjugate Production and Determination of Modification. 16
2.4.2 Monoclonal Antibody Production ....................................... 18
2.4.3 Classification of ELISA Types ............................................ 20

Examples of ELISAs Currently in Use ............................................ 23

2.6

Chapter 3.

3.1

3.2

3.4

Chapter 4

2.5.2 Acetaldehyde ELISAs ........................................................ 24
2.5.3 Polyclonal Antibody for Hexanal ....................................... 25

Conclusion ...................................................................................... 26

Production of Monoclonal Antibodies Against Hexanal-Protein

Conjugates and Characterization of the Final Clone ...................... 27
Abstract .......................................................................................... 27
Introduction .................................................................................... 28
Methods and Materials .................................................................... 31
3.3.1 Materials ............................................................................. 31
3.3.2 Conjugate Preparation ......................................................... 32
3.3.3 Production of Monoclonal Antibodies ................................. 33

3.3.4 Titer Determination and Screening by Indirect ELISA ....... 35

3.3.5 Protein Specificity by Indirect Competitive ELISA ............ 36
3.3.6 Precision .............................................................................. 38
3.3.7 Antibody Cross-Reactivity by IC-ELISA ........................... 38
3.3.8 Statistics .............................................................................. 40
Results and Discussion .................................................................... 40
3.4.1 Antigen Preparation ............................................................. 40
3.4.2 Hybridoma Production ........................................................ 42
3.4.3 Protein Specificity ............................................................... 48
3.4.4 ELISA Optimization ............................................................ 52
3.4.5 Cross-Reactivity Study ........................................................ 53

Verification of a Monoclonal Antibody Based ELISA for Monitoring
Lipid Oxidation by Correlation to Gas Chromotography ................ 64

vi

4. 1 Abstract .......................................................................................... 64

4.2 Introduction .................................................................................... 65
4.3 Methods and Materials .................................................................... 68
4.3.1 Materials ............................................................................. 68
4.3.2 Hexanal Modification of C SA ............................................. 68
4.3.3 Hexanal Concentration by CI-ELISA .................................. 69

4.3.4 Isolation and Concentration of Volatiles for Gas
Chromotography .................................................................. 70
4.3.5 Hexanal Concentration by Gas Chromatography ................ 70
4.3.6 Statistical Analysis ............................................................... 71

4.3.7 Western Blot of Hexanal Modified Salt Soluble Protein ..... 71

4.3.8 Proximate Analysis of Chicken ............................................ 72
4.3.9 Frozen Storage Study (Preliminary Experiments) ................ 73
4.3.10 Optimization of Extraction Procedure for Cooked
Thigh Meat .......................................................................... 73
4.3.1 1 Cooked Storage Study (Preliminary Experiments) .............. 75
4.3.12 TBA-RS Assay .................................................................... 76
4.4 Results and Discussion .................................................................... 76
4.4.1 Differential Modification of CSA ......................................... 76

4.4.2 Electrophoresis and Western Blot ofSalt Soluble Proteins. 79

4.4.3 Frozen Storage Study .......................................................... 82
4.4.4 Enzyme Conditions for Cooked Meat Extraction ................. 85
4.4.5 Cooked Storage Study ......................................................... 88

Chapter 5 ............................................................................................................... 92

Chapter 6 ............................................................................................................... 93

Appendix A ........................................................................................................... 94
Appendix B .......................................................................................................... 95
References .............................................................................................................. 96

viii

List Of Tables

Table
2.1 Unsaturated fatty acid content of chicken fat ........................................ 12

3.1 Amino acid content of native and hexanal modified (hex-) chicken serum
albumin (CSA), bovine serum albumin (BSA) and keyhole limpet
hemocyanin (KLH) ....................................................................... 41

3.2 Sera titers against hexanal modified bovine serum albumin (hex-BSA) or
keyhole limpet hemocyanin (hex-KLH) determined 5 weeks after the
initial immunization ..................................................................... 44

3.3 Reproducibility of the indirect competitive ELISA was determined for
both within a run (intra-assay) and between runs (inter—assay) .................... 55

3.4 Loss of reactive amino groups in CSA solutions was measured by
trinitribenezesulfonic acid (TNBS) assay ............................................ 56

3.5 Cross-reactivity of the monoclonal antibody to aldehyde modified chicken
serum albumin (CSA) by indirect competitive ELISA ........................... 59

4.1 Effect of enzyme concentration on the extraction of salt soluble proteins
from cooked chicken thigh at 50°C for 2 h.. .................................................... 86

4.2 Effect of extraction time on the salt soluble protein concentration of cooked
chicken thigh at 0.5% enzyme concentration and 50°C ..................................... 87

Figure

2.1

2.2

3.1

3.2.

3.3

3.4

3.5

3.6

3.7

3.8

List of Figures

Condensation reaction of hexanal and a primary amine that results in a Schiff
base ....................................................................................... 1?

Diagram of an indirect competitive ELISA ......................................... 21

Production of antibodies against hexanal modified bovine serum albumin
(BSA) was determined by indirect ELISA 43

Production of antibodies against hexanal modified keyole limpet
hemocyanin (KLH) was determined by indirect ELISA .......................... 45

Antibody production by hexanal modified bovine serum albumin
(hex-BSA) injected mice was determined by indirect competitive ELISA, using
serum from the first bleeding ........................................................... 47

Specificity of the monoclonal antibody to hexanal modified proteins was
determined by indirect competitive ELISA ........................................ 50

Effect of hexanal concentration on the binding inhibition of hexanal modified
proteins was determined by indirect competitive ELISA. . . . . . . . 51

Representative standard curve showing the working range of the indirect
competitive ELISA for hexanal ...................................................... 54

Specificity of the monoclonal antibody to aliphatic aldehydes conjugated
to chicken serum albumin (CSA) ................................................... 58

Specificity of the monoclonal antibody to hexanal modified amino acids was
determined by indirect competitive ELISA ........................................ 62

4.1

4.2

4.3

4.4

4.5

Concentration of hexanal added to prepare chicken serum albumin

conjugates and the percent inhibition produced in the ELISA by three
dilutions of the conjugates ..............................................................................

Hexanal concentration of differentially modified protein conjugates as
measured by ELISA and gas chromatography (GC)............... . ..

Detection of hexanal modified and native chicken breast salt soluble
proteins by western blotting following sodium dodecyl sulfate
polyacrylamide gel electrophoresis (SDS-PAGE). . .. ....................................

The hexanal concentration of raw chicken thigh stored at -20°C was

determined by indiret competitive ELISA and gas chromatography
(GC) ...............................................................................................................

Hexanal concentration and 2-thiobarituric acid reactive substances (TBARS)
number of cooked chicken thigh meat stored at 4° C ......................................

78

80

81

83

90

Chapter 1

INTRODUCTION

Coronary heart disease, strokes and cancer have all been found to be related to
dietary fat intake. Atherosclerosis, a major cause of coronary heart disease and a
contributor to strokes, has been linked to high fat intake, especially of saturated fats,
(Pearson et al., 1983). It has been suggested that polyunsaturated fatty acids, which are
more susceptible to autoxidation, may increase the risk of cancer. Studies have shown
that restricting calories, polyunsaturated fatty acids and protein can decrease tumor
incidence (Pearson et al., 1983). Lipid oxidation has recently received interest as to the
possible role it plays in human disease and toxicology (Gutteridge and Halliwell, 1990).
Increased levels of malonaldehyde are often associated with certain diseases such as
atherosclerosis. diabetes and myocardial infarction (Esterbauer, 1993). Malonaldehyde
may also play a role in carcinogenisis by acting as an initiator or promoter (Esterbauer,
1993). Oxidized cholesterol is believed to be a strong atherogenic agent, causing
atherogenesis (Kubow, 1993).

Lipid oxidation has also long been recognized as a problem in the food industry.
Oxidative rancidity as a result of lipid oxidation is a problem during both the preparation
and the storage of food. Both meat and poultry are products susceptible to lipid oxidation.
Flavor, color and nutritional changes in raw and cooked meat can occur under any kind of

storage conditions (Love and Pearson, 1971).

Lipid oxidation begins with a hydrogen atom being abstracted from the methylene
carbon of a fatty acid side chain. Removing the hydrogen atom leaves behind an unpaired
electron on the carbon atom, which generally reacts with oxygen to give a peroxyl
radical. This radical can combine with other radicals, resulting in termination, or can
continue the chain reaction by attacking membrane proteins or removing hydrogens from
other fatty acids (Gutteridge and Halliwell, 1990). The primary products of this reaction
are hydroperoxides. Their breakdown leads to a mixture of low molecular weight
compounds such as alkanes, alkenes, aldehydes, ketones and alcohols (Gray and
Monahan, 1992).

This mixture of low molecular weight compounds causes the off flavors
associated with rancid or oxidized meat (Gray and Monahan, 1992). One commonly used
term for describing oxidized flavors in cooked meats is warmed-over-flavor (WOF).
Warmed-over-flavor is described as cardboardy, rancid, stale and metallic and is a
problem in both meat and poultry (St. Angelo et al., 1987). Color deterioration also
occurs as a result of oxidative reactions. Consumers view the brown color arising from
oxidation of ferric hemes as undesirable (Love and Pearson, 1971 ). Health implications of
lipid oxidation have also started receiving attention. Protein and membrane damage may
be caused by lipid hydroperoxides and their decomposition products, some of which are
believed to be chemical toxicants (Ladikos and Lougovois, 1990). Free radicals that are
formed during lipid peroxidation may also have an adverse effect by attacking membrane
proteins, leading to impairment of membrane functions (Gutteridge and Halliwell, 1990).

The most commonly used test for measuring lipid oxidation is the 2-thiobarbituric

acid (TBA) test. This method was originally believed to only measure malonaldehyde

found in oxidized food (Botsoglou et al., 1994). Since a number of other compounds,
such as sucrose, woodsmoke and acetaldehyde, have also been found to react with the
acid, the test is now referred to as the thiobarbituric acid reactive substances test, known
as TBARS. Because TBARS lacks sensitivity and specificity to MA, it has been widely
criticized. Gray and Monahan (1992) suggest using the TBARS test, but to use a second,
complementary method (such as hexanal concentration) and to correlate both tests back to
sensory scores.

Hexanal is one of the major secondary products formed during lipid oxidation of
linoleic acid (Gray and Monahan, 1992). A number of researchers have been studying
hexanal as a possible indicator of lipid oxidation. The concentration of hexanal has been
found to correlate significantly with sensory scores and TBARS for cooked pork (Shahidi
et al., 1987), cooked beef (St. Angelo et al., 1987) and in restructured chicken nuggets
(Lai et al., 1995). All of these studies used gas chromatography (GC) to measure hexanal
concentration. The drawbacks to this method were outlined by Ajuyah et a1. (1993).
These include the necessity of dedicated equipment. inconvenient sample analysis (only
one sample at a time can be analyzed) and the possibility of peroxide decomposition from
the heat used to drive the volatiles onto the GC column.

Enzyme-linked immunosorbent assays (ELISAs) have been developed to test for a
variety of substances in food such as bacteria, mycotoxins, pesticides, anabolic agents
and adulterants (Samarajeewa et al., 1991). The advantages of ELISAs over GC include
reduction in assay time, sample size, cleaning steps and equipment expense.
(Samarajeewa et al., 1991). An ELISA for hexanal would provide a reproducible, fast and

simple test for monitoring lipid oxidation in meat products.

The overall objective of this project was to develop a monoclonal antibody based
ELISA to quantify hexanal in meat to monitor lipid oxidation. The specific objectives of
this project were as follows:

a. To produce a monoclonal antibody to hexanal.

b. To develop an indirect competitive ELISA, using that monoclonal antibody,

and to test the specificity and sensitivity of the indirect ELISA.

c. To develop a reproducible extraction procedure to extract hexanal from raw

and cooked meat and quantify, by ELISA, the amount of hexanal extracted.

d. To correlate the concentration of hexanal, as measured by ELISA, to current

accepted protocols, such as GC, in storage studies using meat.

Chapter 2

LITERATURE REVIEW

2.1. Mechanism of Lipid Oxidation

The proposed mechanism of lipid oxidation consists of three phases: initiation,
propagation and termination. Initiation involves the removal of a hydrogen atom from the
methylene carbon in the side chain of a fatty acid. This results in a radical which
preferentially (in an aerobic system) reacts with oxygen to give a peroxyl radical. These
radicals can react with each other (termination) or remove hydrogens from other fatty

acid side chains, thereby propagating the reaction (Gutteridge and Halliwell, 1990). The

reaction was summarized by Asghar et al. (1988) as such :

Initiation: LH+O3 ——> 3L0 + OOOH
Propagation: L- + 02 —> LOO.

LH + LOO. -—> LOOH + L0

LOOH —> L0. + oOI-I
Termination: L0 + L0 —> L-L

Lo + LOO. —> LOOL

Loo. + Loo. —> LOOL + 02

with LH representing the unsaturated fatty acid, LOO- representing the peroxy radical,
LOO representing the alkoxy radical and LOOH representing the hydroperoxide.
Unsaturated fatty acids are more susceptible to oxidation than saturated fatty acids
due to the allylic CH bonds (Kaur and Perkins, 1991). The bond dissociation energy of
the allylic CH structure is 20% less than the bond dissociation energy of a similiar
compound that lacks allylic bonds. As a result, the peroxyl radical can readily abstract the
hydrogen from the central methylene carbon in a conjugated diene unit (-CH=CH-CH2-

CH=CH), but can not easily remove a hydrogen from a saturated hydrocarbon (Kaur and

Perkins, 1991).

2.1.1. Initiation of Lipid Oxidation

The initiation of lipid oxidation begins with the formation of radicals. This can
occur through homolysis of weak bonds, such as the 0-0 bond in a peroxide (Kaur and
Perkins, 1991). The classic example is the reduction of hydrogen peroxide by Fe(II),
known as the Fenton reaction:

Fe2+ + H203 —> Fe3+ +HOO +HO'

The resulting hydroxyl radical is extremely reactive and is capable of removing a
hydrogen atom from an unsaturated lipid. This reaction is an example of an induced
peroxide decomposition.

The source of these metal catalysts is controversial. Heme proteins were found to
have a greater effect on lipid oxidation (measured by TBARS) in a pork model system

than inorganic iron (Monahan et al., 1993). Ahn et a1. (1993), however, found that free

ionic iron had more of an effect on the generation of TBARS in raw turkey meat than
ferritin and transferrin, but that heme protein had no effect.

Another way for homolysis of weak bonds to occur is by reacting with singlet
oxygen that is produced from absorption of ultraviolet light (Frankel, 1991). Ultraviolet
light can also transform carbonyl groups into reactive species that can then behave as

reactive free radicals (Kaur and Perkins, 1991).

2.2. Current Methods For Monitoring Lipid Oxidation

Hydroperoxides are the primary products of lipid oxidation. They are colorless,
tasteless and odorless. It is their breakdown into a mixture of low molecular weight
compounds that results in the rancid and off-flavor characteristics that are associated with
oxidized meat. These low molecular weight compounds include alkanes. alkenes,
ketones, alcohols, esters and acids (Gray and Monahan, 1992). Sensory analysis is often
used to evaluate the extent of lipid oxidation in meat. Warrned-over-flavor, or WOF, is a
term used to describe the off flavors resulting from lipid oxidation in cooked meat. This
term was first introduced by Tims and Watts (1958). Since the consumer uses
organoleptic evaluation when judging the quality of foods, sensory analysis can be a
valuable tool. However, sensory experiments can be very time consuming and have poor
reproducibility due to differences in odor and taste sensitivity (Gray, 1978). Another
problem encountered with sensory testing is that the vocabulary used to describe flavor
defects can vary greatly (Frankel, 1991). Some of the terms used are musty, rancid, stale,
sour, mettalic and bitter (Civille and Dus, 1992). Also, various storage conditions can

result in different flavor descriptions. One technique that has been developed to avoid

some of these problems is quantitative descriptive analysis (Stone et al., 1974). However,
this method still involves the need for careful subject selection, training and repeated
judgments and a lot of statistical analysis. While any method used to measure lipid
oxidation will ultimately need to be correlated to a sensory method (Gray, 1978), a less

subjective method needs to be used as the primary determinant of lipid oxidation.

2.2.1. TBA Test

The most common chemical method for measuring lipid oxidation in food is the
2-thiobarbituric acid (TBA) test. This method was developed to measure malonaldehyde
(MA). Malonaldehyde is formed in lipid materials either as the by-product of enzymatic
thromboxane A2 production from arachidonic acid or as an end product from
nonenzymatic oxidative degradation of polyunsaturated fatty acids (Rahaijo and Sofos,
1993). One mole of MA reacts with two moles of thiobarbituric acid to form a pink
complex which can then be measured spectrophotometrically at 532 nm (Botsoglou et al.,
1994). The extent of oxidation is usually expressed as milligrams of malonaldehyde per
kilogram of sample.

The TBA test can be performed (a) by directly heating the sample with the TBA
reagent, then measuring the soluble pink complex formed (Turner et al., 1954; Sinnhuber
and Yu, 1958), (b) by distillation of the sample to remove volatiles and then reacting the
distillate with TBA (Tarladgis et al., 1960; Rhee, 1978), (c) using organic solvents to
extract the lipid from the sample and reacting the extract with TBA (Pikul et al., 1983,
1989) or ((1) using aqueous trichloroacetic acid (Witte et al., 1970; Newburg and Concon,

1980) or using perchloric acid (Salih et al., 1987; Pikul et al., 1989) to extract the MA

and then reacting the MA with TBA. These methods have been found to correlate well
with sensory scores of oxidized meats; however, this test is often criticized because TBA
has been found to react with substances other than MDA. For that reason, it is now
referred to as the thiobarbituric acid-reactive substances (TBARS) test (Gray and
Monahan, 1992).

Several variations of the TBA test have been developed to improve sensitivity and
specificity. Botsoglou et a1. (1994) have optimized the aqueous extraction method by
using third-derivative spectrophotometry in place of conventional spectrophotometry.
Third derivative spectrophotometry is used when spectral interference obscures the
analytical band. This method has allowed them to develop a rapid, aqueous acid
extraction procedure which is sensitive and specific. Another variation of the TBA test
was developed which allowed for the measurement of saturated aldehydes. A yellow
pigment is formed following the reaction of one mole of aldehyde with one mole of TBA.

The reaction can then be measured spectrophotometrically at 455nm (Kosugi and

Kikugawa, 1986).

2.2.2. Hexanal

Due to the limitations of the TBARS method. alternative ways of monitoring lipid
oxidation have been investigated. Hexanal is a six carbon aldehyde that is one of the
major secondary products formed by the oxidation of linoleic acid (Gray and Monahan,
1992). Tamura et a1. (1991) measured the amount of reactive products formed in a model

oxidation system containing Fe2+ and 11302 and different unsaturated fatty acids. The

products were derivatized. N-methylhydrazine was used for the 0t,B-unsaturated
aldehydes and the B-dicarbonyl compounds and cysteamine for the saturated normal
aldehydes, like hexanal. The quantity of the derivative was measured by GC. Hexanal
was one of the major products produced from arachidonic acid (43 nmol/mg) and linoleic
acid (141 nmol/mg).

Hexanal concentration has been found to correlate with sensory scores and
TBARS values in a variety of products undergoing lipid oxidation. The hexanal
concentration and TBA numbers of cooked ground pork, stored for 35 days at 4°C was
found to have a correlation coefficient of 0.995 and that as TBA and hexanal numbers
went up, flavor acceptability decreased (Shahidi et al., 1987). Hexanal and TBA numbers
of freshly cooked, stored and reheated beef muscle showed a high degree of correlation to
sensory scores (St. Angelo et al., 1987). Sensory scores from that study indicated that
hexanal could be used a marker compound for lipid oxidation. The TBA numbers and the
areas of three major GC peaks. one of which was determined to be hexanal. of chicken
patties showed a significant positive correlation (r = 0.97) (Ang and Young, 1989). The
patties had been frozen for about 6 months at -34°C, then thawed, cooked and stored at
4°C for 5 days. A positive correlation was also found between the hexanal concentration,
TBA values and sensory scores of restructured chicken nuggets stored at -20°C for 6
months (Lai et al., 1995). These studies demonstrate that hexanal has the potential to be

used as an indicator of lipid oxidation in meat.

II

2.3. Oxidation of Chicken
2.3.1. Fatty Acid Composition of Chicken

More than half of the fatty acids (68.3%) found in chicken fat are unsaturated (Table
2.1.) Since unsaturated fatty acids are highly susceptible to oxidation, it can be expected
that chicken meat will oxidize easily and rapidly. The major unsaturated fatty acids in
chicken are oleic and linoleic. Formaldehyde is the prominent oxidation breakdown
product from oleic acid (Tamura et al., 1991). Hexanal is the major secondary oxidation
product of linoleic acid (141 nmol/mg of free fatty acid), followed by malonaldehyde as
the next highest product (67.3 nmol/mg of free fatty acid) (Tamura et al., 1991). Because
chicken consists of a high concentration of linoleic acid (19.6 % of total fat), it is a good
source for hexanal and malonaldehyde during oxidation. As a result, the two methods

most commonly used to monitor lipid oxidation in chicken are TBA test and hexanal

concentration.

2.3.2. TBA Values for Oxidized Chicken

Rhee et al. (1996) examined the lipid oxidation potential of beef, chicken and
pork. They found chicken thigh to have the highest fat content (5.99% raw and 7.46%
cooked) and chicken breast the lowest (1.41% raw and 1.92% cooked) of all the meat cuts
examined. The TBA value for cooked chicken thigh was 14.5 mg malonaldehyde/kg meat
while the TBA value for cooked chicken breast was 7.5 mg malonaldehyde/kg meat, after
6 days storage at 4°C. It has also been determined that the diet of the animal can have an

effect on the oxidation potential of meat. The composition of lipids in the subcellular

12

Table 2.1. Unsaturated fatty acid content of chicken fat.

 

MML—

 

Abbreviation Common Name Content (% of total fat)‘
14:1 Myristoleic acid 0.3

16:1 Palmitoleic acid 5.7

18:1 Oleic acid 37.3

18:2 Linoleic acid 19.6

18:3 a-Linolenic acid 1.0

20:1 Eicosenoic acid 1.1

20:4 Arachidonic acid 0.1

 

'Adapted from USDA (1979).

13

membranes is influenced by the fat intake of the diet and these changes reflect a change
in susceptibility to peroxidation (Asghar et. al.. 1988).

Whang and Peng (1987) measured lipid oxidation in raw chicken skin and muscle.
The TBA numbers for chicken thigh increased by a factor of 10 while the TBA numbers
for skin and breast doubled after 8 days storage at 4°C. Chicken thigh had a higher
concentration of unsaturated fatty acids and oxidized faster than the breast meat.

Freezing and cooking can also have an effect on the rate of lipid oxidation. Pikul
et al. (1984) examined the effect of frozen storage and cooking on lipid oxidation in
chicken meat. They found a 2.5 fold increase in TBA numbers for both chicken breast
and thigh after 6 months storage at -20°C. While freezing decreased the rate of oxidation,
oxidation did still occur. The study also showed cooking had an effect on lipid oxidation.
The TBA values for the cooked meat. stored at -20°C for 6 months prior to cooking, was
83% higher in malonaldehyde concentration compared to the uncooked sample that had
also been stored for 6 months. Freezing rates had no effect on the amount of oxidation in
chicken breasts (Tomas and Anon. 1990).

Akamittath et a1. (1990) examined the effect of salt. as well as combinations of
salt with phosphates and antioxidants. They found that salt had a significant effect on the
oxidation of restructured turkey steaks. The TBA number of turkey steaks without salt
increased from 0.49 mg malonaldehyde/kg meat to 0.90 mg malonaldehyde/kg meat over
6 days storage at -10°C. The TBA numbers for steaks with salt increased from 1.34 mg

malonaldehyde/kg to 3.23 mg malonaldehyde/kg during the same storage period.

Phosphates have some antioxidative effect in turkey steaks, but the prooxidant effect of

salt and other free ions was eventually able to overcome the antioxidant effect.

2.3.2. Hexanal Concentration in Oxidized Chicken

Although a large number of researchers have used hexanal as a marker compound
for oxidation in chicken, most have used GC to quantify hexanal. The values obtained
have varied widely, mainly due to differences in cooking, storage conditions and
additives, all of which affect the extent of lipid oxidation, as discussed above. The GC
methods most commonly used for monitoring flavors in foods were outlined by
Reineccius (1996). The most commonly used method to isolate flavor compounds prior to
GC anaysis is headspace concentration, in which the sample is purged with an inert gas
and the volatiles trapped. The volatiles are then removed from the trap either thermally,
by heating, or chemically, using a solvent.

Ramarathnam et a1. (1991) used headspace concentration followed by solvent
extraction of the volatiles using n-pentane. They determined the hexanal concentration.
after 24 hour storage at 4° C, to be 9.84 ug/g meat for uncured chicken (ground chicken
cooked in a water bath with no additives) and 0.11 mg/kg for cured chicken (salt, sugar,
sodium ascorbate. sodium tripolyphosphate and sodium nitrate were added to the chicken
prior to cooking). The hexanal concentration of cooked white and dark meat of chicken
was found to have increased from 64 ng/g meat at day zero to 314 ng/g meat after 15 days
at 4°C for the white meat and from 15 ng/g meat to 1395 ng/g meat after 10 days for the

dark meat (Ajuyah et al., 1993). This study used a modified headspace technique. During

15

the purging step, the volatiles were trapped in a flask submerged in liquid nitrogen. The
resulting distillate was then allowed to thaw at room temperature prior to injection onto
the GC. Another study looked at the concentration of hexanal in roast chicken after 5
days storage at 4°C (Dupuy et al., 1987). The concentration was measured using a
combination of headspace concentration and direct injection. The concentrations were
measured on days 0, l, 3 and 5 and were found to be 0.09, 6.9, 10.8 and 14.6 jig/g meat
respectively. The hexanal concentration of chicken nuggets, which had been fried, then
frozen for 6 months without any antioxidants, was found to be 18.5 ug/g meat (Lai et al.,
1995). Hexanal concentration was determined using headspace concentration with
thermal desorption of the volatiles from the traps.

Although hexanal concentration by GC has been used by a number of researchers,
analysis by GC can be a long and involved process. Some of the drawbacks of GC
include expensive equipment, being able to analyze only one sample at a time and the
possibility of peroxide decomposition to additional volatiles by the heat used to drive the
extracted volatiles onto the GC (Ajuyah et al., 1993). An alternative to GC would be an
enzyme linked immunosorbent assay (ELISA) to measure hexanal concentration. These
assays have the advantages of reduced assay time, sample size and equipment expense
compared to GC (Samarajeewa et al., 1991). An ELISA for hexanal would provide a

faster and simpler way to monitor lipid oxidation.

16

2.4. ELISA Development

Antibodies are proteins produced by the body in response to foreign molecules.
An antibody binds specifically to an antigen and this characteristic is the reason a variety
of methods have been developed utilizing this interaction (Harlow and Lane, 1988). One
kind of immunoassay that is commonly used is ELISA. ELISAs involve the use of
enzyme labeled antibodies or antigens and take advantage of the antibody-antigen

complex that is formed. The first part of developing an ELISA is producing an antibody.

2.4.1. Conjugate Production and Determination of Modification

In order for a compound to elicit an immune response, which is the first step in
antibody production, it must be immunogenic. Size is generally the determining factor of
immunogenicity. Compounds with a molecular mass of less than 5000 daltons generally
are too small to elicit an immune response and are referred to as haptens (Harlow and
Lane, 1988). Haptens can be made immunogenic by conjugating them to proteins. This
conjugate is then large enough to cause an immune response. Hexanal, at 100 daltons, is a
hapten and would require conjugation to a protein to become immunogenic.

Ketones and aldehydes, like hexanal. can react with primary amines to form
imines, which are also known as Schiff bases (Wade, 1987). The reaction is a
condensation, involving the joining of two organic molecules with the elimination of
water (Figure 2.1). However, the double bond in the Schiff base is unstable. Sodium
cyanoborohydride will reduce this double bond to a single bond (Jentoft and Dearbom,

1979). This results in a stable conjugate. Means and Feeney (1968) determined that it is

17

025323255 8368 53> Emmi—.8 38%
a 8 3262 E 83 onm 2; .83 onm m E 8:62 35 65:8 .0255 a can 355; Lo .8388 :osmmcoucoo ._.N oczwi

2%:38 633m oEErEoSnocio Eaton omnm tEom

fun:uazohmuamuioza $ M:mzuaz + a:Damon:oazoazoionzé

ommpkzzom 38on oEEm mecm

of + funnies:okzoazuionzi i M:U-N:u-£o-£u-£u-:ouo + £221

only the e-aminolysine groups of the proteins that are modified when aliphatic aldehydes
are added with sodium cyanoborohydride at pH 9 and 0°C.

Hexanal can be conjugated to proteins through the Schiff base reaction. A
research group looked at the ability of acetaldehyde, an aldehyde smaller than hexanal, to
bind to bovine serum albumin (Donohue et al., 1983), and found both stable and unstable
adducts were formed. The unstable adducts were reduced to form stable adducts by using
sodium cyanoborohydride. The optimal reaction conditions for reducing the unstable
adducts to stable ones were determined to be a pH of 9.5, about 0.1 mM sodium
cyanoborohydride and a reaction time greater then 60 min at room temperature.

To use these conjugates for immunizations or as antigens in immunoassays, it is
necessary to know how much of the protein has been modified. as an indirect way to
determine the concentration of the compound conjugated to the protein. One way to do
this is to determine the number of free amino groups that are left on the protein after
modification (Habeeb. 1966). This is accomplished by using 2,4.6-trinitrobezenesulfonic
acid (TNBS) to react with free amino groups. The absorbance is read at 335 nm and it

was found that there is a linear relationship between absorbance and free amino group

concentration.

2.4.2. Monoclonal Antibody Production
Once a suitable conjugate has been made, or if the compound of interest is
immunogenic without conjugation, monoclonal antibody production can be started. There

are three characteristics of monoclonal antibodies that define their usefulness- specificity

for binding with an antigen, homogeneity (they are produced from only one clone and
are not a mixture of antibodies) and ability to produce monoclonal antibodies in unlimited
quantities (Harlow and Lane, 1988). Monoclonal antibodies were first reported by Kohler
and Milstein (1975). Monoclonal antibodies are produced by hybridoma cells, which are
hybrids between myeloma cells and spleen cells. The spleen cells are obtained by
removing the spleen from mice or rats which have been immunized with the antigen and
homogenizing the tissue. Generally, the easiest to use are BALB/c mice or LOU rats
because most of the myelomas commonly used for fusions are derivatives of mylomas
from these species (Galfre and Milstein, 1981). The first step in the process of producing
a monoclonal antibody is to begin immunizing mice or rats. The injection is prepared by
emulsifying 1-5 mg/mL of the antigen with an equal volume of Freund’s complete
adjuvant (Galfre and Milstein, 1981). Injections are then repeated at 3-5 week intervals,
using an emulsion of the antigen with F reund’s incomplete adjuvant. About 10 days after
injecting, a sample of blood is taken and tested for antibodies using an immunoassay. The
animals producing antibodies with high specificity and avidity are used in the fusion. The
animals are sacrificed and their spleens removed and homogenized. Spleen cells are not
able to survive in tissue culture, but by fusing them with myeloma cells, which can
survive in tissue culture, a hybrid that can survive in culture indefinitely is obtained.
Polyethylene glycol (PEG) is used to dissolve the cell walls and allow the fusion to occur.
The fused cells are selected using medium containing hypoxanthine, aminopterin and
thymidine (HAT medium). The myelomas cells are mutants that lack either the enzyme
hypoxanthine guanine ribosyltransferase or thymidine kinase. These mutants cannot grow

in a medium containing aminopterin supplemented with hypoxanthine and thymidine.

20

Since unfused spleen cells can not survive in culture, it is only the fused cells that are

capable of growing. After fusion, individual hybridomas producing the desired antibodies

can be selected and further cloned.

2.4.3. Classification of ELISA Types

Once an antibody has been produced, an ELISA format needs to be selected.
ELISAs are generally classed as one of two types. They are either competitive or non-
competitive (Clark and Engvall, 1980). Competitive ELISAs are ones where the
unlabeled antigen and the enzyme labeled antigen compete for a limited number of
antibody sites. Non competitive ELISAs are where the antigen or antibody that is to be
measured are allowed to react with an excess of the other reactant.

The first kind of competitive ELISA is when the antibody is attached to a solid
support. Any unattached antibody is washed away and a fixed concentration of enzyme
labeled antigen is added with either a known concentration of standard antigen or an
unknown concentration of test antigen. After washing again to remove the unreacted
antigen, the enzyme labeled antibody-antigen complex is incubated with a substrate. The
enzyme reaction is stOpped and the concentration determined spectrophotometrically. The
measured end product is inversely proportional to the concentrations of standard antigen
or test antigen.

The second kind of competitive ELISA involves attaching the antigen to the

support and using an enzyme labeled antibody (Figure 2.2.). The free antigen that is

 

 

 

 

 

 

 
 
 
 
 

.<m3m 025388 82:2: .8 gamma .~.m 05m:—

.EBQ

v5 2: 8388 can 68:33

.3 330:8 Sontag 388
3.33 0835 £3» 8535 .m

Sconces can
295% S 28.8% .23 2335 .m

.5325 2: Lo EoEsomu< .—

22

added as a standard with the labeled antibody competes for the antibody with the attached
antigen. A variation of this is when an enzyme labeled second antibody is used, one that
is specific to the lgG of the animal species the first antibody was produced in. In this
case, unlabeled antibody is added with the free antigen. After washing, the unlabeled
antigen-antibody complex is incubated with enzyme-labeled anti-lgG. A substrate is
added and the concentration is quantitated as above (Clark and Engvall, 1980).

There are two main kinds of non-competitive ELISAs. The first type is a single
site or one antibody assay. The antibody is attached and standard or test antigen is
allowed to react with the antibody. After washing, excess enzyme labeled antigen is
added and will attach to any unreacted antibody. Substrate is added after washing and the
reaction stopped and measured as in the competitive ELISA. The concentration of the
standard or the sample in this ELISA is inversely proportional to the measured end
product. Another kind of single site non-competitive ELISA is when test or standard
antigen is allowed to react with an excess of enzyme labeled antibody. The reaction
mixture is then added to an excess of immobilized antigen. As above, the concentration is
inversely proportional to end product.

The second kind of non-competitive ELISA is a two site or sandwich ELISA
involving two antibodies. Attached antibody is incubated with the test or standard
antigen. After washing, an enzyme labeled second antibody is added. As a result, the
concentration is proportional to the end product (Clark and Engvall, 1980).

The ELISA format chosen depends on the type of antibody available and the

purity of the antibody, as well as whether it is the presence or quantity of antigen or

23

antibody that is to be measured (Harlow and Lane, 1988). Competitive ELISAs are highly
specific, however, their use is limited when the test solution contains serum, urine or
tissue extracts (Clark and Engvall, 1980). These solutions may contain enzymes such as
proteases and noncompetitive enzyme inhibitors that may alter the activity of the enzyme

labeled antigen or antibody if they are incubated with it. This problem is avoided by

using a noncompetitive ELISA.

2.5. Examples of ELISAs Currently in Use
2.5.1. ELISAs in the Food Industry

ELISAs have been used to detect a variety of compounds in food (Samarajeewa et
al., 1991) such as bacteria, molds, viruses, pesticides, anabolic agents, and adulterants. In
the meat industry, ELISAs have been used for detection of the meat species or minor
meat protein, detection of non-meat products in meat. as well as testing for Salmonella
and Staphylococcal enterotoxins in meat products (Fukal, 1991). Concern about proper
cooking has led to the development of ELISAs to enzymes that denature as they are
heated. Monoclonal antibodies were used in a sandwich ELISA to monitor endpoint
cooking of poultry breasts by testing for lactate dehydrogenase. a protein that denatures
during cooking and can therefore be used as an indicator of the final temperature
(Abouzied et al., 1993).

Other food contaminants, such as pesticides and mycotoxins, can also be tested
for by ELISA. Mycotoxins are secondary metabolites produced by molds. They are of

low molecular weight and must be conjugated to a carrier protein before they are capable

24

of producing an immune response (Pestka et al., 1995). Even though the compounds are
small, a large number of ELISA kits are commercially available for detecting a wide

range of mycotoxins, such as aflatoxin, ochratoxin, deoxynivalenol, and zearalenone.

2.5.2. Acetaldehyde ELISAs

Outside of the food industry, ELISAs are being used for detection of acetaldehyde
adducts that are formed when acetaldehyde condenses with plasma proteins (Israel et al.,
1992). Acetaldehyde is the primary product of ethanol metabolism and is believed to
cause hepatic injury by binding to macromolecules (Donohue et al., 1983). This research
group examined the ability of acetaldehyde to bind to proteins, specifically, bovine serum
albumin (BSA), and found that both stable (reduced bonds) and unstable (unreduced
bonds) adducts were formed between acetaldehyde and BSA.

Monoclonal and polyclonal antibodies have been produced to acetaldehyde
conjugated to human plasma proteins (Israel et al.. 1986). This group determined that
both monoclonal and polyclonal antibodies specifically recognized the conjugated
proteins, with no recognition of unmodified erythrocyte proteins. They also found that the
adducts were formed with the lysine side chains of the proteins and that the polyclonal
antibodies were specific to the lysine-aldehyde adduct, with no recognition of adducts
formed with tyrosine or valine.

Another monoclonal antibody to acetaldehyde adducts was produced by a
different research group (Klassen et al., 1994). This group determined whether the
antibody could differentiate between an adduct formed under reducing conditions as

opposed to non-reducing conditions. Sodium cyanoborohydride was used as the reducing

25

agent. They found one clone that was able to recognize all adducts formed under reducing
conditions, regardless of the carrier protein. This same antibody was then further tested to
determine its specificity (Thiele et al., 1994). The antibody was found to be specific to
acetaldehyde adducts that were formed with N-ethyl lysine under reducing conditions.
The antibody did not recognize adducts formed with arginine, ethylamine or lysine under
reducing conditions or with proteins modified by acetaldehyde under non-reducing
conditions. The theory that the N-ethyl lysine adduct is the epitope was further tested
using mouse and human growth factor (EGF). Both contain one alpha amino group, but

only the human-EGF has lysine residues with epsilon amino groups. The antibody reacted

with the human-EGF only.

2.5.3. Polyclonal Antibody ELISA for Hexanal

A polyclonal antibody and ELISA for the detection of hexanal-protein conjugates
was developed by Smith (1997). The antibody was characterized and found to cross-react
with heptanal (86.3%) and pentanal (11.8%). These two compounds are the aliphatic
aldehydes with one carbon more and less than hexanal. The limit of detection of the
polyclonal antibody ELISA was determined to be 7.4 ng hexanal/mL. This ELISA was
then used to monitor lipid oxidation in a model meat system and the results were
compared to those from a TBA assay and a GC method for hexanal. The ELISA was

found to correlate well with both the TBA (r = 0.85) and the GC (r = 0.89) methods.

26

2.6. Conclusion

One of the major secondary products of linoleic and arachidonic acid oxidation,
hexanal, has been shown to be a suitable marker of lipid oxidation in different meats. An
ELISA to hexanal would allow for rapid detection of hexanal and would be a useful
alternative to GC. Polyclonal antibodies have already been produced to hexanal, but
cross-react with two other aliphatic aldehydes. Monoclonal antibodies for hexanal could
prove to be more specific to hexanal, as well as more sensitive. Monoclonal antibodies
that are both specific and sensitive have been produced to acetaldehyde, an aldehyde that
is smaller than hexanal. The development of a competitive, monoclonal antibody ELISA
to monitor lipid oxidation is feasible and could prove to be an invaluable asset to

producers and researchers.

Chapter 3
PRODUCTION AND CHARACTERIZATION OF A MONOCLONAL

ANTIBODY AGAINST A HEXANAL-PROTEIN CONJUGATE

3.1. ABSTRACT

Monoclonal antibodies were produced to hexanal conjugated to bovine serum
albumin (BSA). The final clone was designated 45P1-D9 and determined to be an
immunoglobulin G I isotype. An indirect competitive ELISA was developed and used to
determine the epitope recognized by the antibody. The working range of the ELISA was 1
to 50 ng hexanal/mL. Hexanal conjugated to three different proteins [BSA. chicken serum
albumin (CSA) and keyhole limpet hemocyanin (KLH)] was recognized. while free
hexanal and the native proteins were not. The antibody cross-reacted with pentanal.
heptanal and 2-t-hexenal conjugated to CSA with cross-reactivities of 37.9%. 76.6% and
45.0%. respectively. There was less than 10% binding inhibition (at 100 ng/mL of
aldehyde) with propanal. butanal. octanal and nonanal conjugated to CSA. Conjugates
made from binding the branched aldehydes. alcohols and a ketone to CSA also gave less
than 10% inhibition of binding at the same concentration. The antibody showed no
recognition of the glycine or arginine conjugates or the unmodified amino acids (less than
3% inhibition of binding at 2 mM). The antibody showed recognition of the lysine

conjugate (17% inhibition at 2 mM). but showed higher recognition of the e-amino

N

28

caproic acid conjugate (47% inhibition at 2 mM). e-Amino caproic acid is a lysine
derivative without the an amine group, simulating a lysine contained in a protein. The
results indicate that the antibody recognizes the lysine-hexanal complex and is specific to

this complex to within plus or minus one carbon.

3.2. INTRODUCTION

Lipid oxidation can result in objectionable flavors which are generally the
deciding factor in the storage life of foods (Frankel, 1993). Lipid oxidation can also cause
changes in color and nutritional value under a range of storage conditions (Love and
Pearson, 1971). These changes are usually viewed by consumers as undesirable. The
safety of oxidized products has also been questioned. Malonaldehyde (Esterbauer, 1993)
and oxidized cholesterol (Kubow,l993) are both being investigated for their possible
roles in disease. The radicals formed during oxidation are capable of attacking
membranes. leading to impairment of membrane function (Gutteridge and Halliwell,
1990). Both the safety concerns and the economic losses, resulting from consumer
rejection of products, show there is a need for a fast and reliable indicator of oxidation in
food.

Lipid oxidation begins with the removal of a hydrogen atom from the side chain
of a fatty acid. This results in a radical which can combine with other radicals
(termination) or can continue to react by removing hydrogens from other fatty acid side
chains (Gutteridge and Halliwell, 1990). Hydroperoxides are the primary products of

lipid oxidation and are tasteless, colorless and odorless. The breakdown of

29

hydroperoxides produce the low molecular weight compounds which give oxidized food
its characteristic odor and flavor (Gray and Monahan, 1992). One of the major secondary
low weight molecules that is formed from the oxidation of linoleic acid is hexanal (Gray
and Monahan, 1992). Monitoring hexanal concentration was recommended as a possible
alternative to the thiobarbituric acid (TBA) test (Gray and Monahan, 1992) and has been
used successfully by a number of researchers for assessing the extent of lipid oxidation in
foods.

Hexanal concentration has been correlated to sensory scores for cooked pork
(Shahidi et al., 1987), cooked beef (St. Angelo et al., 1987), restructured chicken nuggets
(Lai et al., 1995), oat, corn and wheat cereals (Fritsch and Gale, 1977), peanuts (Bett and
Boylston, 1992) and vegetable oil (Warner et al., 1978). These studies used gas
chromatography (GC) to measure the concentration of hexanal. Gas chromatographic
methods have a number of drawbacks including the need for dedicated equipment,
inconvenience of large sample analysis (only one sample can be analyzed during each
run) and possible decomposition of peroxides to additional volatiles from the heat used in
the analysis of the sample (Ajuyah et al., 1993).

ELISAs have been developed to detect or quantify a variety of substances in food
including pesticides, adulterants and mycotoxins (Samarajeewa et al., 1991) and are
accepted as alternatives to GC. The advantages of ELISAs over GC include reduction in
assay time, sample size, cleaning steps and equipment expense (Samarajeewa et al.,

1991). An ELISA to quantify hexanal would provide a fast and simple test for monitoring

lipid oxidation in foods.

30

Polyclonal antibodies have been produced to hexanal-protein conjugates prepared
by conjugating hexanal to protein via Schiff base reactions with the lysine side chains
(Smith, 1997). These antibodies were characterized using an indirect competitive ELISA.
The antibodies recognized hexanal-protein conjugates but did not recognize free hexanal
or the native protein. The immunodominant epitope was proposed to be the hexanal
modified lysine. The antibody was also found to cross—react with the pentanal, 2-t-
hexenal, (heptanal) and (octanal) (5 to 8 carbon aliphatic aldehydes), as well as 2-
methylbutanal, a branched aldehyde. The amount of cross-reactivity was 11.8%, 100%,
86.3%, 2.2% and 1.1%, respectively. The results indicated the polyclonal antibody was
most specific to the aliphatic aldehydes within one carbon length of hexanal but did not
recognize the branched aldehydes with similar carbon numbers, indicating that these
antibodies were fairly specific. The limit of detection for the polyclonal antibody was 7.4
ng hexanal/mL.

A monoclonal antibody might be more sensitive and specific for hexanal than the
polyclonal antibody. Monoclonal antibodies have been produced to acetaldehyde, a two
carbon aldehyde (Klassen et al., 1994). One of the clones recognized reduced
acetaldehyde-protein conjugates, regardless of the carrier protein, but did not recognize
unreduced conjugates or unmodified proteins when used in an indirect ELISA. This
antibody was found to have a limit of detection of 10 ng/mL. This clone was further
analyzed and it was found that the antibody recognized only the N-ethyl lysine adducts
that were formed when the protein was modified by acetaldehyde (Thiele et al., 1994).

These studies demonstrate that it is possible to produce a monoclonal antibody specific to

a low carbon number aldehyde.

31

The purpose of this study was to produce monoclonal antibodies to hexanal and
then to develop and optimize an ELISA using antibodies from the clone that showed the
highest sensitivity and specificity. Following optimization, the epitope the antibody
recognized was characterized to determine if the immunoassay might be used to monitor

lipid oxidation in actual food systems.

3.3. METHODS AND MATERIALS
3.3.1. Materials

Female mice (BALB/c, 6-8 weeks old) were purchased from Charles River
Laboratories (Wilmington, MA). Freund’s complete and incomplete adjuvants were
purchased from Difco Laboratories (Detroit, MI). The myeloma cell line P3/NS 1/1-Ag4
(NS-1) (ATCC TIB 18) was obtained from American Type Culture Collection
(Rockville, MD). Tissue culture plastic ware was purchased from Corning Laboratory
Science Co. (Corning, NY). Microtiter wells (lmmunolon-2 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). The TMB (3,3’,5,5‘ tetramethyl benzidine) substrate
and stopping buffer were purchased from Di-Agra (Sterling Heights, MI). Propanal,
butanal. pentanal, heptanal, nonanal, 2-methylpentanal, 2-methylbutanal, 3-
methylpentanal, 2-hexanol, hexanol and 2-heptanone were all purchased from Aldrich
Chemical Co. (Milwaukee, WI). All other materials were purchased from Sigma

Chemical Co. (St. Louis, MO).

32

3.3.2. Conjugate Preparation

The conjugates wereprepared as described by Smith (1997), except that the final
sodium cyanoborohydride concentration was reduced to 20 mM and as a result, no
dialysis was needed. The proteins used in this study were bovine serum albumin (BSA),
chicken serum albumin (CSA) and keyhole limpet hemocyanin (KLH). The hexanal
modified proteins were produced by adding 246 uL of hexanal to 9 mL of 16.7 mg/mL
protein dissolved in phosphate buffered saline (PBS, pH 7.4; 0.1 M sodium chloride, 0.01
M sodium phosphate, pH 7.4). After vortexing, 1 mL of 200 mM sodium
cyanoborohydride (NaCNBH3) dissolved in 0.1 N NaOH was added to the hexanal-
protein solution to reduce Schiff bases. The final concentrations of this reaction mixture
were: 15 mg/mL protein, 200 mM hexanal and 20 mM NaCNBH3. The unreduced
conjugates were prepared the same way except that no NaCNBH3 was added. The
reaction was allowed to proceed overnight at room temperature. Modification was
determined by measuring the loss of reactive amino groups in the conjugates, as
compared to the corresponding native protein, by the trinitrobenzenesulfonic (TNBS)
acid test of Habeeb (1966) as modified by Smith (1997). Conjugates were diluted to 1
mg/mL with PBS (pH 7.2; 0.1 M sodium chloride. 0.01 M sodium phosphate), aliquotted
and frozen at -20°C.

Amino acid analysis was performed by Dr. Nathalie Trottier’s lab (Department of
Animal Science) using high performance liquid chromatography. The amino acid content
of hexanal modified CSA, BSA and KLH, as well as the native form of these proteins,

was determined following derivatization of the sample with phenylisothiocyanate (Cohen

33

et al., 1989). Briefly, a volume of each conjugate or native protein, equaling 20 mg of
protein, was hydrolyzed with 2.5 mL of 4 mM norleucine and 25 mL of 6N HCl by
autoclaving the sample mixtures for 24 h. The samples were then brought up to a volume
of 40 mL using HPLC grade water. For the conjugation, phenylisothiocyanate was added
to 25pL of the hydrolyzed sample. The samples were then analyzed by HPLC to
determine the amino acid content. The limit of detection for this method was 1 picomole.

Results were expressed as molar percentage for each amino acid. Experiments were run

in duplicate.

3.3.3. Production of Monoclonal Antibodies

Two groups of 6-8 week old Balb/c mice were injected subcutaneously with 50 ug
of hexanal modified protein conjugates (conjugates were produced and checked for
modification as described in above section) in 0.1 mL 0.8% saline and emulsified with
0.1 mL of Freund’s complete adjuvant. Mice were injected with a total volume of 0.5 mL.
Five mice were injected with the BSA-hexanal conjugate and five with the KLH-hexanal
conjugate. Booster injections were given at 2 week intervals with the same injection
except that Freund’s incomplete adjuvant was used instead of the complete adjuvant. Two
weeks after the second booster injection, the mice were bled (from the tail), titers were
determined and the sera tested for antibody by an indirect competitive ELISA. The two
mice whose serum showed the lowest recognition of the native BSA, when tested by an

indirect competitive ELISA, were selected for the fusion. A final injection, containing 50

34

ug of hexanal modified protein in 0.1 mL 0.8% saline, was given three days prior to the
fusion.

The fusion procedure followed was performed as described in Galfre and Milstein
(1981) and as modified by Abouzied et a1. (1990). Briefly, the mice were sacrificed and
the spleen removed. Mouse spleen cells (2x108) were mixed with 2x107 NS-l cells using
50% polyethylene glycol as the fusing reagent. Following fusion, the cells were
suspended in Dulbecco’s modified medium containing 20% fetal bovine serum (PBS),
1% NCTC medium, 10 mM MEM sodium pyruvate solution and penicillin/streptomycin
solution (100 U/mL) and distributed into 96 well culture plates. The plates were
incubated at 37°C in a humid atmosphere containing 7% C02. After 24 h of incubation,
half of the media in the wells was removed and replaced with an equal volume of
hypoxanthine, aminopterin and thymidine selective medium (HAT medium). The cells
were fed in this manner every three days. After two weeks, the HAT medium was
replaced with HT medium (HAT medium without the aminopterin). The wells showing
growth and color change were tested for antibody production using an indirect
competitive ELISA. Hybridomas that showed production of antibodies to hexanal-protein
conjugates were expanded and cloned twice by limiting dilution (Goding, 1980). The
hybridomas whose antibodies showed the highest percent inhibition were transferred to
larger growth vessels.

Hybridomas were stored in FBS-dimethyl sulfoxide (9:1) under liquid nitrogen
(Abouzied et al., 1990). Supernatant from the cell culture of the final clone was collected

and the antibody was purified by precipitation with 50% ammonium sulfate (Hebert et al.,

35

1973) then dialyzed for two days against PBS (pH 7.2) at 4°C. The purified antibody was
then aliquotted and frozen at -20°C. The immunoglobulin subclass of the antibody
secreted by the final clone was determined by Sigma ImmunotypeTM Mouse Monoclonal

Antibody Isotyping Kit.

3.3.4. Titer Determination and Screening by Indirect ELISA

An indirect ELISA (Abouzied et al., 1990) used for titer determination was
performed by coating microtiter plates overnight at 4°C with 100 uL hexanal-KLH or
hexanal-BSA conjugate (lug protein/mL), diluted in 0.1 M carbonate buffer (pH 9.6).
Plates were washed 4 times with PBS (pH 7.2) containing 0.05% Tween-20 (PBS-T). The
non-specific binding sites were then blocked by adding 300 uL of 0.5% casein in PBS
(pH 7.2) to each well and incubating 30 min at 37°C. After washing 4 times with PBS-T,
50 uL of serially diluted sera (in PBS; pH 7.2) were added and incubated for 30 min at
37°C. After incubation. the plates were again washed 4 times with PBS-T. To each well,
100 uL of GAM-IgG HRP. diluted 1:500 in PBS with 0.5% casein, was added. After
incubating 30 min at 37°C, the plate was washed 8 times with PBS-T. Peroxidase binding
was determined using 2,2’-azinobis (3-ethylbenzothiazoline-6-sulfonic acid)-H202
substrate as described by Pestka et a1. (1982). Absorbance was measured at 405 nm using
a THERMOmax plate reader (Molecular Devices, Menlo Park, CA). The titer of the
serum was arbitrarily determined as the dilution showing an absorbance that was twice

the absorbance of the pre-immune serum at the same dilution.

36

The indirect competitive ELISA used for screening the mice and hybridomas for
antibody production was the same as above except that after blocking the nonspecific
sites, 50 pL of cell culture supernatant or sera were added with either 50 pL of PBS (pH
7.2) as the zero concentration or with 50 pL hexanal modified KLH at 10 pg protein/mL
as the competition. The percent inhibition at 10 pg protein/mL was determined by

dividing the absorbance at that concentration by the absorbance at zero concentration then

subtracting from 1.

3.3.5. Protein Specificity by Indirect Competitive ELISA

Hexanal modified and native CSA, BSA and KLH were tested for antibody
recognition. The indirect competitive (IC) ELISA used was performed by coating
microtiter plates overnight at 4°C with 100 pL hexanal-CSA conjugate (1 pg protein/mL),
diluted in 0.1M carbonate buffer (pH 9.6). Plates were washed 4 times with PBS (pH 7.2)
containing 0.05% Tween-20 (PBS-T). Purified antibody (50 pL of 1.6 pg protein/mL in
PBS, pH 7.2) was added with 50 pL of hexanal modified or native protein at
concentrations of 0 to 10 pg protein/mL and incubated for 15 min at room temperature.
After incubation, the plates were again washed 4 times with PBS-T. To each well, 100 pL
of GAM-lgG HRP, diluted 1:500 in PBS with 0.5% casein, was added. After incubating
15 min at room temperature, the plate was washed 8 times with PBS-T. Peroxidase
binding was determined using a TMB buffer. After 10 min, the reaction was stopped with
stopping buffer and the absorbance was read at 450 nm. A standard curve was generated

by plotting the concentration of hexanal modified protein against binding inhibition.

37

Since each protein contains different amounts of lysine, the hexanal concentration
was calculated from the protein concentration, based on the moles lysine/mole protein
and the percent modification. The following formula was used, showing the conversion of

protein concentration to hexanal concentration:

 

pg protein x moles lysine x lmole protein x lgram x fl°gmoles = gmoles lysine
mL 1 mole protein MW (g/mole) protein 10" pg lmole mL

The pmoles of lysine are multiplied by the % modification to adjust for differences in
lysine modification during the conjugation. Since one mole of lysine will bind one mole

of hexanal, these concentrations are equal :

gmole hexanal x 100g(MW of hexanal) x lmole x l_Q°_r_ig = ng hexanal
mL 1 mole 10° pmole 1 g mL

The moles of lysine in each protein were calculated from the amino acid analysis. The
percent mole 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 (minus 18, for the loss of a water when it is a part of a protein) to obtain the moles
of lysine per mole of protein. The molecular weight used for KLH was 1700 kD, the
average range for the commercial protein. The molecular weights used for BSA and CSA
were 66,000 and 66,300 daltons, respectively. An example of this formula, converting 1

pg protein/mL concentration of hex-CSA to hexanal concentration is shown in Appendix

38

A. Standard curves were generated by plotting the concentration of hexanal against the
binding inhibition produced by each protein conjugate.

The antibody was also tested for recognition of free hexanal using the above
ELISA format. Different aliquots of hexanal were added to 30% methanol so that the
final concentration of free hexanal was 0, 0.05, 0.1, 0.25, 0.5 and 1.0 pg/mL. Unreduced
conjugates (prepared as stated under the conjugate preparation section) were also tested

for recognition in the indirect competitive ELISA at concentrations of 0 to 5 pg

protein/mL.

3.3.6. Precision

lntra-assay and inter-assay precision, or reproducibility, of the IC-ELISA was
determined (Deshpande, 1996). The concentrations used for determining the precision
were 0, 5, 10, 20 and 40 ng/mL of hexanal, which corresponds to 100, 80, 65, 55 and
50% of the maximum binding (80 ), respectively. The average, standard deviation and
coefficient of variation was determined for all replicates at each concentration. The intra-
assay variability was determined using 16 replicate wells and the inter-assay variability

was determined using 32 replicate wells from eight different plates.

3.3.7. Antibody Cross-Reactivity by IC-ELISA
Conjugates were produced, using the conjugation procedure in section 3.3.2., by
replacing either the hexanal or the protein portion of the conjugate. CSA was modified

with the following compounds in place of hexanal: propanal, butanal, pentanal, heptanal,

39

octanal, nonanal, 2-methylpentanal, 2-methylbutanal, 3-methylpentanal, 2-t-hexenal, 2-
hexanol, hexyl alcohol and 2-heptanone. The final concentrations of these reactions were:
15 mg/mL protein, 200 mM hexanal substitute and 20 mM NaCNBH,. The standard
curve used in the ELISA was converted from CSA concentration to aldehyde, alcohol or
ketone concentration in order to compare the ELISA results to those obtained by GC,
which is generally expressed as microgram or milligram of hexanal per gram of sample.
The formula for converting protein concentration to hexanal concentration was used, but
substituting the molecular weight of each reactant for that of hexanal (100 daltons).

Hexanal was used to modify the following amino acids: glycine, L-arginine, L-
lysine and s-aminocaproic acid. The final concentrations of these reactions were: 200 mM
amino acid, 200 mM hexanal and 20 mM NaCNBH3. The amount of modification of all
conjugates was determined using the TNBS assay.

The cross-reactivity of the monoclonal antibody was determined by testing the
aldehyde, alcohol, ketone and amino acid conjugates in the IC-ELISA described in
section 3.3.5. The conjugates were added with the antibody at concentrations of 0 to 100
ng/ml of aldehyde. alcohol or ketone, or 0 to 2 mM amino acid. Cross-reactivity was
determined by dividing the concentration of hexanal-CSA required for 50% binding
inhibition by the concentration of aldehyde, alcohol, ketone or amino acid conjugate

required for 50% binding inhibition multiplied by 100. All determinations were made in

at least triplicate.

40

3.3.8. Statistics
Standard error of the means was determined for each curve, at each concentration,

using one way analysis of variance (SAS, SAS Institute, Inc., Cary, NC).

3.4. RESULTS AND DISCUSSION
3.4.1. Antigen Preparation

Amino acid analysis of the native and modified CSA, BSA and KLH showed the
lysine content of the hexanal modified proteins to have decreased to an undetectable
concentration. The lysine content of native CSA, BSA and KLH was 8.84, 10.79 and 4.37
mole percent, respectively (Table 3.1.). The calculated moles of lysine per mole of
protien, based on the amino acid analysis, were 580, 56 and 46 for KLH, BSA and CSA,
respectively. Harlow and Lane (1988) and Alaiz and Giron (1994) reported 59 moles of
lysine in BSA. Block (1956) reported 49 lysine residues in CSA. Therefore, the
calculated values for these proteins are reasonable.

The amino acid values of the native proteins are comparable to those reported in
Alaiz and Giron (1994) for BSA and Block (1956) for CSA. The serine and glycine
contents of the KLH conjugate also decreased as compared to native KLH, but did not
show a significant change. Since serine and glycine do not have reactive sidechains that
are susceptible to reacting with lipid oxidation products (Hall, 1987), it is highly unlikely
they are being modified by hexanal.

The amino acid analysis suggests that the conjugates were highly modified. The

percent modification of the hex-CSA, hex-BSA and hex-KLH conjugates was determined

Table 3.1.

41

Amino acid content of native and hexanal modified (hex-) chicken serum

albumin (CSA), bovine serum albumin (BSA) and keyhole limpet

 

 

 

hemocyanin (KLH).
Mole Percentl

Amino Native Hex- Native Hex- Native Hex-
Acid CSA CSA BSA BSA KLH KLH
Asp 9.20: 0.35 12.71:0.04 11.01: 0.05 12.33:0.06 13.08:0.40 13.99:0.04
Glu 15.06:0.04 17.24: 0.68 14.58: 0.11 16.47:0.ll 10.75:0.97 11.81:0.08
Ser 6.84: 0.05 7.58: 0.11 4.78: 0.0 5.44:0.07 8.12:2.57 6.65:0.07
Gly 5.01: 0.10 5.34: 0.06 3.62:0.10 3.89:0.05 7.35:1.07 7.12:0.00
His 2.33: 0.06 2.44: 0.51 2.97:0.13 3.55:0.03 5.23:0.32 5.95:0.04
Arg 4.84: 0.09 5.18: 0.39 4.47: 0.06 5.00:0.32 4.62:0.35 5.10:0.12
Thr 3.70: 0.25 4.17: 0.26 5.41: 0.15 6.36:0.60 4.93:0.11 5.15:0.07
Ala 8.67: 0.33 9.21: 0.13 8.95: 0.11 9.80:0.32 7.48:0.20 7.95:0.38
Pro 6.11: 0.02 6.72: 0.07 6.61: 0.02 7.32:0.15 6.02:0.15 6.31:0.16
Tyr 3.77: 0.02 4.03: 0.07 3.71: 0.04 4.30:0.08 4.49:0.02 4.78:0.04
Val 5.79: 0.03 6.02: 0.02 5.79: 0.04 6.31:0.08 5.48:0.11 5.43:0.07
Met 2.98: 0.06 3.11: 0.04 uni-MHZ "Hunt: 2 uni-nun 2 neutral-u" 2
Ile 4.49: 0.02 4.61:0.05 1.90: 0.04 2.28:0.11 4.46:0.52 4.64:0.106
Leu 7.05: 0.06 7.33: 0.41 10.31: 0.01 1 1.14:0.14 7.86:0.38 8.65:0.18
Phe 5.27: 0.02 5.38: 0.20 4.95: 0.01 5.44:0.02 5.79:0.25 6.33:0.24
Lys 8.84: 0.02 ******* 2 10.79:0.01 ******** 2 4.37:0.14 "*""" 2

 

2 Values below the limit of detection.

Values are expressed as mean : standard deviation.

42

by TNBS assay to be 95.3%, 94.8% and 86.8%, respectively. Previous amino acid
analysis of the unmodified chicken serum albumin (CSA) and the hexanal modified CSA
(hex-CSA) showed the available lysine content to be 11.6 and 0.2 mole %, respectively
(Smith, 1997). These results indicated a 99% reduction in available lysine which
corresponded to the 99% modification of the conjugate that was measured by the TNBS
assay. Based on the results from both of these studies, the TNBS results were believed to

be an accurate measurement of the amount of modification.

3.4.2. Hybridoma Production

Hexanal modified keyhole limpet hemocyanin (hex-KLH) and hexanal modified
bovine serum albumin (hex-BSA) used for immunization of the mice were determined to
have 74.6% and 97.4%, respectively, of their lysine groups modified by hexanal. All of
the mice showed an immunogenic response one week after the second boost. Titers were
determined by indirect ELISA using serial dilutions of the mouse sera obtained through
tail bleeding 5 weeks after the initial immunization. There was a problem with the
bleeding of the mice prior to immunization with the hex-BSA conjugate and not enough
serum was removed from each mouse to allow for individual testing, so the sera was
pooled and then tested (Figure 3.1.). The titers for the hex-BSA injected mice ranged
from 4.0x104 to 5.3x10‘, 5 weeks after the initial injection (Table 3.2.). There was enough
serum removed from each of the hex-KLH injected mice to allow for individual testing of
the sera obtained prior to injection (Figure 3.2.) The titers for the hex-KLH injected mice
were higher than the hex-BSA injected mice, ranging from 6.9x104 to 9.5x10‘ (Table

3.2.). An indirect competitive ELISA was performed using the sera from the first

Absorbance (405 nm)

Figure 3.1.

0.9 »

0.8
0.7

0.6

0.5 .

0.4

0.3

0.2

0.1

43

 

+Mouse A
t +Mouse B
+Mouse C

+Mouse D

-6— Pre-lmmune

 

 

 

 

 

Dilution ( x 103 )

Production of antibodies against hexanal modified bovine serum albumin
(BSA) was determined by indirect ELISA. Sera were obtained from the
mice by tail bleeding 5 weeks after the initial injection. The initial
injection and booster injections at 2 and 4 weeks contained 50 pg of
hexanal modified BSA. Serum for mouse E was lost. Due to problems
with pre-immune bleeding, sera from all 5 mice were pooled so as to have
enough sera to perform the ELISA. Microtiter wells were coated with

100 pL of 5 pg/mL of hexanal modified keyhole limpet hemocyanin
(KLH) and incubated with different dilutions of sera.

44

Table 3.2. Sera titers against hexanal modified bovine serum albumin (hex-BSA) or
keyhole limpet hemocyanin (hex-KLH) determined 5 weeks after the initial

 

 

 

immunizationl
Titer (x104)
Mouse Hex-BSA Hex-KLH
A 4.0 8 7
B 4.0 7 2
C 4.6 6.9
D 5.3 8 l
E 7 9.5

 

2

Titers were determined 5 weeks after the initial injection. The mice
received two booster injections, at week 2 and 4, with 50 pg of hexanal
modified protein. The titer was arbitrarily determined to be the dilution of
serum in which the absorbance is twice the absorbance of the pre-immune
serum at the same dilution as measured in an indirect ELISA. Microtiter
wells were coated with 100 pL of 5 pg/mL of hexanal modified protein
(hexanal modified KLH was used to test BSA injected mouse sera and
hexanal modified BSA was used to test KLH injected mouse sera). Plates
were incubated with different dilutions of sera.

Serum for mouse E was lost during storage.

2.5
E
C
m

g 2
O
0
S
0
II
2

1

0

Figure 3.2.

1.5 ,

45

 

  
 
 
 
 
 
 
   

 

+Pre-A
+Mouse A
+Pre-B
+Mouse B
+Pre~C
+Mouse C
+Pre-D
+Mouse D
+Pre-E
-)(—Mouse E

 

 

 

 

Dilution (x103)

Production of antibodies against hexanal modified keyhole limpet
hemocyanin (KLH) was determined by indirect ELISA. Sera was obtained
from the mice by tail bleeding 5 weeks after the initial injection. The
initial injection and booster injections at 2 and 4 weeks contained 50 pg of
hexanal modified KLH. Microtiter wells were coated with 100 pL of 5
pg/mL of hexanal modified bovine serum albumin (BSA) and incubated
with different dilutions of serum. Pre-x refers to the pre-immune sera
which was removed from each mouse prior to the first injection.

46

bleeding of the hex-BSA injected mice and hex-BSA at concentrations of 0, 0.01, 0.1, 1,
10 and 100 pg protein/mL (Figure 3.3.). Sera from all of the mice gave greater than 80%
binding inhibition at 100 pg protein/mL. This indicates that the mice were producing
antibody that could recognize not only the hex-BSA attached to the plate, but also free
hexanal-BSA.

Although the hex-KLH mice had higher titers, the hex-BSA injected mice were
used for the fusion since the hex-BSA mice showed positive results in the indirect
competitive ELISA prior to the first bleeding of the hex-KLH injected mice. Mouse B
and mouse D were selected for the fusion because their sera showed the lowest
recognition of native BSA when tested in an indirect ELISA. However, mouse A might
have been a better choice than mouse D because serum from mouse A gave higher
binding inhibition in the competitive ELISA. At 10 pg protein/mL, mouse A gave a
binding inhibition of 73% whereas mouse D was 46%. In the indirect ELISA for
determining recognition of the native BSA, at 1:1000 serum dilution, mouse D had an
absorbance of 0.694 while mouse A had an absorbance of 1.087. However, at the next
higher dilution, 1:3160, the absorbances for mouse D and A were 0.298 and 0.493,
respectively. This is a small difference in the recognition of the native protein by the sera
from these mice, so the indirect competitive ELISA results were probably a better
indication of antibody production.

Following the fusion, the cells were plated out and allowed to grow. The fusion
efficiency (the number of wells showing growth/number of wells seeded) of the plates

was determined and found to be 89% (425/480). Eight positive hybridomas were

Inhibition (%

Figure 3.3

47

 

Protein Concentration (uglmL)

Antibody production by hexanal modified bovine serum albumin (hex-
BSA) injected mice was determined by indirect competitive ELISA, using
serum from the first bleeding. Mice were tail bled five weeks after being
injected with 50 pg hex-BSA. Boosters of 50 pg hex-BSA were given at
weeks 2 and 4. Microtiter wells were coated with 100 pL of hex-BSA (5
pg protein/mL). A 1/1000 dilution of sera was incubated with free hex-
BSA at concentrations of 0.01, 0.1, 1, 10 and 100 pg/mL of BSA.

48

produced from both fusions and 3 were further cloned. The hybridoma which showed the
highest binding inhibition in the indirect competitive ELISA from the first cloning was
cloned a second time. The hybridoma from the second cloning showing the highest
binding inhibition and the least cross-reactivity with aliphatic aldehydes other than
hexanal was identified and designated as 45P1-D9. The isotype of this clone was IgG,.
The monoclonal antibody secreted by this cell line was used for the optimization of the

indirect competitive ELISA, as well as for the epitope studies.

3.4.3. Protein Specificity

The antibody was first checked, by indirect competitive ELISA, for recognition of
free hexanal. There was no recognition of free hexanal in 30% methanol at
concentrations up to 1000 ng hexanal/mL. Methanol was added to help make the free
hexanal more soluble. This concentration of methanol had been previously determined
not to cause interference with the ELISA using a polyclonal antibody (Smith, 1997) and
was found not to cause interference with the monoclonal antibody ELISA.

After determining that the antibody did not recognize free hexanal, it was checked
for recognition of hexanal conjugated to different carrier proteins. The degrees of
modification for the hex-CSA, hex-BSA and hex-KLH were 99.1%, 95.3% and 90.9%,
respectively. There was less than 5% binding inhibition with 5 pg protein/mL of the
native protein (CSA. BSA, and KLH dissolved in PBS; pH 7.2) or the protein blank
(CSA, BSA and KLH that had been through the conjugation reaction but without any

hexanal added). At 10 pg protein/mL, the hex-CSA conjugate showed 88% binding

49

inhibition, the hex-BSA conjugate showed slightly less inhibition, 78%, and the hex-KLH
conjugate showed only 42% (Figure 3.4.) The difference in the amount of binding
inhibition for each of the conjugates, at the same protein concentration, is partly due to
the moles of lysine in each protein. Since hexanal binds to the e-amino group of lysine,
the protein with the most available lysine should bind the most hexanal. Native KLH was
determined, by amino acid analysis, to have 4.37 mole % lysine as compared to 8.84 and
10.79 mole %, respectively, for CSA and BSA. Both BSA and CSA have more than twice
the number of lysine residues as KLH. Also, KLH has a large molecular weight,
compared to BSA and CSA and as a result, there are fewer moles of KLH at the same
protein concentration as BSA and CSA, as well as fewer lysines.

The hexanal concentration of the modified proteins was calculated for each
protein concentration, using the moles of lysine and percent modification. The calculated
moles of lysine per mole of protein were 580, 56 and 46 for KLH, BSA and CSA,
respectively. Although the difference in percent modification and the moles of lysine had
been adjusted for. the protein conjugates gave different binding inhibitions at the same
hexanal concentration (Figure 3.5.) At 8 ng hexanal/mL. binding inhibition for the KLH,
BSA and CSA conjugates were 7, 16 and 19%, respectively. This shows that the type of
protein has an effect on the binding inhibition. This may be due to the antibody showing
some recognition of the carrier protein. Since the mice were immunized with hex-BSA
and hex-KLH, hex-CSA was used as the competitor in the screening assay. As a result,

the antibodies selected may show a preference for the hex-CSA conjugate. Although the

Inhibition (7.)

Figure 3.4.

50

 

    

+HEX-CSA

 

+ HEX-BSA
+ HEX-KLH

 

 

0 2 4 6 8 10

Protein Concentration (ugImL)

Specificity of monoclonal antibody to hexanal modified proteins was
determined by indirect competitive ELISA. Microtiter wells were coated
with l pg/mL of hexanal modified chicken serum albumin (hex-CSA). A
1/1000 dilution of antibody (1.6 pg/mL) was incubated with competing
native or hexanal modified CSA, bovine serum albumin (BSA) or keyhole
limpet hemocyanin (KLH) at concentrations of 0.1, 0.25, 0.5, l, 2.5, 5 and
10 pg/mL protein concentration. Each data point represents 6 replicates.
The amount of modification for BSA, CSA and KLH, as determined by
TNBS assay, was 95.3%, 99.1% and 90.9% respectively. The standard
error of the means, at all concentrations, were 0.63, 0.78 and 1.43 for hex-
CSA, hex-BSA and hex-KLH, respectively.

80»

90
7o 4
§
c 60 «
0
:e
a 50
.C
S
g 40
‘6
C
a 30
20
10
0
Figure 3.5.

51

 

   
 

 

+ Hex-CSA

+ Hex-BSA

+ Hex-KLH

 

 

1 00 200 300 400 500 600 700 800

Hexanal Concentration (nglmL)

Effect of hexanal concentration on the binding inhibition of hexanal
modified proteins was determined by indirect competitive ELISA.
Microtiter wells were coated with 1 pg/mL of hexanal modified chicken
serum albumin (hex-CSA). A 1/1000 dilution of antibody (1.6 pg/mL)
was incubated with competing hexanal modified C SA, bovine serum
albumin (BSA) or keyhole limpet hemocyanin (KLH) conjugates at
concentrations of 0 to 800 ng hexanal/mL. Each data point represents 6
replicates. The amount of modification for BSA, CSA and KLH, as
determined by TNBS assay, was 95.3%, 99.1% and 90.9% respectively.

52

antibodies may preferentially bind to hex-CSA, nevertheless, they do not recognize native
CSA, indicating the antibody is recognizing the hexanal attached to CSA, not CSA.

Sodium cyanoborohydride was added to reduce the double bond that is formed by
the Schiff base reaction between the lysine side chains of the protein and hexanal. It was
also determined by indirect competitive ELISA that the antibody recognized only the
conjugates that have had this bond reduced. Klassen et a1. (1994) found their monoclonal
antibody to acetaldehyde adducts also recognized only reduced conjugates.

These results suggest that the antibody does not recognize free hexanal but
recognizes hexanal modified proteins, regardless of the protein carrier. These results are
consistent with the results found using polyclonal antibodies for hexanal (Smith, 1997).
Polyclonal antibodies did not recognize free hexanal, but needed the protein carrier.
Studies using monoclonal antibodies to acetaldehyde adducts have also reported similar
results. Israel et a1. (1986) developed an antibody to acetaldehyde which recognized the
acetaldehyde modified erythrocytes, but not the unmodified erythrocytes. Klassen et al.
(1994) found their monoclonal antibody recognized acetaldehyde conjugated to BSA,

KLH or actin, but did not recognize unreduced conjugates or unmodified proteins.

3.4.4. ELISA Optimization

The hexanal-CSA conjugate was used to optimize the ELISA because it was the
most highly modified and gave the best binding inhibition over a range of concentrations.
Protein concentrations used in the standard curve were converted to hexanal
concentration as described in the methods and materials. A representative standard curve

for the indirect competitive ELISA that was used for the epitope studies shows the

53

working range was 1 to 50 ng hexanal/mL (Figure 3.6.). The limit of detection for this
assay was 1 ng hexanal/mL. The limit of detection of the polyclonal antibody ELISA was
determined to be 7.4 ng/ml (Smith, 1997), with a working range of 7.4-740 ng
hexanal/mL. With a working range of 1 to 50 ng/mL, the monoclonal based ELISA is
more sensitive for hexanal conjugates than the polyclonal antibody assay used by Smith
(1997)

The reproducibility of the ELISA, within and between assays, was also
determined (Table 3.3.). Deshpande (1996) recommended using at least 3 analyte
concentrations that represent low (80% Ba), midrange (50%80) and high (20% Ba)
binding inhibition when evaluating precision. 80 is equal to the binding of the zero
standard, or maximum binding. For this ELISA format, Sng/mL and 40 ng/mL
represented the low and midrange binding, respectively. A higher concentration that was
equal to 80% 80 was not used because it was outside of the working range of the assay (1
to 50 ng/mL). The 10 and 20 ng/mL concentrations were used for a broader range of
concentrations within the working range of the standard curve. The coefficient of
variation (%CV) ranged from 4.15 to 6.43 for the intra- assay variability and from 5.07 to

7.71 for inter-assay variability.

3.4.5. Cross-reactivity Study
The cross-reactivity of the antibody with other aliphatic aldehydes was
determined using CSA conjugates in an indirect competitive ELISA. CSA was modified

as outlined in section 3.3.2. with different aldehydes and the extent of modification was

Binding Inhibition (7.)

Figure 3.6.

70

60-

50

40

30

20

10

54

 

 

 

 

1 0 20 30 40 50 60

Hexanal Concentration (nglmL)

Representative standard curve showing the working range of the indirect
competitive ELISA for hexanal. The working range was determined to be
1 to 50 ng hexanal/mL. Each data point represents 8 replications.
Microtiter wells were coated with 100 pL of lug/mL hexanal modified
CSA. A 1/1000 dilution of antibody (1.6 pg/mL) was incubated with hex-
CSA at concentrations of 0, 1, 5, 10 and 50 ng/mL of hexanal. The
standard error of the means was 0.869 for all concentrations except for 40
ng/mL, which was 0.929.

55

Table 3.3. Reproducibility of the indirect competitive ELISA was determined for both
within a run (intra-assay) and between runs (inter-assay).

 

Coefficient of Variation'

 

 

Concentration2 Intra-assay2 Inter-assay3
(ng hex/mL) (%) (%)

O 4.40 5.07

5 5.60 5.28

I O 6.22 6.56

20 6.43 7.71

40 4.15 7.3 I

 

Coefficient of variation was calculated by dividing the standard deviation
of the data series by the mean and multiplying by 100.

lntra-assay variability was calculated using 16 replicate wells.

Inter-assay variability was calculated using 32 replicates from eight
different plates.

56

Table 3.4. Loss of reactive amino groups in CSA solutions was measured by
trinitrobenezesulfonic acid (TNBS) assay

 

 

Solution Loss (%)l
Propanal-CSA 96. 1
Butanal-CSA 97.3
Pentanal-CSA 97.2
Hexanal-CSA 97.2
Heptanal-CSA 97.3
Octanal-CSA 97.4
Nonanal-CSA 97.4
2-t-hexenal-C SA 81 .4
2-methylpentanal-C SA 96. l
2-methylbutanal-CSA 96.0
3-methylpentanal-C SA 96.0

 

Defined as percentage decrease in TNBS reactive amino groups of native protein.

57

expressed as percent loss of reactive amino groups (Table 3.4.) The antibody did not
cross-react with the 3 (propanal), 4 (butanal) and 9 (nonanal) carbon aldehydes when
conjugated to CSA (Figure 3.7.). There was less than 8% binding inhibition at 100 ng
aldehyde/mL of these three conjugates. Although there was some recognition of the 8
(octanal) carbon aldehyde, the antibody did not cross-react with it because the binding
inhibition did not reach 50%. Antibody cross-reactivity with the 5 (pentanal) and 7
(heptanal) carbon aldehyde-protein conjugates was 37.9% and 76.6%, respectively (Table
3.5.). Since all of these aldehyde conjugates showed greater than 90% modification, the
differences in cross-reactivity were caused by differences in antibody recognition and not
by a difference in the extent of modification.

The antibody also cross-reacted (45.0%) with 2-t-hexenal when conjugated to
CSA (Table 3.5.). 2-t-Hexenal is a six carbon aldehyde with a double bond between the
second and third carbons. Jentoff and Dearbom (1979) demonstrated that NaCNBH3
could be used to reduce the double bonds in Schiff bases. The double bond of 2-t-hexenal
may also be reduced, along with the Schiff base formed between 2-t-hexenal and the
lysine residues of the protein, in the presence of NaCNBH3. This would result in the
formation of hexanal during the conjugation step. The lower cross-reactivity of the
antibody with this conjugate may be due to incomplete reduction of the double bond to
form hexanal. The lower cross-reactivity may also be due to lower lysine modification
(87% as measured by TNBS) of the protein, resulting in fewer moles of aldehyde in the
conjugate. Smith (1997) determined that the polyclonal antibody to hexanal cross-reacted

100% with a 2-t-hexenal conjugate. The conjugate was determined to be 96% modified,

58

 

, +Hexanal

 

80
70 1"
601

if. 50*
C
O
39

a 40
2
s
O

.s so
U
5
m

20

10

o

Figure3.7.

-I- Propanal
+Butanal
+Pentanal

 
 
 
 
 
 
   

+Heptanal
+Octanal

+ Nonanal
+2-t—hexenal

 

1 0 20 30 40 50 60 70 80 90 1 00

Aldehyde Concentration (nglmL)

Specificity of the monoclonal antibody to aliphatic aldehydes conjugated
to chicken serum albumin (CSA). Each data point represents 4
replications. Microtiter wells were coated with 100 pL of lpg/mL hexanal
modified chicken serum albumin (hex-CSA). A 1/1000 dilution of
antibody (1.6 pg/mL) was incubated with aldehdye-CSA conjugates at
concentrations of 0, 5, 10, 50 and 100 ng aldehyde/mL. The standard error
of the means, at all concentrations, were 2.21, 1.52,0.68, 1.09, 0.73, 0.76,
1.10 and 0.57 for the propanal, butanal, pentanal, hexanal, heptanal,
octanal, nonanal and 2-t-hexenal CSA conjugates, respectively, at all
concentrations.

59

Table 3.5. Cross-reactivity of the monoclonal antibody to aldehyde modified chicken
serum albumin (CSA) by indirect competitive ELISA.

 

 

Aldehyde Conjugate Cross-
Structure Reactivityl
(%)
Hexanal CSA-NH(CH2),C H 3 100
Heptanal CSA-NH(CH2),,CH3 76.7
2-t-hexenal CSA-NHCHZCH=CH(CH3)2CH3 45.0
Pentanal CSA-NH(CH2),CH3 37.9

 

Defined as the concentration of hexanal-CSA required for 50% inhibition of binding

divided by the concentration of aldehyde-CSA required for 50% inhibition of binding
multiplied by 100.

60

by TNBS assay, indicating the low cross-reactivity of this monoclonal antibody to the 2-
t-hexenal conjugate may be due to lower modification.

Two compounds, 2-methylbutanal and 2-methylpentanal, are Strecker degradation
products. The Strecker degradation pathway involves the interaction of a-dicarbonyl
compounds with (it-amino acids resulting in volatile aldehydes, pyrazine and sugar
fragments. This reaction often occurs during cooking of meat. These compounds, as well
as a structurally related compound, 3-methylbutanal, and a 7 carbon ketone (Z-heptanone)
were also checked for cross-reactivity. There was less than 5% binding inhibition at 100
ng/mL concentration of 2-methylbutanal and 3-methylbutanal conjugates, and only
slightly more binding inhibition of the 2-methylpentanal conjugate (9% inhibition of
binding at 100 ng/mL concentration). The low binding inhibition indicates that
conjugates made with branched aldehydes were not recognized by the antibody. The
antibody showed no recognition of the 2-heptanone conjugate (95.9% modified).

CSA was also modified by n- and 2-hexanol. The TNBS assay showed the lysine
residues of the n-hexanol-CSA and 2-hexanol-CSA conjugates to be 78.0% and 77.0%
modified, respectively. Although it was theorized that the alcohols might be reduced to
aldehydes during the conjugation procedure, there was less than 8% binding inhibition at
100 ng alcohol/mL. The reason may be due to using sodium cyanoborohydride as a
reducing agent. Sodium cyanoborohydride is a weak reducing agent that readily reduces
Schiff bases, but does not reduce aldehydes (Jentoft and Dearbom, 1979). Sodium

cyanoborohydride, therefore, may not be strong enough to reduce the alcohols to

aldehydes.

61

These results indicate that the antibody is most specific to the straight chain, 6
carbon aldehyde, when conjugated to CSA. These results are similar to results obtained
by Smith (1997), where the cross-reactivity of the polyclonal antibody was found to
increase as the carbon number in the aliphatic aldehyde approached six. Although the
polyclonal antibody did cross-react with the branched aldehyde and ketone conjugates,
the amounts were small (less than 1.1%). The results indicated that the polyclonal
antibody was also most specific to 6 carbon aliphatic aldehydes, conjugated to CSA.

The next part of the study was designed to determine if the monoclonal antibody
could recognize hexanal when conjugated to amino acids. The proposed epitope
recognized by the monoclonal antibody is the lysine-hexanal complex, based on previous
work by Smith (1997) and that this antibody does not recognize free hexanal. Hexanal
was conjugated to lysine, glycine and arginine, as well as s-amino caproic acid, a lysine
derivative, to see what functional groups were required for recognition by the antibody.
The TNBS assay showed the available reactive sites on the lysine. glycine. arginine and
e-amino caproic conjugates to be 97.6%, 97.4%, 97.5% and 97.4% modified,
respectively, by hexanal. There was less than 3% binding inhibition at 2 mM amino acid
concentration with the hexanal modified glycine, arginine and the unmodified amino
acids. The antibody showed the highest recognition of the hexanal modified e-amino
caproic acid. At the 2 mM concentration, the e-amino caproic acid conjugate caused a
46% binding inhibition while the hexanal modified lysine conjugate showed 17% binding
inhibition at the same concentration (Figure 3.8.). This was as expected since e-amino

caproic acid has only the 8 amino terminal group available to bind hexanal and is

Inhibition (%)

Figure 3.8.

62

 

45 r +Lys-Hex
4o +Gly-Hex
35 +Arg-ch
30 +AC-Hex
25

20 r

 

 

 

 

0 0.5 l 1.5 2

Concentration of Amino Acid (mM)

Specificity of the monoclonal antibody to hexanal modified amino acids
was determined by indirect competitive ELISA. Each data point represents
4 replications. Microtiter wells were coated with 100 pL of hex-CSA (1
pg protein/mL). A 1/1000 dilution of antibody (1.6 pg/mL) was incubated
with 0, 0.05, 0.1, 0.5, 1 and 2 mM amino acid. The standard error of the
means, at all concentrations, were 2.31, 1.82, 1.77 and 5.51 for hexanal
modified lysine, glycine, arginine and amino caproic acid, respectively, at
all concentrations.

63

representative of the lysine side chains found in protein. The free lysine has both the or
and 8 amino groups available for binding to hexanal and may form a conjugate the
antibody cannot recognize. It is possible that the antibody does not recognize hexanal
bound to only the on amino group of lysine or if hexanal is bound to both the or and 8
amino groups.

Our results are similar to those of other researchers. Smith (1997) also found that
the s-amino caproic acid conjugate was a better competitor in the polyclonal ELISA for
hexanal than the lysine conjugate was. Thiele et a1. (1997) characterized a monoclonal
antibody that recognized acetaldehyde adducts formed with N—ethyl lysine. The cross-
reactivity of the antibody with acetaldehyde adducts formed with other amino acids
(arginine, ethylamine and lysine) was determined; however, this antibody was found to be
specific to only the acetaldehyde adduct formed with N—ethyl lysine. The antibody used in
our study also appears to be specific to only one adduct, the hexanal modified lysines
within a protein chain.

These results indicate that this monoclonal antibody was most specific to the
hexanal modified lysine side chain of proteins. The ELISA developed using this antibody
was sensitive to lng hexanal/mL. Based on its specificity and sensitivity, this ELISA
shows promise for being useful in actual systems. Although the ELISA will detect some
pentanal and heptanal, both these compounds are lipid oxidation products. The ELISA

could be used to detect not just hexanal, but to monitor lipid oxidation by measuring all

three products.

Chapter 4

VERIFICATION OF A MONOCLONAL ANTIBODY BASED ELISA FOR
MONITORING LIPID OXIDATION BY CORRELATION TO GAS

CHROMATOGRAPHY

4.1. Abstract

An indirect competitive (IC) ELISA for monitoring lipid oxidation was compared
with a commonly used gas chromatography (GC) method by measuring the hexanal
concentration of chicken serum albumin (CSA) that was differentially modified with
hexanal. The conjugates were tested in the ELISA at 3 dilutions (1/500, 1/1000 and
1/2000) and all 3 dilutions correlated well with the GC method (r = 0.975, 0.952 and
0.978, respectively). Native and hexanal modified proteins extracted from raw chicken
breast were separated by SDS-PAGE and used in a western blot. The antibody recognized
only the hexanal modified proteins. A frozen storage study was conducted using raw,
ground chicken meat. The hexanal concentration, as measured by IC-ELISA, peaked at 3
months with 3.3 pg hexanal/g breast meat and 4.02 pg hexanal/g thigh meat. Although
hexanal concentration by GC increased, the concentration was still low even after 9
months of storage. The breast contained only 0.42 pg hexanal/g meat and the thigh

contained 2.05 pg hexanal/g meat. A refrigerated storage study was conducted using

64

65

ground, cooked chicken. An alkaline protease was used as an aid in solublizing the
cooked chicken proteins. The maximum hexanal concentration, by GC, was 12.41 pg/g of
thigh meat after 60 h storage and 7.06 pg/g of breast meat after 3 days storage. Using the
lC-ELISA, the concentration of hexanal in breast meat did not change, whereas the
concentration of the thigh reached a maximum of 1.41 pg hexanal/g meat after 3 days.
Although the study with the differentially modified protein showed the IC-ELISA and
GC to correlate well, the storage studies with the chicken muscle were inconclusive.
Further studies of the reactions occurring between hexanal and meat proteins need to be

conducted so the IC-ELISA may be further optimized for use with a meat system.

4.2. Introduction

Oxidation plays a major role in consumer acceptance of muscle foods. Oxidation
can affect a range of meat components including proteins, lipids, vitamins. carbohydrates
and pigments (Kanner. 1994) resulting in changes in color, flavor and nutritive value of
the meat (Love and Pearson. 1971). Intrinsic factors can have an effect on the extent of
oxidation. Rhee et a1. (1996) found the higher the concentration of heme iron and
unsaturated fatty acids in a meat cut, the more likely the meat was to oxidize. Extrinsic
factors, such as cooking or storage temperature, can also influence oxidation. Cooked
meats develop rancid odors and flavors faster than raw meat (Ladikos and Lougovois,
1990)

The 2-thiobarbituric acid (TBA) is the most commonly used method for

monitoring lipid oxidation in foods. Two moles of TBA react with one mole of

66

malonaldehyde (MA) to form a pink complex that is measured spectrophotometrically at
532 nm (Botsoglou et al., 1994). TBA results are expressed as milligrams of MA per
kilogram of sample (Melton, 1983). The TBA test is now often referred to as the
thiobarbituric acid reactive substances (TBARS) test because it lacks specificity to MA.
As a result of this lack in specificity, TBARS has been widely criticized. Gray and
Monahan (1992) suggest using the TBARS test, but to keep in mind its limitations and
use a second method, such as hexanal concentration, for monitoring lipid oxidation.

Hexanal is a six carbon aldehyde that is the major secondary product formed
when linoleic acid is oxidized (Gray and Monahan, 1992). Hexanal concentration is
generally measured using gas chromatography (GC). The most commonly used GC
method is head space (HS) concentration (Reineccius, 1996). This method involves
purging the sample with an inert gas and trapping the volatiles. The volatiles are removed
from the trap by heating (thermal desorption) or by using a solvent (chemical desorption)
prior to injection into the GC. Hexanal concentration, as measured by HS-GC, has been
shown to correlate with sensory scores and TBARS numbers for cooked pork (Shahidi et
al., 1987), cooked beef (St. Angelo et al., 1987) and restructured chicken nuggets (Lai et
aL,1995)

Although GC is widely used, this method has several drawbacks including the
need for dedicated equipment, inconvenience of analyzing only one sample at a time and
possible peroxide decomposition to additional volatiles from the heat used to drive
volatiles onto the column (Ajuyah et al., 1993). An alternative to GC is an enzyme linked
immunosorbent assay (ELISA) for determining hexanal concentration. A monoclonal

antibody against hexanal was developed and an ELISA devised (Chapter 3, section

67

3.3.5.). The antibody was characterized to determine the usefulness of the ELISA. The
antibodies recognized hexanal conjugated to bovine serum albumin (BSA), chicken
serum albumin (CSA) and keyhole limpet hemocyanin (KLH) but did not recognize free
hexanal, native proteins or free amino acids. The antibody cross-reacted with pentanal,
heptanal and 2-t-hexenal (37.9%, 76.6% and 45.0%, respectively) conjugated to CSA, as
well as reacting strongly with hexanal modified amino caproic acid. Based on these
results, we proposed that the antibody recognized the hexanal-lysine complex formed via
Shiff base reactions.

Although the antibody cross-reacted with pentanal, heptanal and 2-t-hexenal
conjugates, as well as recognizing hexanal conjugates, this should not interference with
the usefulness of the ELISA. Pentanal and heptanal have much lower odor thresholds
than hexanal (Brewer and Vega, 1995), and are also important contributors to oxidized
flavors. The ELISA could be used to monitor lipid oxidation by measuring the
concentration of not just hexanal, but also pentanal and heptanal.

Another indirect ELISA for monitoring hexanal concentration, using this same
monoclonal antibody, was devised by Smith (1997). The ELISA was used, along with a
GC method and a TBARS assay, to monitor lipid oxidation in a model meat system.
Hexanal concentration measured by that monoclonal ELISA was found to correlate with
hexanal concentration by GC and TBARS numbers (r = 0.81 and 0.77, respectively).

The purpose of this study was to determine if hexanal concentrations quantified
by the monoclonal antibody based ELISA devised in the previous chapter would
correlate with hexanal concentrations measured by a common GC method. Chicken

serum albumin was modified with different amounts of hexanal and the hexanal

68

concentration of the conjugates, measured by ELISA and GC, was correlated.
Extraction procedures for raw and cooked meat were devised. Storage studies were then

designed to see if the hexanal concentration in an actual meat system, measured by GC

and ELISA, could be correlated.

4.3. Methods and Materials
4.3.1. Materials

The goat anti-mouse immunoglobulin G conjugated to horseradish peroxidase
(lgG-HRP) was purchased from Cappel Laboratories (West Chester, PA). The 60/80
mesh Tenax-TA was purchased from Alltech Associates, Inc. (Deerfield, IL) and the
silane treated glass wool was from Supelco, Inc. (Bellefonte, PA). The 10.2 cm long
metal traps were from MSU Biochemistry Shop (East Lansing, MI). The solvent, 2-
methyl butane, was purchased from Aldrich Chemical Co., Inc. (Milwaukee. WI). Whirl
Pak bags were from Nasco (Fort Atkinson, WI) and resealable storage bags were from
Gorden Food Service (Grand Rapids, MI). Enzeco® Alkaline Protease-L FG ( minimum
enzyme activity was 560,000 Delft units per gram, standardized with glycerol) was from
Enzyme Development Corporation (New York, NY). Chicken was purchased from local

retailers. All other chemicals were purchased from Sigma Chemical Co. (St. Louis, MO).

4.3.2. Hexanal Modification of CSA
Chicken serum albumin (CSA) was modified with different amounts of hexanal as

described in Chapter 3, section 3.3.2., except that NaCNBH3 was not used when

69

preparing samples for GC analysis. Briefly, 0.62, 1.23, 2.46, 6.15, 12.3, 18.5, 24.6, 37.0
or 49.2 pL of hexanal were added to 8 mL of CSA (12.5 mg/mL) diluted in phosphate
buffered saline (PBS; 0.1M sodium chloride, 0.01M sodium phosphate, pH 7.4). The
volume was brought up to 9 mL with PBS (pH 7.4). Half of this solution (4.5 mL) was
mixed with 0.5 mL of PBS (pH 7.4) and used for gas chromatography. The final reactant
concentrations were 10 mg/mL of protein and 0.5, 1, 2, 5, 10, 15, 20, 30 and 40 mM of
hexanal. The other half (4.5 mL) of the solution was prepared for ELISA by adding 0.5
mL of sodium cyanoborohydride, diluted in 0.1N NaOH, to a final concentration of 20
mM sodium cyanoborohydride. The final reactant concentrations of the protein and
hexanal were the same as the GC samples. The reaction was allowed to proceed
overnight. The percentage of protein modification of the conjugates used in the ELISA

was determined using a trinitrobenzenesulfonic acid (TNBS) assay (Habeeb. 1966) as

modified by Smith (1997).

4.3.3. Hexanal Concentration by CI-ELISA

The hexanal concentration of the conjugates was determined by indirect
competitive (IC) ELISA as described in Chapter 3, section 3.3.5. A standard curve was
generated using hexanal-CSA at concentrations of 0 to 500 ng hexanal/mL. The
differentially modified protein conjugates were diluted 1/2000, 1/1000 and 1/500 in PBS
(7.2) prior to ELISA analysis. These dilutions were equivalent to 5, 10 and 20 pg
protein/mL, respectively. Hexanal concentration was expressed as milligram of hexanal

per gram of protein. Six replicates were performed for each concentration.

70

4.3.4. Isolation and Concentration of Volatiles for Gas Chromatography

Prior to analysis by GC, the volatiles present in the sample, including hexanal,
were removed using a purge and trap apparatus. Volatiles from the samples were
absorbed onto Tenax and desorbed using 2-methylbutane following the methods of Liu
et a1. (1992) and Koelsch et a1. (1991) as modified by Smith (1997). Briefly, the 10.2 cm
long metal traps were packed with 200 mg of 60/80 mesh Tenax-TA and the ends
plugged with silane treated glass wool. The traps were conditioned for 24 h with nitrogen
gas at a flow rate of 25 mL/min at 150°C. After cooling, the traps were attached to a
purge and trap apparatus. For the determination of hexanal concentration in the pure
protein system (CSA), 1 mL of the solution and 50 mL of water were added to a round
bottomed, 500 mL flask. In the storage studies, 10 g of chicken and 50 mL of water were
used. The flask was then heated at 60°C for 90 min while sparging with a 30 mL/min
stream of nitrogen. To remove the volatiles from the Tenax, 2-methylbutane was added
dropwise and eluted with centrifugation at 80 x g for 5 min intervals until 1.5 to 2 mL

eluent was collected. The extract was then concentrated under nitrogen to 0.5 mL and

stored at -20°C.

4.3.5. Hexanal Concentration by Gas Chromatography

Hexanal concentration was determined using a Hewlett-Packard 5890A gas
chromatograph with a flame ionization detector. The volatiles were separated on a DB-
225 fused silica capillary column (30m x 0.25 mm i.d., 0.25 pm film thickness; J&W

Scientific Inc., Rancho Cordova, CA). Samples (5 pL) were injected and run according to

71

the program used by Smith (1997) except that the helium flow rate was increased to 50
mL/min and a split ratio of 66 was used. Hewlett-Packard 3365 Series II Chem Station
Software (Minneapolis, MN) was used for peak integration. Peak retention times of
hexanal, pentanal and heptanal references were used to identify elution peaks. The results
were calculated as milligrams of hexanal per gram of CSA. Each concentration of

differentially modified conjugate was tested in triplicate.

4.3.6. Statistical Analysis
Correlations between the hexanal concentration, measured by GC and each

protein concentration used in the ELISA, were established using linear regression

analysis (SAS, SAS Institute. Inc., Cary, NC).

4.3.7. Western Blot of Hexanal Modified Salt Soluble Proteins

Boneless, skinless chicken breast meat, purchased from a local retail store, was
ground twice with a Hobart Kitchen-Aid food grinder (Model KF-A, Troy, OH) using a 4
mm plate. The ground muscle was then homogenized with 3 volumes of 0.6 M NaCl; 50
mM sodium phosphate buffer using a Polytron homogenizer (Model PT 10/35;
Brinkmann Instruments Co., Westbury, NY). After straining the mixture through two
layers of cheesecloth, the proteins were modified with hexanal, using the method
described in Chapter 3, section 3.3.2. The final reactant concentrations were 200 mM
hexanal, 15 mg/mL of protein and 20 mM sodium cyanoborohydride. Both modified and
unmodified chicken proteins were separated by sodium dodecyl sulfate polyacrylamide

gel electrophoresis (SDS-PAGE) (Laemmli, 1970) using a Mini-Protean II

 

72

electrophoresis unit (Bio-Rad Laboratories, Richmond, CA). The stacking and separating
gels were 4 and 12% acrylamide, respectively, with 45 pg of protein loaded in each lane.
Protein bands were stained with Coomassie Brillant Blue R 250. The western blot was
performed as described by Wang et a1. (1992) except casein was used in place of BSA for
blocking and goat anti-mouse IgG peroxidase conjugate was used in place of the goat
anti-rabbit conjugate. Briefly, the proteins were transferred from an unstained SDS-
PAGE gel to a nitrocellulose membrane using a Mini Trans-Blot unit (Bio-Rad). After
transferring, the membrane was washed and blocked with 3% casein in phosphate buffer
saline (PBS; pH 7.2). After blocking, the membrane was again washed and monoclonal
antibody against hexanal (1/1000) was added and allowed to incubate for 30 min. The
unbound antibody was removed by washing and goat anti-mouse IgG peroxidase
(1/2000) was added and again incubated for 30 min. The membrane was washed a final
time and the bound peroxidase determined with 2,2’ azino-bis(3-ethylbenz-thiazoline-6-

sulfonic acid) substrate. The staining was stopped with water.

4.3.8. Proximate Analysis of Chicken
Protein, moisture and fat content of the raw and cooked chicken used in the

storage studies was determined in triplicate following AOAC methods (1990) 981.10,

950.46 B and 991.36, respectively.

73

4.3.9. Frozen Storage Study (Preliminary Experiments)

Boneless, skinless chicken thighs and breasts were purchased from a local retail
store. The muscle was ground twice with a Hobart Kitchen-Aid food grinder (Model KF-
A, Troy, OH) using a 4 mm plate. Salt (1% of total meat weight) was dissolved in 10%
added water, mixed into the ground muscle and ground twice. Meat (25 g) was placed in
24 oz. Whirl Pak bags and stored at -20°C for up to 9 months. The zero time samples
were tested prior to freezing. Six bags (3 bags of each muscle) were removed at intervals
and allowed to thaw at 4°C overnight. Meat from each bag was divided into two 10 g
portions for testing by ELISA and GC.

The hexanal concentration was determined by GC as described above. Meat for
ELISA analysis was extracted in 90 mL of 0.6 M NaCl buffer with 50 mM sodium borate
(pH 9) by stirring for 1 h on a stir plate. The samples were reduced with sodium
cyanoborohydride (20 mM final concentration) dissolved in 0.1N NaOH and left
overnight at ambient temperature. The mixture was strained through two layers of
cheesecloth, diluted 1:80 in PBS (pH 7.2) and tested by the ELISA method described
above. The protein concentration of the meat extracts used in the ELISA was about 250
pg protein/mL. Hexanal concentration by both ELISA and GC was expressed as

micrograms hexanal per gram of meat.

4.3.10. Optimization of Extraction Procedure for Cooked Thigh Meat
Boneless chicken thighs were ground once with a Hobart Kitchen-Aid food

grinder (Model KF-A, Troy, OH) using a 4 mm plate. After grinding, 30 g meat was

74

placed into a 50 mL centrifuge tube and held in a circulating water bath (Model 1268—52,
Cole-Parmer, Chicago, IL) at 73°C. A thermocouple (RTD, 1.6 mm diameter : 0.5°C)
connected to a Solomat MPM 2000 Modumeter (Stanford, CT) was placed in one tube
containing meat. When the temperature of the meat reached 71°C, the tubes were
removed and placed in ice water for 15 min. After cooking, the meat was reground twice.

An enzyme solution was prepared by diluting 25 pL of the alkaline protease in 50
mL of the extraction buffer (0.6 M sodium chloride, 50 mM sodium borate, pH 9.0). The
enzyme concentration necessary to maximize protein solubility was determined by adding
2.4, 4.8, 24 or 48 pL of the enzyme solution (which corresponded to 0.05, 0.1, 0.5 and
1% (w/w) of the protein concentration) to 3 g cooked chicken and 27 mL extraction
buffer (pH 9.0) in a 50 mL centrifuge tube. After shaking for 30 s, the tubes were placed
in a water bath at 50°C for 2 h with shaking. The solutions were filtered through two
layers of cheesecloth. The filtrate was considered the soluble protein solution. The protein
concentration of the solution was then determined using AOAC method (1990) 981.10.
Enzyme concentration was plotted against the percent of soluble protein in the extracted
solution to determine the enzyme concentration that extracted the most protein.

Extraction time was determined using 3 g cooked chicken, 27 mL extraction
buffer and an enzyme concentration equal to 0.5% (w/w) of the total protein. Centrifuge
tubes were placed in the water bath at 50°C and removed at 0.25, 0.5, 1, 2 or 4 h. After
filtering through cheesecloth, the soluble protein solution concentration was again

determined using AOAC method (1990) 981.10. Extraction time was plotted against the

75

percent protein concentration to determine which extraction time extracted the most

protein.

4.3.11. Cooked Storage Study (Preliminary Experiments)

Boneless, skinless chicken thighs and breasts were ground twice with a Hobart
Kitchen-Aid food grinder using a 4 mm plate. Salt (1% of total meat weight) was
dissolved in 10% added water and mixed in with the ground meat prior to grinding a
second time. Three replicate batches of ground thigh meat and three batches of ground
breast meat, were cooked in a circulating water bath as described in the previous section.
Prior to cooking, a raw sample was removed and tested for hexanal. The zero time sample
was removed immediately after cooking and tested prior to storage at 4°C. The remaining
cooked meat was placed in resealable storage bags and stored at 4°C. Each replicate was
tested for hexanal by IC-ELISA and GC at 12 h intervals for the thigh and 24 h intervals
for the breast.

The samples were extracted for determination by IC-ELISA by placing 3 g cooked
meat in a 50 mL test tube with 26 mL of extraction buffer (0.6M NaCl. 50 mM sodium
borate; pH 9.0), 1 mL of 60 mM NaCNBH3 diluted in 0.1 N NaOH and 24 pL of enzyme
solution (prepared by diluting 25 pL of the enzyme in 50 mL of the extraction buffer).
The tube was placed in a shaking water bath at 50°C for 3 h. After extraction, the solution
was filtered through 2 layers of cheesecloth and diluted 1:10 with PBS (pH 7.2) prior to

being tested in the ELISA. Protein concentration used in the ELISA, described above,

 

76

was about 2 mg protein/mL. Hexanal concentration was expressed as micrograms of

hexanal per gram of meat.

4.3.12. TBA-RS Assay

Thiobarbituric acid-reactive substances of the cooked meat during storage were
determined by the method of Buege and Aust (1978). After the protein was extracted
from the cooked chicken and filtered through cheesecloth, 2 mL of the extract was
transferred to a 15 mL centrifuge tube and 4 mL of the TBA reagent was added. The TBA
reagent was comprised of 15% w/v TBA, 0.375% w/v trichloroacetic acid and 0.25N
hydrochloric acid. The samples were heated in boiling water for 15 min then centrifuged
at 1000 x g for 15 min. The absorbance of the supernatant was read at 535 nm and the
results expressed as milligram of malonaldehyde per kilogram of meat using a molar

extinction coefficient of 1.56 x 10'5 L x mole‘l x cm". Each sample was tested in

triplicate.

4.4. Results and Discussion
4.4.1. Differential Modification of CSA

The binding inhibition produced by the differentially modified hex-CSA
conjugates, when tested in the ELISA at the three dilutions, were graphed against the
concentration of hexanal added to CSA (Figure 4.1). The range of binding inhibition of
the conjugates at the 1/2000 dilution (5 pg CSA/mL) was the largest, ranging from 1% to

70%. The range of inhibition for the 1/1000 dilution (10 pg CSA/mL) was smaller,

 

77

ranging from 26% to 80%. The smaller range of binding inhibition may be a result of
increased protein concentration which gives an increase in the concentration of hexanal
that is added to the wells of the ELISA, resulting in increased inhibition. At the 1/500
dilution (20 pg/mL), the binding inhibition ranged from 12% to 79%. Although the
protein concentration is doubled, the binding inhibition at the 1/500 dilution is similar to
the 1/1000 dilution when CSA was modified with greater than 20 mg of hexanal/g CSA.
This may be due to the protein concentration becoming too high and causing interference.
Although the protein is denatured during modification, some cross-linking may still be
occurring at higher protein concentrations and blocking the antibody from binding to the
hexanal.

Diluting the extract from a meat sample to a volume that would contain
approximately 10 pg protein/mL would be the recommended dilution to use in the ELISA
for storage studies measuring lower quantities of hexanal (less than 25 mg hexanal/g
protein). At this protein concentration, the ELISA has been shown to be more sensitive to
lower concentrations of hexanal, giving a higher binding inhibition at of CSA lower
concentrations conjugates. However, in studies measuring larger concentrations of
hexanal, dilutions yielding 5 pg protein/mL would be better, as it gives a wider range of
binding inhibition with changes in hexanal concentration. If the concentration of hexanal
is expected to exceed that used for this differential modification study, the use of a lower
protein concentration in the ELISA should be considered.

Similar results were found by Smith (1997) using ovalbumin as a carrier protein

in place of CSA. The study showed a large increase in binding inhibition, 11.4% to

60 ..

90
80 «
70 .
.3
5
a 50
D
E
3 40
C
"E
a 30
20
10
0
Figure 4.1.

78

 

 

 

 

 

/ 7" __
/
/

+1/2000 dilution
+111000 dilution
+1/500 dilution

0 50 100 150 200 250 300 350 400

Concentration of Hexanal Used to Prepare Conjugate (mglg protein)
Concentration of hexanal added to prepare chicken serum albumin conjugates

and the percent inhibition produced in the ELISA by three dilutions of the
conjugates. Conjugates were prepared by adding 0-400 mg hexanal/g protein,
then diluted 1/500, 1/1000 and 1/2000 prior to being added to the ELISA as a
competitor. The calculated protein concentration of each dilution was 5
pg/mL (1/2000 dilution), 10 pg/mL (1/1000 dilution) and 20 pg/mL (1/500
dilution). Each point represents three replicates.

79

91.7%, as the concentration of added hexanal was increased by 25 fold. The protein
concentration used in that study was 20 pg ovalbumin/mL.

The hexanal concentration of the differentially modified CSA conjugates was
measured by IC-ELISA and compared to the concentration measured by GC. Hexanal
concentration, measured by GC, ranged from 13.61 to 89.23 mg hexanal/g protein when
4.74 to 379.1 mg hexanal/g CSA was used for the modification. The hexanal
concentration, measured by the ELISA, was varied, depending on the dilution of the
conjugate (Figure 4.2.) The concentration ranged from 0.15 to 30.18 mg/g at the 1/2000
dilution. A 1/1000 dilution gave a similar range of 0.59 to 26.1 mg/g. At the lowest
dilution, the concentration ranged from 0.08 to 11.69 mg/g. Although the hexanal
concentration measured by GC was much higher than that by ELISA, statistical analysis
of the results showed good correlation between the hexanal concentration measured by
GC and the ELISA. A correlation of 0.98, 0.98 and 0.95 were calculated for the 1/2000,

1/1000 and 1/500 dilutions of the conjugates. respectively.

4.4.2 Electrophoresis and Western Blot of Salt Soluble Proteins

The hexanal modified salt soluble proteins (SSP) did not migrate as far through
the gel as the corresponding unmodified proteins, suggesting the modified proteins had
higher molecular weights than the native ones due to the addition of hexanal (Figure
4.3.). The antibody showed recognition of the modified chicken SSP, indicated by the
presence of bands on the western blot, but showed no recognition of the native chicken

proteins. It should also be noted that hexanal was conjugated to several chicken proteins.

80

 

100
+122000 ELISA dilution
so +1 :1000 ELISA dilution

+1500 ELISA dilution

 

Measured Hexanal Concentration (mg/g protein)

 

 

 

 

0 50 1 00 1 50 200 250 300 350 400

Concentration of Hexanal Used to Prepare Conjugate (mg/g protein)

Figure 4.2. Hexanal concentration of differentially modified protein conjugates as
measured by ELISA and gas chromatography (GC). The conjugates were
diluted 1/500, 1/1000 and 1/2000 prior to being added to the ELISA as a
competitor. The calculated protein concentration of each dilution is 5 pg/mL
(1/2000 dilution), 10 pg/mL (1/1000 dilution) and 20 pg/mL (1/500 dilution).
Each point represents 3 replicates.

 

81

 

 

Figure 4.3. Detection of hexanal modified and native chicken breast salt soluble proteins
by western blotting following sodium dodecyl sulfate polyacrylamide gel
electrophoresis (SDS-PAGE). The SDS-PAGE gel was stained with
Commassie Blue. The western blot was performed using monoclonal
antibodies against hexanal at a 1/1000 dilution. a: SDS-PAGE of hexanal
modified salt soluble proteins, b: SDS-PAGE of native proteins, c: western
blot of hexanal modified salt soluble proteins and d: western blot of native
proteins.

82

4.4.3. Frozen Storage Study

Chicken thigh used in this study contained 73.3 :0.69 % moisture, 5.29:0.1 % fat
and 17.4:1.21 % protein, whereas the breast contained 74.3:5.25 % moisture, 0.93:0.07
% fat and 21 6:13 % protein. The hexanal concentration of the thigh, as measured by IC-
ELISA, increased rapidly compared to that measured by GC (Figure 4.4). The
concentration measured by the ELISA peaked at 4.02 pg hexanal/ g meat afier 3 months
storage. At 9 months, the hexanal concentration by GC was 2.05 pg hexanan meat.
Although the hexanal concentration of the breast by ELISA had increased to 3.3 pg
hexanal/g meat after 3 months storage, the hexanal concentration by GC had only
increased to 0.42 pg hexanal / g meat after 9 months storage (Appendix A).

The difference between the hexanal concentration measured by IC-ELISA and GC
may be due to changes in free and bound hexanal within the meat throughout the study.
Hexanal will readily bind to protein (Gardner, 1979; Gutheil and Bailey, 1992). Since the
ELISA only detects hexanal once it is bound to protein, the ELISA might initially give
higher values for hexanal than GC as protein binding sites for hexanal should be readily
available. Although the meat extract used in GC is heated to remove the volatiles, this
temperature will remove some bound volatiles, but not all of them (Reineccius, 1996). As
the oxidation reaction progresses, additional hexanal is produced, thereby resulting in
more free hexanal as the protein binding sites are saturated. Since GC measures free
hexanal, this would result in an increased hexanal concentration at longer storage times.

Other aldehdyes will also bind to protein. Malonaldehdye has been shown to bind to

83

4.5

 

+ GC
3.5

+8.8A

2.5

1.5

Hexanal Concentration (ug hexanal/g meat)

 

 

 

Storage Time (months)

Figure 4.4. The hexanal concentration of raw chicken thigh stored at -20°C was
determined by indirect competitive ELISA and gas chromatography (GC).

84

lysine (Crawford et a1. 1967; Buttkus, 1967). It is possible that additional aldehdyes, such
as malonaldehyde, are being produced over time and are displacing the hexanal, resulting
in increased free and increased GC values. The ELISA values would decrease due to lack
of available binding sites on the protein.

The results of another study also showed low lipid oxidation in frozen, raw
chicken. Rhee et a1. (1996) measured TBA numbers for raw chicken thigh and breast
stored at -20°C for up to 120 days (just over 5 months). The TBA numbers for both the
breast and thigh increased a small amount, 0.1 mg MA/kg meat, indicating very little
lipid oxidation. They hypothesized that, although chicken muscle is high in unsaturated
fatty acids, the low amount of oxidation in the breast may be due to a low concentration
of metmyoglobin, an initiator of oxidation when activated by H202. The low oxidation of
the thigh may be due to a high level of catalase activity, which limits H202 production,
even though the thigh is more pigmented than the breast and contains more
metmyoglobin.

The study by Rhee et a1. (1996) also showed the oxidation of raw. frozen pork and
beef. The TBA numbers for pork longissimus dorsi increased, from 0.05 to 1 mg MA/kg
sample, after 75 days storage at 4°C. The TBA numbers for beef longissimus dorsi also
increased, from 0.6 to 1.4 mg MA/kg sample, after 150 days storage at 4°C. Either pork
or beef might be better meats to use for another frozen storage study to determine if the

hexanal concentration measured by ELISA and GC can be correlated.

85

4.4.4. Enzyme Conditions for Cooked Meat Extraction

The concentration of alkaline protease and extraction time needed to maximize
protein solubility of the cooked chicken proteins was determined. The soluble protein
concentration of the cooked meat extract was 0.88% without enzyme and increased to
1.69% when enzyme concentration was increased to 0.5% of the total protein
concentration and incubated at 50° C for 2 h (Table 2.1.). Total protein concentration of
chicken thigh is 20.08% (USDA, 1978). The meat is diluted 1/10 during the extraction
process, giving a theoretical maximum protein concentration of 2%. Soluble protein
concentration increased only slightly (a 0.12% increase) when enzyme concentration was
increased to 1% of total protein. Although the enzyme is further diluted (1/50) prior to
ELISA analysis, thereby decreasing the concentration, and would have been at a pH and
temperature outside of its optimal range (Enzyme Development Corporation information
sheet), the 0.5% enzyme concentration was used to be certain there would be no
proteolysis of the antibody by the enzyme.

Using a 0.5% (wt/vol) concentration of the enzyme, the effect of extraction time
at 50°C on soluble protein concentration was determined. The soluble protein
concentration increased from 0.86 to 1.99% during the 0.25 to 4 h incubation (Table 2.2.).
A 3 h extraction time was chosen so that the ELISA could be completed within a period
of time comparable to the GC procedure. The final extraction conditions were set at 50°C

for 3 h using an enzyme concentration of 0.5% of the total protein.

 

86

Table 4. 1. Effect of enzyme concentration on the extraction of salt soluble protein
from cooked chicken thigh at 50°C for 2 h.

 

 

Enzyme Concentration Soluble Protein

(%)| (%)Z

0 0.88 : 0.09
0.05 1.02 : 0.03
0.1 1.23 : 0.05
0.2 1.36 : 0.18
0.5 1.64 : 0.11
1.0 1.79 : 0.04

 

Enzyme concentration was expressed as % of total protein in chicken thigh.
Theoretical maximum protein concentration of the extract was 2%.

87

Table 4.2. Effect of extraction time on the salt soluble protein concentration of cooked
chicken thigh at 0.5% enzyme concentration and 50°C.

 

 

Time Soluble Protein
0!) (%)'
0.25 0.86 : 0.05
0.5 1.14 : 0.04
1.0 1.42 : 0.13
1.5 1.55 : 0.06
2.0 1.68 : 0.13
4.0 1.99 : 0.13

 

' Theoretical maximum protein concentration of the extract is 2%.
Enzyme concentration was expressed as % of total in chicken thigh.

88

4.4.5. Cooked Storage Study
Prior to cooking, the chicken breast in this study contained 70.7: 1.67%
moisture, 1.34: .03% fat and 21.2: 0.35% protein. Although the breast meat showed an
increase in hexanal concentration during storage when measured by GC, it did not show
an increase by ELISA. The hexanal concentrations obtained for the breast by GC and
ELISA, as well as the TBARS numbers are shown in Appendix A.
Chicken thigh in this study contained 73.3:0.69% moisture, 5.29: 0.1% fat and

16.8: 1.21% protein. The hexanal concentration of the thigh, measured by GC, increased
rapidly from 2.48 to 12.41 pg hexanal/g meat over 60 h of storage. The hexanal values by
GC for the thigh in this study are comparable to the values reported by other researchers.
Dupuy et al. (1987) reported the hexanal concentration of dark meat from roast turkey to
be 11.7 pg hexanal/g meat after 2 days storage at 4°C. The hexanal concentration of
uncured, cooked chicken after 24 h storage at 4°C was 9.84 pg hexanal/g meat
(Ramarathnam et al., 1991).

The TBARS value also showed an increase similar to the GC results, from 1.17 to
13.01 mg MA/kg meat after 60 h storage at 4°C. These values are also comparable to
values reported by other researchers. The TBA number of cooked chicken thigh stored for
6 days at 4°C was 14.5 mg MA/kg sample (Rhee et al., 1996). The TBA number of raw
chicken thigh increased by a factor of 10 after 8 days storage at 4°C (Whang and Peng,
1987).

The hexanal concentration measured by ELISA increased from 0.71 pg hexanal/g

meat to 1.41 pg hexanal/g meat after 12 h of storage (Figure 4.7.). Hexanal concentration

 

89

by ELISA for the cooked chicken may be lower than GC due to the loss of free lysine
sidechains due to binding with other aldehydes and ketones formed during storage. The
TBARS numbers of the cooked thigh increased rapidly, a possible indication that
malonaldehdye was being formed. Crawford et al. (1967) found the 8 amino group of
lysine in bovine serum albumin was blocked after exposure to MA. Buttkus (1967) found
that, after exposure to MA, ninhydrin reactive amino groups in trout myosin were lost.
Ninhydrin will react only with free amino groups, especially the 8 amino of lysine.
Another study noted that as 2-thiobarbituric acid (TBA) values increased, free 8 amino
groups decreased (Kuusi et al., 1975). It is possible that malonaldehdye is binding to the
lysine side chains.

A possible solution to the problem of lysine not being available for binding to
hexanal would be to add in a pure protein. such as C SA, during extraction. Free hexanal
could then bind to the CSA and the hexanal concentration could be determined based on
the amount of CSA added. This could then be converted to the hexanal concentration in
the meat. Another possible source of lysine would be to add c-amino caproic acid. Since
the antibody did not cross-react 100% with e-amino caproic acid, the lower amount of
cross reactivity could be adjusted for when calculating the hexanal concentration.

Another possible reason for the lower hexanal concentration measured by the
ELISA may be not enough sodium cyanoborohydride to sufficiently reduce the double
bonds formed between hexanal and the meat proteins. Additional studies will need to be

conducted to determine if the concentration of sodium borohydride needs to be increased.

Concentration of hexanal (ug hex/g protein)

Figure 4.5. Hexanal

90

+GC
12 ~ -e— ELISA
+TBARS

Storage Time (h)

-10

Concentration of malonaldehyde
(mg MAI kg meat)

60

concentration and 2-thiobarbituric acid reactive substances

(TBARS) number of cooked chicken thigh meat stored at 4°C. Hexanal
concentration was determined by both ELISA and gas chromatography.

 

91

The results of the study with the differentially modified CSA suggest that the
ELISA could be used in place of the GC method. The correlation between GC and all
three dilutions of the conjugates was high. However, when the ELISA was used in actual
meat storage studies, the results were not conclusive. No correlation was found between
the two methods. Further studies are required to determine if the ELISA can be used as a

reliable indicator of lipid oxidation in meat systems.

Chapter 5

CONCLUSIONS

We produced highly sensitive monoclonal antibodies to hexanal modified bovine
serum albumin. Sodium cyanoborohydride was used as the reducing agent,
allowing us to reduce the Schiff bases formed between hexanal and lysine. The

antibody was found to recognize the hexanal-lysine complex.

We found three compounds, in addition to hexanal, that the antibody cross-reacted
with. The compounds were pentanal, 2-t-hexenal and heptanal, aliphatic
aldehydes with 5, 6 and 7 carbons. Although the monoclonal antibody cross-
reacted with these three compounds, the assay is still useful for monitoring lipid
oxidation, as both pentanal and heptanal are lipid oxidation products that
contribute to oxidized flavors. also, the concentrations of these three compounds
that have been measured in foods are insignificant compared to the levels of

hexanal commonly found.

We determined that the monoclonal antibody ELISA correlated well with a
commonly used GC method when using differentially modified chicken serum

albumin conjugates.

92

Chapter 6

FUTURE RESEARCH

Although the ELISA showed promise by correlating with GC using the
differentially modified conjugates, the results of the storage study were inconclusive.
More work needs to be done to better understand the binding of hexanal to chicken
proteins and the measurement of hexanal concentrations by ELISA.

A frozen storage study may be more useful than a refrigerated study as oxidation
occurs slower. By using a meat product, such as beef or pork, that oxidizes quickly at
frozen temperatures. the ELISA method could be perfected. After a better understanding
of the ELISA is determined, a refrigerated study could be undertaken. Further research is
needed to determine the availability of lysine during oxidation of chicken meat at 4° C.
The possibility of adding in a set concentration of protein needs to be examined as a
source of lysine. Once the ELISA method has been verified by GC for an actual meat
system, the ELISA could then be verified through sensory analysis.

In addition to detecting hexanal concentration in meats, the ELISA could be
examined for use in detection of hexanal concentration in non-meat products. Protein
would need to be added as a carrier if the product has a low protein concentration. If
additional protein is added for more lysine in the meat system, it could then be easily

used in another product.

93

APPENDIX A

APPENDIX A

Conversion Example

An example of the conversion of 1 pg protein/mL concentration of hex-CSA to hexanal
concentration is shown here.

1 pg protein x 46 moles lysine x 1 mole protein x 1 gram x l_0_° moles = 7.4 x 10“ pmoles lysine
mL 1 mole protein 66,300 gram 10° pg 1 mole mL

The lysine concentration is multiplied by 99.1% to account for the % modification. Since

one mole of lysine will bind one mole of hexanal, the moles are considered equivalent.

Therefore, there are 7.3x10"‘ pmoles of hexanal.

 

7.3x10‘4Lnole hexanal x 100 gram x 1 mole x E" ng = 73 ng hexanal
mL 1 mole 10° p mole 1 gram mL

94

APPENDIX B

APPENDIX B

Values Obtained for Oxidation of Breast

Frozen Study Results

Hexanal Concentration (pg /g meat)

 

 

 

Month GC ELISA
0 0 0.018
0.5 0 4.15
1 0 2.13
1.5 0 2.62
2 0 2.38
3 0 3.30
5 0 1.96
6 0 2.20
8 0.51 1.81
9 0.42 0.37
Cooked Study

Hexanal Concentration (112/le
Day GC ELISA TBARS (mg MA/kg sample)
0 1.92 0.43 3.31
1 5.25 0.46 5.48
2 6.01 0.24 6.62
3 7.06 0.03 6.89
4 5.84 0.64 8.27
5 4.1 0.29 7.55
7 1.82 0.07 7.79

95

LIST OF REFERENCES

LIST OF REFERENCES

AOAC. 1990. Meat and meat products. Ch. 39 in W, K.
Helrich (Ed.), p.931-948. Association of Official Analytical Chemists, Inc.,
Arlington, VA.

Abouzied, M.A., Asghar, A., Pearson, A.M., Gray, J .I., Miller, ER. and Pestka, J.J. 1990.
Monoclonal antibody-based enzyme-linked immunosorbent assay for Clg'Al6"
steroids in sera of boar pigs. J. Agric. Food Chem. 38:331-335.

Abouzied, M.A., Wang, C.H., Pestka, J.J. and Smith, D.M. 1993. Lactate dehydrogensae
as a safe endpoint cooking indicator in poultry breast rolls: Development of
monoclonal antibodies and application to sandwich enzyme-linked
immunosorbent assay (ELISA). J. Food Prot. 56: 120-124,129.

Ahn, D.U., Wolfe, EB. and Sim, J .S. 1993. The effect of free and bound iron on lipid
peroxidation in turkey meat. Poultry Sci. 72:209-215.

Ajuyah, A.O., F enton, T.W., Hardin, R.T. and Sim, J .S. 1993. Measuring lipid oxidation
volatiles in meats. J. Food Sci. 58:270-277.

Akamittath, J .G., Brekke, OJ. and Schanus. E.G. 1990. Lipid oxidation and color

stability in restructured meat systems during frozen storage. J. Food Sci. 55:1513-
1517.

Alaiz, M. and Giron, J. 1994. Modification of histidine residues in bovine serum albumin
by reaction with (E)-2-octenal. J. Agric. Food Chem. 42:2094-2098.

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.

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

96

97

Bett, K. L. and Boylston, T. D. 1992. Effect of storage on roasted peanut quality:
Descriptive sensory analysis and gas chromatographic techniques. Ch. 19 in Lipid.

WW, A..J St. Angelo (Ed.,) p. 322- 343. American Chemical
Society, Washington, DC.

Block, RI. 1956. MW, Charles C. Thomas, Springfield, IL.

Botsoglou, N.A., Fletouris, D.J., Papageorgiou, G.E., Vassilopoulos, V.N., Mantis, A.J.
and Trakatellis, A.G. 1994. Rapid, sensitive, and specific thiobarbituric acid
method for measuring lipid peroxidation in animal tissue, food, and feedstuff
samples. J. Agric. Food Chem. 42:1931-1937.

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

Buttkus, H. 1967. Reaction of myosin with malonaldehdye. J. Food Sci. 32:432-434.

Chan, H.W.S. 1987. The mechanism of autoxidation. In Wad
Lipids, H.W.S. Chan (Ed.), p.1-16. Academic Press, London, England.

Civille, G. V. and Dus, C. A. 1992. Sensory evaluation of lipid oxidation in foods. Ch. 15

ianpiaLQn‘datLanzLEQQd, A.J. St. Angelo (Ed.,) p. 279- 289. American
Chemical Society, Washington, DC.

Clark, B.R. And Engvall, E. 1980. Enzyme-linked immunosorbent assay (ELISA):

Theoretical and practical aspects. Ch. 8 in W, E.T. Maggio
(Ed.), p. 167-196. CRC Press, Boca Raton, FL.

Cohen, S..,A Meys, M. and Tarvin, TL. 1989. WM
AflancetfliechmauemLAminaAfldAnm Millipore Corp” Bedford MA

Crawford, D.L., Yu, TC. and Sinnhuber, RD. 1967. Reaction of malonaldehdye with
protein. J. Food Sci. 32:332-335.

Deshpande, S. S. 1996. Assay development, evaluation and validation. Ch. 9m Enzymg

W p 275- 359 Chapman
and Hall, New York, NY.

Donohue, T.M., Tuma, DJ. and Sorrell, M.F. 1983. Acetaldehyde adducts with proteins:

Binding of [”C] acetaldehyde to serum albumin. Arch. Biochem. Biophysics
220:239-246.

Dupuy, H.P., Bailey, M.E., St. Angelo, A.J., Vercellotti, JR and Legendre, MG. 1987.
Instrumental analysis of volatiles related to warmed-over flavor of cooked meats.

98

In W A.J- St. M8610 and
ME. Bailey, (Ed.), p.165-190. Academic Press, Orlando, FL.

Esterbauer, H. 1993. Cytotoxicity and genotoxicity of lipid oxidation products. Am. J.
Clin. Nutr. 57:779S-78SS.

Frankel, EN. 1991. Review- Recent adavancements in lipid oxidation. J. Sci. Food
Agric. 54:495-511.

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

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.

Galfre, G. and Milstein, C. 1981. Preparation of monoclonal antibodies: Strategies and
procedures. Methods Enzymol. 73:3-46.

Gardner, H.W. 1979. Lipid hydroperoxide reactivity with proteins and amino acids: A
review. J. Agric. Food Chem. 27:220-229.

Goding, J .W. 1980. Antibody production by hybridoma. J. Immunol. Methods. 39:285-
308.

Gray, J .1. Measurement of lipid oxidation: A review. 1978. J. Am. 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.

Gutheil, RA and Bailey, ME. 1992. A method for determining binding of hexanal by
myosin and actin using equilibrium headspace sampling gas chromatography. In

H! '1 ' en 0.1141 A. "u' D' ' uu't ' He- '1' ',G.
Charalambous (Ed.), p.783-815. Elsevier Science Publishers, Amsterdam,
Netherlands.

Gutteridge, 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.

99

Hall, G. 1987. Interactions between products of lipid oxidation and proteins. Food Sci.
Technol. Today 1:155-158.

Harlow, E. and Lane, D. 1988. WWW. Cold Spring Harbor
Laboratory, Cold Spring Harbor, NY.

Hebert, G.A., Pelham, PL. and Bittman, B. 1973. Determination of the optimal
ammonium sulfate concentration for the fraction of rabbit, sheep, horse and goat
antisera. Appl. Mirobiol. 25:26-36.

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.

Israel, Y., MacDonald, A., Niemela, 0., Zamel, D., Sharni, E., Zywulko, M., Klajner, F.
and Borgono, C. 1992. Hypersensitivity to acetaldehyde-protein adducts. Molec.
Pharmacol. 42:711-717.

Jentoft, N. and Dearbom, D.G. 1979. Labeling of proteins by reductive methylation using
sodium cyanoborohydride. J. Biol. Chem. 254:4359-4365.

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

Kaur, H. and Perkins, M. J. 1991. The free radical chemistry of food additives. Ch. 2 1n

Wm.01 Aruoma andB. Hallrwell (Ed) p. 17-35.
Taylor and Francis, Ltd, Bristol, PA.

Klassen, L.W., Tuma, D.J., Sorrell, M.F., McDonald, T.L., DeVasure, J .M. and Thiele,
G.M. 1994. Detection of reduced acetaldehyde protein adducts using a unique
monoclonal antibody. Alcoh. Clin. Exper. Research 1 8: 164-1 71 .

Koelsch, C. M., Downes, T. W. and Labuza, T. P. 1991. Hexanal formation via lipid

oxidation as a function of oxygen concentration: Measurement and kinetics. J.
Food Sci. 56:816-820, 834.

Kohler, G. and Milstein, C. 1975. Continuous cultures of fused cells secreting antibody of
predefined specificity. Nature 256:495-497.

Kosugi, H. and Kikugawa, K. 1986. Reaction of thiobarituric acid with saturated
aldehydes. Lipids 21:537-542.

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

100

Kuusi, T., Nikkila, O.E., Savolainen, K. 1975. Formation of malonaldehdye in frozen
Baltic herring and its influence on the changes in proteins. Z. Lebensm.-Unters.
Forsch. 159:285-286.

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

Laemmli, UK. 1970. Cleavage of structural proteins during the asembly of the head or
bacteriophage T4. Nature 227:680-685.

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.

Liu, H.F., Booren, A.M., Gray, J .I. and Crackel, R.L. 1992. Antioxidant efficacy of

oleoresin rosemary and sodium tripolyphosphate in restructured pork steaks. J.
Food Sci. 57:803-806.

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

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

Melton, S.L. 1983. Methodology for following lipid oxidation in muscle foods. Food
Technol. 42(7):105-111, 116.

Monahan, F .J ., Crackel, R.L., Gray, J.I., Buckley, DJ. and Morrissey, PA. 1993.

Catalysis of lipid oxidation in muscle model systems by haem and inorganic iron.
Meat Sci. 34:95-106.

Newburg, D.S. and Concon. J.M. 1980. Malonaldehyde concentrations in food are
affected by cooking conditions. J. Food Sci. 45: 1681-1687.

Pearson, A.M., Gray, J .I., Wolzak, A.M. and Horenstein, NA. 1983. Safety implications
of oxidized lipids in muscle foods. Food Technol. 37(7):]21-129.

Pestka, J .J ., Lee, Y.K. and Chu, F .S. 1982. Reactivity of aflatoxin 82, antibody with

aflatoxin Bl modified DNA and related metabolites. Appl. Environ. Microbiol.
44:1159-1165.

Pestka, J .J ., Abouzied, MA. and Sutikno. 1995. Immunological assays for mycotoxin
detection. Food Technol. 49(2): 120: 128.

101

Pikul, J ., Leszczynski, DE. and Kummerow, EA. 1983. Elimination of sample
autoxidation by butylated hydroxytoluene additions before thiobarituric acid assay
for malonaldehyde in fat from chicken meat. J. Agric. Food Chem. 31 :1338-1342.

Pikul, J., Leszczynski, D.E., Bechtel, P]. and Kummerow, FA. 1984. Effects of storage
and cooking on lipid oxidation in chicken meat. J. Food Sci. 49:838-843.

Pikul, J., Leszczynski, DE. and Kummerow, RA. 1989. Evaluation of three modified
TBA methods for measuring lipid oxidation in chicken meat. J. Agric. Food
Chem. 37:1309-1313.

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

Ramarathnam, 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: 1839-1 847.

Reineccius, GA. 1996. Instrumental means of monitoring the flavor quality of foods. Ch.

22 in W. T.C. Lee and HJ. Kim
(Ed.) p.241-252. American Chemical Society, Washington, DC.

Rhee, KS. 1978. Minimization of further lipid peroxidation in the distillation 2-
thiobarituric acid test of fish and meat. J. Food Sci. 43:1776-1779.

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

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

Salih, A.M., Smith, D.M., Price, J .F. and Dawson, LE. 1987. Modified extraction 2-

thiobarituric acid method for measuring lipid oxidation in poultry. Poult. Sci.
66:1483-1488.

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

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

 

102

Sinnhuber, RC. and Yu, TC. 1958. 2-Thiobarituric acid method for the measurement of
rancidity in fishery products. 211. The quantitative determination of
malonaldehdye. Food Technol. 1229-12.

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.

Stone, H., Sidel, J. Oliver, S. Woolsey, A. and Singleton, RC. 1974. Sensory evaluation
by quantitative descriptive analysis. Food Technol. 28(1 1):11-13.

Tamura, H., Kitta, K. and Shibamoto, T. 1991. Formation of reactive aldehydes from
fatty acids in a Fe2+/ H202 Oxidation System. Am. Chem. Soc. 39:439-442.

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

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.
Phann. 48:183-189.

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

Tomas, M.C.and Anon, MC. 1990. Study on the influence of freezing rate on lipid

oxidation in fish (salmon) and chicken breast muscles. Inter. J. Food Sci. Technol.
25:718-721.

Turner, E.W., Paynter, W.D., Montie, E.J., Bessert, M.W., Struck, G.M. and Olsen, EC.

1954. Use of the 2-thiobarituric acid reagent to measure rancidity in frozen pork.
Food Technol. 82326-330.

USDA. 1979. Composition of foods: Poultry products. AW,
United States Department of Agriculture, Washington, DC.

Wade, Jr., L.G. 1987. W- Prentice-Hall, Englewood Cliffs, NJ.

Wang, C.H., Abouzied, M.M., Pestka, J .J . and Smith, D.M. 1992. Antibody development
and enzyme-linked immunosorbent assay for the protein marker lactate

dehydrogensae to determine safe cooking end-point temperatures of turkey rolls.
J. Agri. Food Chem. 40:1671-1676.

103

Warner, K., Evans, CD, List, G.R., Dupuy, H.P., Wadsworth, J.I. and Goheen, GE.
1978. Flavor score correlation with pentanal and hexanal contents of vegetable
oil. J. Amer. Chem. Soc. 55:252-256.

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

Witte, V.C., Krause, GT. and Bailey, ME. 1970. A new extraction method for
determining 2-thiobarituric acid values of pork and beef during storage. J. Food
Sci. 35:582-585.

"1111111111111111“