. _ .

(4va o
.. o

.M.
. .
vh.’ « I‘Ofic ,KI ' 0
MWan h wlqr

..

.4. .73...

v.4: I
. Tun.
.despvbut n .2. .
96f. Jr .3
r .

\..

Hf.- --

42:21.2; . .. .‘

 

w
0-0“... .
«11'? .
Elfin...
.i... ,
vomflfii... ..

+3

4'.
f"

"I

'J

.-

.. I»

.flwffirih...“ ma. 4.
W... . x : .

c

.2
.
..
c

. umw

. . 2 .3. a?
..+.§mr.

u. bin

.4
. mug»...

slut o. c

34..

r
:3: .

“m. m... .55 but. .
.._ WM” #3“; ..
..... . 3 $1

.627 AVG§U~Wx .

Tim. st.

.338"? -
. Lug... ..

35w...“ v.52 on.
may...“ .
EL. . .. v

. . v...

I... .11 1 1m. -
Mun..-» rwfsvruidnq

Puffinrvlt

: Want... hf .. «$51....»

.....Lw 1 .3th hum...

.3? pm.

To

. thrl

.
i v
. .2911... :4».
BY. 11... . 3: r...
... flmuruflf. . . a..." ... .........hmm?w.va.§

t. t!» . .1. Al\

rlrhfiyrwiLulnli
p . ..:...1...,
. v.» a...

Q .03..
I x .v‘:

0..
.Q

4.3.. .3."
a... .
. . . tsumuél.
wfinhwfiméfi
............ p“... $43.3... .2... .

.n.

L an
9....

- fundhm... 9...”... 51... a... g .
1.4.31... .3. 1 o. 3.. i -u
v. .. fl... Kumb..mw..v. wk.

ol
5|!

I: .3
A... V
. (u
.
.

, .v .1 .1

at- r ._....v~.w.3

.1 o¥.r......§........ .. . .
. . \._w .d 3

:0 1-..4 $.56.
flaw... . It»...
. I c. mfg ‘1’.

Wm... .

.

«V

l.-

..»m. . .

. .61. I...

k...

:2. kayak,

qumufi .
any”: _. .i ..

o l 1 4k. 0. k .
.Vxfioém. .. 3.3 war. 3%

i 51 A.

.6 :1
3

.Mm
.. .0.

kwwwfinpfi. nu EB,

.7135

um... i. “a. is

7.1!

. .1011..- L
gor- v
.. v vaiobl.

I! . ..
II”. .v-
vi. JJ It!
.. NJ 19.1.3»;

14.03.!

.rvuJ.

w'.. 5..
l'.(‘wvl \l!

x‘i .v.‘ 1. .
£10321 .

.I

1......
i. .

filo}. I;4
. . . .

haml

.s..I.I: ill?

.05;

3!...P...

A...

1.!11

I:

. och—hwy“. 313:.
r... 13%
. Wfinflum-

#8 .

I . z’...

1.. :3...
J. J! .. flit-3...... . .

.u, Huh... .

if. ..

1, C L;
13...... @v.

\ ....~ “.5. t

....n.

«Univ
.5747:
. . .. .

I. 1

. vt‘rlrt‘ivov ’Lv. vi
. ) v'l.xu.tuxo‘o .vn...,.v:| N9). ..

. . . . l\.v!‘.V.v I. .
. 3137.! {55v $341.31...

. ..-\t v.

...u an» L.

:{A‘

 

3......
.. .....!..3..o|.
3.. . 3.423% e
. .(fiozfltupxnt
xio A
:3 ,... 0o -3...
r... 1.37. I...
.3 5'7.
. . . 0|! 1 I.
.II.
I...

. . 2| f
71. ‘1 ~04
..

. .129. Jan,
. . 7 t
.. .thythwA

.\ 0" .1:

.. ul. I.
. .IAI-I.".- -‘
11.9-];

. .. .....,.I..-_y...
. Iflfo.’ . u

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Hi Iii/TI!Sl'iilliii/ililiii/Will

3 1293 01563 9143 f

This is to certify that the

dissertation entitled

Production and Characterization of Polyclonal Antibodies to
Hexanal-Lysine Adducts for Use in an ELISA to Monitor
Lipid Oxidation in a Meat Model System

presented by

Stephanie A. Smith

has been accepted towards fulfillment
of the requirements for

Ph.D. degree in Food Science .

 

 

Major professor

Date June 26, 1997

 

MSU is an Affirmative Action/Equal Opportunity Institution 0-12771

 

 

*LiaaAnv
Michigan State
University

 

 

 

~ PLACE ll RETURN BOX to remove thle checkout from your record.
TO AVOID FINES return on or before due due.

DATE DUE DATE DUE DATE DUE

    
 

     

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

CW
[:34

E

MSU Ie An Affirmative AotiorVEquel Opportunity Institution
Wanna

 

 

 

 

 

 

 

 

 

 

 

 

 

PRODUCTION AND CHARACTERIZATION OF POLYCLONAL ANTIBODIES TO
HEXANAL-LYSINE ADDUCTS FOR USE IN AN ELISA TO MONITOR
LIPID OXIDATION IN A MEAT MODEL SYSTEM
By

Stephanie A Smith

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTORATE OF PHILOSOPHY

Department of Food Science and Human Nutrition

1997

ABSTRACT
PRODUCTION AND CHARACTERIZATION OF POLYCLONAL ANTIBODIES
To HEXANAL-LYSINE ADDUCTS FOR USE IN AN ELISA To MONITOR
LIPID OXIDATION IN A MEAT MODEL SYSTEM
By

Stephanie A. Smith

Hexanal content is a popular index of lipid oxidation in foods, including meats.
Lipid oxidation in meat is generally associated with development of objectionable flavors
potentially limiting shelf-life and causing loss of consumer acceptability. Interest in new
methods for measuring the extent of lipid oxidation has arisen due to limitations of
standard approaches, such as thiobarbituric acid (TBA) assay and headspace gas
chromatography (HS-GC). While TBA assays are sensitive and simple, they are non-
specific. While HS-GC methods are sensitive and specific, they are typically time-
consuming and often difficult. Enzyme-linked immunosorbent assays (ELISAS) are
considered sensitive, specific, rapid and easy-to-use. The goal of this study was to
develop an ELISA incorporating hexanal-specific antibodies to measure hexanal in meat.

We developed and characterized highly sensitive and specific polyclonal antibodies
against hexanal-modified bovine serum albumin. The resultant antibodies recognized
hexanal conjugates, but not free hexanal in 0, 10, 20 or 30% methanol. Reactivity

occurred with hexanal conjugates containing a protein as large as keyhole limpet

hemocyanin (400 kD) or a compound as small as e-aminocaproic acid, a lysine derivative
(131 D). In contrast, the antibodies reacted much more specifically with the “hapten-end”
of conjugates. When chicken serum albumin was modified with other aldehydes, ketones,
and alcohols, instead of hexanal, only two compounds showed substantial cross-reactivity.
Heptanal, the aliphatic aldehyde with one more carbon than hexanal, and pentanal, with
one less carbon than hexanal, had cross-reactivities of 86.3% and 11.8%, respectively.
This polyclonal-based ELISA had a detection limit of 1.5 nanogram hexanal per milliliter
solution with a mean interplate coefficient of variation of 9.3% (n = 281). The assay was
used to assess the extent of lipid oxidation resulting from incubation of chicken thigh
muscle homogenate at 50°C. Comparison of these results with results obtained by a TBA
assay and a dynamic HS-GC method showed high correlations (r = 0.85 and r = 0.89,
respectively). A monoclonal antibody to hexanal-lysine adducts, produced in a separate
study, was also incorporated into an ELISA and used to measure hexanal content. These
results correlated similarly with standard methods. Further study of lipid oxidation in

muscle foods using immunoassays is recommended.

 

To my parents, Drs. Patricia G. and George L. Smith,
for all their love and support.

iv

ACKNOWLEDGMENTS

For their valuable input and support, I would like to recognize the Michigan State University
faculty members who served on my doctoral committee, Drs. .1 Ian Gray, L. Patrick Hart.

James J. Pestka and, my advisor, Dr. Denise M Smith.
For its financial support, I would like to recognize the MSU Crop & Food Processing Center.

For their technical assistance, I would like to thank Dr. Virginia Warner. Dr. Enayat Gomaa.
Matthew Rarick. Brian Chadwick, Donald Herrington and, especially, Drs. James Clarke and

Robert T empleman.

For their continued support through the years, I would like to recognize the Ohio State University

faculty members, Drs. Grady Chism, Mike Mangino, David B. Min and Emil Mikolajcik.

For making my doctoral experience much more than an academic one, I would like to thank my

many friends in the department, in particular. Alicia Orta-Ramirez.

I am especially grateful to Drs. Ivy Hsu and Arti Arora for sharing the dissertation-writing process

and to my lab-mates, Arnie Sair, Manee Vittayanont, Dr. Giri Veeramuthu and Tammy Zielinski.

Finally, I would like to thank my very dear friend, Steven J. Ramey, who made me decide to pursue

a doctoral degree and who has had confidence in me throughout my quest.

TABLE OF CONTENTS

List of Tables .................................................................................. ix

List of Figures ................................................................................. x

Chapter 1. Introduction ........................... . ...................................... 1

Chapter2. LiteratureReview..................................L ......................... 5

2.1 Lipid Oxidation Mechanisms ............................................... 5

2.2 Factors Influencing Lipid Oxidation in Food ............................. 11

2.3 Flavor Significance of Oxidative Products ............................... 14

2.4 Hexanal as an Indicator of Lipid Oxidation .............................. 15

2.5 Measurement of Lipid Oxidation .......................................... 18

2.6 Immunoassays ................................................................ 24

2.7 Epitope Structure ............................................................ 27
Chapter 3. Production and Characterization of Polyclonal Antibodies to

Hexanal-Lysine Adducts ..................................................... 33

3.1 Abstract ....................................................................... 33

3.2 Introduction .................................................................. 34

3.3 Materials and Methods ...................................................... 37

3.3.1 Protein-Hapten Conjugation by Reductive Alkylation ......... 37

3 .3 .2 Protein Labeling by Reductive Alkylation ....................... 40

3.4

Chapter 4.

4.1
4.2

4.3

3.3.3 Polyclonal Antibody Production .................................. 40

3 .3 .4 Indirect ELISA for Serum Titer Detemiination ................. 41
3.3.5 Antibody Specificity Determination by CI-ELISA .............. 43
3 .3 .6 Assay Sensitivity and Precision .................................... 44
3.3.7 Western blot .......................................................... 44
Results and Discussion ...................................................... 45
3.4.1 Preparation of Hexanal Conjugates ............................... 45
3.4.2 Antiserum Titers ..................................................... 47
3.4.3 Assay Sensitivity and Precision .................................... 49
3.4.4 Protein Specificity ................................................... 49
3.4.5 Aldehyde Specificity ................................................ 52
3.4.6 Amino Acid Specificity ............................................. 59
ELISA to Monitor Lipid Oxidation in a Meat Model System by

Measurement of Hexanal Content ......................................... 66
Abstract ....................................................................... 66
Introduction .................................................................. 67
Materials and Methods ...................................................... 70
4.3.1 Preparation of Muscle Homogenate ............................. 70
4.3.2 Hexanal Modification of Protein Solutions ..................... 71
4.3.3 Accelerated Lipid Oxidation Model System .................... 72
4.3.4 Hexanal Measurement by CI-ELISA ............................ 73
4.3.5 TBA-RS Assay ..................................................... 73
4.3.6 Isolation and Concentration of Volatile Compounds .......... 74

vii

4.3.7 Hexanal Measurement by Gas Chromatography ............... 75
4.3.8 Statistical Analysis .................................................. 76

4.4 Results and Discussion ...................................................... 76

4.4.1 Hexanal Measurement by Polyclonal Antibody-based

CI-ELISA ......................................................... 76
4.4.2 Comparison of a Monoclonal versus Polyclonal Antibody-
based CI-ELISA .................................................. 81
4.4.3 Detection of Hexanal in Oxidized Chicken Thigh
Homogenate ....................................................... 83
Chapter 5 Conclusions ................................................................. 90
Chapter 6 Future Research ................................. . .......................... 92
List of References ............................ ' ................................................ 95

viii

LIST OF TABLES

Table
2.1 Fatty acid precursors, hydroperoxide intermediates and B-scission reaction

pathways of alkoxyl radicals forming various saturated aldehydes ............. 8

2.2 Studies using hexanal measurement to assess lipid oxidation in foods
during processing or storage ........................................................ 17

2.3 Studies characterizing the dominant epitopes of antibodies against
immunogens formed by reaction of aldehyde and protein using sodium
cyanoborohydride as reducing agent ................................................ 32

3.1 Amino acid composition (mole percent) of native and hexanal-modified
chicken serum albumin (CSA) ...................................................... 46

3.2 Serum titers of polyclonal antibodies against hexanal-modified bovine
serum albumin (BSA) in rabbits .................................................... 48

3.3 Concentration of reactive amino groups in native and modified chicken
serum albumin (CSA) solutions as measured by trinitrobenezenesulfonic
acid (TNBS) assay ................................................................... 55

3 .4 Cross-reactivity of polyclonal antibodies against hexanal-modified bovine
serum albumin with aldehyde-modified chicken serum albumin (CSA) by
competitive indirect ELISA ......................................................... 58

3.5 Reactivtity of polyclonal antibodies against hexanal-modified bovine
serum albumin with aldehyde-amino acid conjugates by competitive
indirect ELISA ........................................................................ 63

4.1 Comparison of loss of reactive amino groups and antibody binding
inhibition of difi‘erentially-modified ovalbumin-hexanal conjugates ............ 79

4.2 Correlation coefficients (r) of measurements of the extent of lipid oxidation
of chicken thigh homogenate. Methods were headspace gas chromato-
graphy (HS-GC), thiobarbitun'c acid-reactive substances (TBA-RS) assay
and competitive indirect (Cl) ELISA .............................................. 87

ix

Figure

2.1

2.2

2.3

2.4

3.1

3.2

3.3

3.4

LIST OF FIGURES

Autoxidation of linoleic acid: Breakdown of l3-hydroperoxide to hexanal
via B-scission route b .................................................................. 9

Autoxidation of linoleic acid: Breakdown of 9-hydroperoxide to hexanal
via B-scission route a .................................................................. 10

Competitive indirect enzyme-linked immunosorbent assay ...................... 25

Structures of three types of products resulting from hexanal-protein
reactions: Unstable (Schifi‘ base), stable-reduced and stable-nonreduced 28

Standard curve of hexanal-modified chicken serum albumin (CSA).
Hexanal concentration determined from protein concentration, based on
amount of tritium incorporated into hexanal-CSA during reductive alkylation.. 50

Protein specificity of antibodies to hexanal-modified bovine serum albumin
(BSA) was determined by CI-ELISA. Microtiter wells were coated with

2 ug/mL hexanal-modified chicken serum albumin (CSA) and incubated

with antiserum diluted 123160. Unmodified and modified CSA, BSA and
keyhole limpet hemocyanin (KLH) competed at protein concentrations of

0.01, 0.1, l, 10 and 100 ug/mL. Determinations were made in triplicate ...... 51

Western blot of unmodified and hexanal-modified chicken serum albumin

(CSA; Lanes 1 and 2), bovine serum albumin (BSA; Lanes 3 and 4) and high
molecular weight standard (HMW S; Lanes 5 and 6) containing myosin

(205 kD), B-galactosidase (116 kD), phosphorylase b (97.4 kD), BSA

(66 kD) and ovalbumin (45 kD) ...................................................... 53

Specificity of antibodies to hexanal-modified bovine serum albumin was
determined by CI-ELISA. Microtiter wells were coated with 2 ug/mL
hexanal-modified chicken serum albumin (C SA) and incubated with antisemm
diluted 13160. Free hexanal (open symbols) or CSA-conjugated hexanal
(closed symbols) competed in 0, 10, 20 or 30% methanol. Hexanal concen-
trations of conjugates were based on tritium labeling results. Determinations
were made in triplicate ................................................................. 54

3.5

3.6

3.7

3.8

4.1

4.2

4.3

Aliphatic aldehyde specificity of antibodies to hexanal-modified bovine serum
albumin was determined by CI-ELISA Microtiter wells were coated with 2
ug/mL hexanal-modified chicken serum albumin (CSA) and incubated with
antiserum diluted 1:3160. Aldehyde-modified CSA competed at protein
concentrations of 0.01, 0.1, 1, 10, 100 and 1,000 ug/mL. Aldehydes were:
propanal (3C), butanal (4C), pentanal (5C), hexanal (6C), heptanal (7C),

octanal (8C) and nonanal (9C). Determinations were made in triplicate ...... 57

Carbonyl compound specificity of antibodies to hexanal-modified bovine

serum albumin was determined by CI-ELISA. Microtiter wells were coated

with 2 ug/mL hexanal-modified chicken serum albumin (CSA) and incubated
with antiserum diluted 1:3160. Modified CSA competed at protein concen-
trations of0.001, 0.01, 0.1, 1, 10, 100 and 1,000 ug/mL. Modifying

compounds were: 2-trans-hexenal(21—HEX-CSA), 3-methyl-butanal (3-
MB-CSA), 2-methylpentanal (2-MP-CSA), 2-methylbutanal (2-MB-CSA)

and 2-hexanone (2-K-CSA). Determinations were made in triplicate ........... 60

Structures of thirteen atom compounds: hexanal-modified e-aminocaproic
acid, octanal-modified y-aminobutanoic acid and tn'decanoic acid ............... 6]

Acid specificity of antibodies to hexanal-modified bovine serum albumin

was determined by CI-ELISA. Microtiter wells were coated with 2 ug/mL
hexanal-modified chicken serum albumin and incubated with antiserum

diluted 1:3160. Conjugates competed at concentrations of 0.001, 0.01, 0.1,

1 and 10 mM. Conjugates were: e-aminocaproic acid (ACA)-hexanal, y-
aminobutanoic acid (GABA)-octanal, lysine (LYS)-hexanal and glycine
(GLY)-hexanal. Determinations were made in triplicate .......................... 62

Effect of protein (hexanal) concentration on polyclonal antibody binding as
measured by CI-ELISA. Solutions of 10 mg/mL protein of chicken serum
albumin (CSA), ovalbumin (0A) and chicken breast homogenate (CBH)

were maximally-modified with excess hexanal, diluted to protein concentra-

tions of0.003, 0.01, 0.03, 0.1, 1, 3, 10, 30 and 100 ug/mL and assayed.
Determinations were made in triplicate ............................................... 77

Efi‘ect of hexanal concentration on polyclonal antibody binding as measured

by CI-ELISA. Solutions of 10 mg/mL protein of ovalbumin (0A) and

chicken breast homogenate (CBH) were differentially-modified with varying
amounts of hexanal. The added hexanal concentrations (prior to reduction,
dialysis and dilution) were 2, 4, 6, 8, 10, 15, 20 and 25 mM. The solutions

were diluted to 20 ug/mL protein and assayed. Determinations were made

in triplicate ............................................................................... 80

Comparison of added and measured hexanal concentrations for hexanal-
modified ovalbumin solutions. Solutions of 10 mg/mL protein were

xi

4.4

4.5

combined with varying amounts of hexanal, reduced, dialyzed and diluted
1:500 to a protein concentration of 20 ug/mL. Determinations were made
in triplicate by polyclonal antibody-based CI-ELISA ...........................

Binding curves of differentially-modified chicken thigh homogenate solutions
diluted to low (2 ug/mL), medium (20 ug/mL) and high (200 ug/mL) protein
concentrations. Solutions had added hexanal concentrations (prior to
reduction and dialysis) of 50, 100, 200, 500, 1000, 2000 and 5000 ug/mL.
Solutions were subjected to monoclonal antibody (MAb; open symbols)
-based and polyclonal antibody (PAb; filled symbols)-based CI-ELISA.
Determinations were made in triplicate .............................................

Efi‘ect of incubation time on hexanal and malonaldehyde concentrations of
chicken thigh homogenate as measured by dynamic headspace gas chromato-
graphy (HS-GC), monoclonal antibody-based competitive indirect ELISA
(mELISA), polyclonal antibody-based competitive indirect ELISA (pELISA)
and thiobarbituric acid—reactive substances (TBA-RS) assay. TBA-RS
results expressed as milligrams malonaldehyde per kilogram meat.
Determinations were made in triplicate .............................................

xii

82

84

85

CHAPTER 1

INTRODUCTION

The poultry industry has grown tremendously especially with the development of
further-processed chicken products (Kinsman, 1994). A dramatic increase in the amount
of chicken consumed domestically has occurred since 1970 (Kinsman, 1994). In addition,
U.S. broiler meat exports hit a record-high of$1.7 billion in 1995 and forecasts are $2.5
billion for 1997 (Thornton, 1996). Consumer acceptance of meat products depends to a
major extent on flavor quality (Ramarathnam et al., 1991a). Poultry muscle is very
susceptible to oxidative deterioration resulting in off-flavor development because of its
relatively high concentration of unsaturated fatty acids and low concentration Of
tocopherols (Ladikos and Lougouvois, 1990). Unfortunately, the potential for lipid
oxidation and rancid or warmed-over flavor development is one of the single greatest
constraints in determining the processing and shelf-life characteristics of muscle foods
(Allen and Foegeding, 1981).

A uniform and satisfactory method to evaluate flavor quality and stability of lipid
foods needs to be developed (Min, 1981). Fifteen years later, the subject of appropriate
measurements continues to be one of considerable debate among researchers. In May
1996, at the American Oil Chemists’ Society (AOCS) Annual Meeting and Expo in

Indianapolis, the AOCS Antioxidant Common Interest Group sponsored a panel

2
discussion on the topic (Cuppett and Warner, 1996). The discussion was organized to

initiate a dialog on accepted protocols for antioxidant research in foods and biological
systems and the panel was composed of a diverse group of international researchers.
While no consensus was reached regarding available protocols, several key points were
made. First, the diversity of methodologies currently in use to evaluate the progress of
oxidation in both food and biological systems has created difficulty in interpreting and
comparing the published data of different researchers. Second, prior to selecting methods,
researchers should clearly define: 1) the study objectives, such as kinetics, reaction
mechanisms, and antioxidant effectiveness; 2) the oxidation system, such as bulk oils,
emulsions, or biological materials; 3) the oxidation targets, such as lipids or proteins; and
4) the oxidation endpoints, such as induction periods, tissue damage or quality changes.
Third, no progress will be made if researchers continue to use archaic tests that measure
nonspecific products in complex food and biological systems.

For purposes of food analysis, it is extremely doubtful that a reliable instrumental
method will replace the human being for sensory evaluation of foods in the near filture
(Reineccius, 1996). While the most commonly used technique for flavor analysis is gas
chromatography, its use for non-research applications, such as quality assurance, has been
limited due to time constraints and method complexity (Reineccius, 1996). Immunoassays
offer a rapid, cost-efi‘ective, and easy-to-use alternative, achieving sensitivity and
specificity without requiring highly trained analysts or sophisticated equipment
(Samarajeewa et al., 1991). As an analytical tool, the immunoassay has already proven to
be very useful for the assessment of a variety of substances in meat products (F ukal,

r991).

3
To assess off-flavor development due to lipid oxidation in foods before an

unacceptable level is reached, a novel analytical approach which is both sensitive and
specific is needed. For testing to be performed easily in a processing or storage facility,
the assay should be simple and rapid as well. The overall goal of this research was to
develop an enzyme-linked immunosorbent assay (ELISA) to assess the extent of lipid
oxidation in muscle foods. The criteria set forth by Gray (1978; Gray and Monahan,
1992) were used to evaluate our rationale. The criteria are as follows:

0 Specificity of property to be measured (hexanal content) for lipid oxidation.

0 Degree to which property represents extent of lipid oxidation.

0 Specificity of method (ELISA) for property.

First, hexanal is a secondary product of lipid oxidation. Hexanal is produced
during autoxidation of all compounds with a double bond in the III-6 position, including 9-
hydroperoxy-IO,12-octadecadienoic acid (9-HPOD), 13-HPOD, 2,4-decadienal, and 2-
octenal (Schieberle and Grosch, 1981). In meats, oleic, linoleic and arachidonic acids are
the primary reactants of lipid oxidation (Ladikos and Lougouvois, 1990) with the latter
two fatty acids degrading to hexanal. Hexanal is also a major product of lipoxygenase-
hydroperoxide lyase activity in a variety of plant tissues (Hsieh, 1994).

Second, hexanal is one of the compounds contributing to the undesirable odor and
flavor associated with rancidity (Pearson et al., 1977). Due to its low odor threshold and
relatively high concentration in oxidizing foods, hexanal has flavor significance as well as
indicating lipid degradation. Over thirty years ago, hexanal was recognized as a useful
indicator of oxidative flavor deterioration in stored foods, specifically in potato granules

(Boggs et al., 1963), uncured, canned ham (Cross and Ziegler, 1965) and fiied chicken

4
(Nonaka and Pippen, 1966). Hexanal has long been used as an index of oxidative

deterioration in potato products (Salinas et al., 1994) and is very commonly used to
monitor hexanal content in vegetable oil due to its well-established correlation with
oxidation in that system (Reineccius, 1996). Hexanal appears to be a sensitive and reliable
indicator for assessment of the oxidative condition and flavor quality of meat and meat
products (Shahidi, 1994). For example, strong correlations have been reported between
sensory scores and hexanal content for cooked ground beef (St. Angelo et al., 1987) and
ground pork (Shahidi et al., 1987) with coemcients of 0.80 and 0.98, respectively.

Third, immunoassays are based on the inherently highly specific immunological
interaction between an antibody and a corresponding antigen (Hefle, 1995). Therefore,
we expect that the use of antibodies against hexanal in an ELISA will yield a detection

system highly specific to hexanal.

The first objective of this study (Chapter 3) was to produce and characterize
sensitive polyclonal antibodies specific to hexanal to be incorporated into an ELISA. The
second objective (Chapter 4) was to compare the ELISA to a “standard” gas
chromatography method for hexanal to monitor lipid oxidation in a model system. A
review of the literature (Chapter 2) further supports our selection of hexanal as an
indicator compound and immunoassay as our analytical technique. The sensitivity and
cross-reactivities of our anti-hexanal polyclonal serum were assessed and are discussed in
Chapter 3. Application of two competitive indirect ELISAs, one using our polyclonal
antiserum and one using a hexanal-specific monoclonal antibody, to measure lipid

oxidation in a chicken muscle model system is shown and discussed in Chapter 4.

CHAPTER 2

LITERATURE REVIEW

2.1 Lipid Oxidation Mechanisms

Oxidative deterioration of unsaturated lipids results in the formation of volatile and
nonvolatile secondary breakdown products via hydroperoxide formation. Hydroperoxide
formation occurs nonenzymatically by one of two mechanisms, autoxidation or
photosensitized oxidation. Autoxidation begins with hydrogen abstraction from allylic
methylenes, in the presence of trace metals, light or heat, producing allylic radicals which
react with molecular oxygen producing peroxy radicals. Subsequent hydrogen abstraction
fiom another unsaturated fatty acid by the peroxy radical results in the formation of a
hydroperoxide. Photoxidation begins as a direct addition of oxygen, activated to the
singlet state by exposure to light and a sensitizer, to the carbon-carbon double bond
yielding hydroperoxides formed at either unsaturated carbon. Each mechanism yields
difierent isomeric distributions of hydroperoxides from the same fatty acid (Frankel, 1984)
and subsequently forms different types and amounts of secondary products.

Hydroperoxide decomposition involves a very complicated set of reaction
pathways (Frankel, 1984) yielding a complex mixture of volatile and nonvolatile secondary
and tertiary reaction products, including acids, alcohols, aldehydes, esters, furans,

hydrocarbons, ketones, and lactones. Monohydroperoxides are very unstable and readily

decompose through homolytic cleavage of the hydroperoxide group to an alkoxy radical
and a hydroxy radical. The alkoxy radical undergoes B-scission of the carbon-carbon
bond via one of two routes. Scission route a is a cleavage of the carbon-carbon bond on
the side of the oxygen-bearing carbon fiirther from the double bond forming an
unsaturated aldehyde and an alkyl radical which can react with a hydroxyl radical to form
an alcohol. Scission route b is a cleavage of the carbon-carbon bond between the double
bond and the oxygen-bearing carbon forming a saturated aldehyde and vinyl radical which
can react with a hydroxyl radical to form a l-enol which tautomerizes to the
corresponding aldehyde. Alternative pathways of hydroperoxide decomposition are
further oxidation, yielding epoxides, cyclic and bicyclic peroxides, and condensation,
yielding dimers and polymers which subsequently may oxidize and decompose into volatile
products (Ladikos and Lougovois, 1990). Malonaldehyde, another common lipid
oxidation product, arises from cyclic endoperoxides produced by autoxidation of fatty
acids containing three or more double bonds. Other tertiary reactions, including
decomposition of non-volatile secondary products and further oxidation of unsaturated
aldehydes and ketones, contribute to the complexity of the flavor profile of an oxidized
food (Kochhar, 1993).

The aliphatic aldehydes are the most important volatile breakdown products
because they are the major contributors of unpleasant odors and flavors in food products
(Kochhar, 1993). Each unsaturated fatty acid gives a characteristic pattern of aldehydes
via oxidation (Esterbauer, 1982). Although the specific production pathways of all the
aldehydes found in various oxidized samples have not been elucidated, most of the

aldehydes are formed by one or more of the chain scission reactions (Esterbauer, 1982).

Oleic acid may generate octanal, nonanal or decanal while linolenic and arachidonic acids
may generate propanal and hexanal, respectively (Table 2.1).

Among the most important precursors of aldehyde compounds is linoleic acid due
to its abundance in foods and high susceptibility to oxidation (Kochhar, 1993).
Autoxidation of linoleic acid produces hexanal via 13-hydroperoxyoctadeca-9,11-dienoic
acid hydroperoxide following homolytic clevage scission route b (Figure 2.1) and 2,4-
decadienal via 9-oxo-nonanoic acid hydroperoxide following route a (Figure 2.2). Further
oxidation of 2,4-decadienal generates hexanal and 2-octenal (Schieberle and Grosch,
1981), while subsequent oxidation of 2-octenal yields heptanal and hexanal. In fact, all
compounds containing an 01-6 double bond produce hexanal (Schieberle and Grosch,
1981).

Experimental determination of the volatile carbonyl compounds arising from
autoxidation of linoleic acid at low and moderate temperatures indicates that hexanal is the
major component comprising 66 mol %, while 2-octenal, 2-heptenal, and 2,4-decadienal
are minor components comprising 18, 6, and 5 mol %, respectively (Badings, 1970). The
fact that hexanal is the only aldehyde to arise from both linoleic acid hydroperoxides, as
well as from some of the aldehydic breakdown products, explains its predominance as a
linoleic acid autoxidation product (Schieberle and Grosch, 1981). In addition, hexanal is
protected against further oxidation to some extent by the more reactive unsaturated
aldehydes present in a mixture (Schieberle and Grosch, 1981), as in the case of a food
system.

Lipid oxidation is a non-enzymic process in that there is no direct reaction of an

enzyme with an unsaturated fatty acid (Kanner, 1994). Lipoxygenase (LOX) does,

Table 2.1

Fatty acid precursors, hydroperoxide intermediates and B-scission reaction

pathways of alkoxyl radicals forming various saturated aldehydes

 

 

Fatty Acid Hydroperoxide Scission Route Aldehyde
Oleic acid 8-OOH b Decanal
9-OOH b Nonanal
10-OOH b Nonanal
ll-OOH b Octanal
Linoleic acid 13-OOH b Hexanal
Linolenic acid l6-OOH a Propanal
Arachidonic acid lS-OOH a Hexanal

 

Adapted from Kochhar (1993)

11

Hsc \/\/\==/\=/\/\/\/\ _.
C CH
\\

13 12 1o 9 o
LINOLEIC ACID
O-OH
H30MWC—OH
\\
13 11 9 o

13-HYDROPEROXYOCTADECA-9,11-DIENOIC ACID

Figure 2.1. Autoxidation of linoleic acid: Breakdown of 13-hydroperoxide to hexanal
via B-scission route b.

\\
13 12 10 9 O
LINOLEIC ACID
12 1O O—OH //0
H,C WM C \ 01-1
3

9-HYDROPEROXYOCTADECA—10.12-DIENOIC ACID

1 scission route a

2,4-DECADIENAL
/ \.

/OH /OH

O O

’LCWWquo KCW/La—ETC-fiuo
WCflzo
H,c

ire/\/\/CH=° ‘/ 2-OCTENAL

HEXANAL

Figure 2.2. Autoxidation of linoleic acid: Breakdown of 9-hydroperoxide to hexanal via
. B-scission route a.

 

11

however, participate in oxidative deterioration of biological systems by catalyzing
oxidation of certain fatty acids (Hsieh, 1994). Activated LOX can abstract hydrogen from
polyunsaturated fatty acids containing a l,4-cis,cis-pentadiene system (Hsieh, 1994). To
become activated, LOX requires fatty acid hydroperoxides (Kanner, 1994) to oxidize its
iron moiety from the ferrous to ferric state (Hsieh and Kinsella, 1989). Various LOX
isozymes exist differing in the position of hydrogen abstraction and oxygen addition, the
direction of double bond shift, and in relative afiinities for difi‘erent fatty acids (Hsieh and
Kinsella, 1989). Hence, they generate different types and amounts of products. For
example, linoleic acid is oxidized by lipoxygenase isozyme 1 (LOX-1) producing
approximately 90% c-13 monohydroperoxide and 10% of the c-9 isomer, whereas LOX-2
and LOX-3 produce 40% and 60% of the isomers, respectively (MacLeod and Ames,
1988). A second enzyme, a hydroperoxide lyase, acts specifically on 13-hydroperoxide of

linoleic acid to produce hexanal.

2.2 Factors Influencing Lipid Oxidation in Food

Lipid oxidation is considered a primary cause of quality deterioration in foods,
especially in muscle tissues (Love and Pearson, 1971; Melton, 1983; Rhee, 1988; Asghar
et al., 1988; Rahargo and Sofos, 1993; and Kanner, 1992). Numerous internal and
external factors influence the rate of lipid oxidation in a food system. The internal factors
are system components, such as fatty acids and metal ions (Pearson et al., 1977), and
system characteristics, such as water activity (F ritsch and Gale, 1977). External factors

1

include storage temperature, time, light, and oxygen availability (Hsieh and Kinsella,

1989)

12

The degree of unsaturation of the fatty acid constituents of the lipids is a primary
factor dictating the reaction rate (Shahidi and Pegg, 1994). In general, the rate of
autoxidation increases as the number of double bonds increases. In meats, linoleic, oleic,
and arachidonic acids are the primary reactants during lipid oxidation (Ladikos and
Lougovois, 1990). The relative oxidative susceptibility of lipids by species, based on the
polyunsaturated fatty acid content, in decreasing order is fish, poultry, pork, beef, and
lamb in both adipose fat (Pearson et al., 1977) and phospholipids (Kanner, 1994). Despite
the low concentration of phospholipids relative to adipose fat in meat tissue, the
combination of the high degree of unsaturation of their constituent fatty acids and their
close proximity to lipid oxidation catalysts make phospholipids primary targets of
oxidative reactions (Drum and Spanier, 1991). Transition-metal ions can stimulate lipid
oxidation by either the generation of initiators, e. g. hydroxyl radical, or the decomposition
of hydroperoxides into peroxy and alkoxyradicals (Halliwell, 1995). Both fiee and
protein-bound iron are capable of catalyzing oxidation of unsaturated fatty acids (Pearson
et al., 1977; Ladikos and Lougovois, 1990). Sources of protein-bound iron existing in
biological tissue include myoglobin, hemoglobin, and cytochromes. Reducing compounds
are the driving forces in metal ion-catalyzed lipid oxidation (Kanner, 1994) and ascorbic
acid is the main election donor for the iron-redox cycle in muscle tissue (St. Angelo,
1996). Ascorbic acid can either promote or inhibit lipid oxidation depending on its
concentration. Low concentrations generally promote oxidation by causing reduction of
metal ions (Ladikos and Lougovois, 1990); high concentrations inhibit oxidation by
donating hydrogen and inactivating fi'ee radicals (Decker and Hultin, 1992). While the

exact ascorbic acid concentration producing a given effect is dependent on the iron

13

concentration, at the concentrations of each found in animal tissue, ascorbic acid is most
likely a prooxidant (Decker and Hultin, 1992). Sodium chloride has a significant
prooxidant efi‘ect, while phosphates act as antioxidants (Gray et al., 1994). Water activity
is another factor influencing lipid oxidation. For most foods, the lowest reaction rate
occurs at a water activity of between 0.2 and 0.3. Water acts as an antioxidant by
decreasing the catalytic activity of transition metals; below 0.1, significant acceleration
occurs (Fritsch, 1994). The storage stability of dry foods such as cookies is mostly limited
by lipid oxidation (Lingnert, 1980).

External factors affecting lipid oxidation during both processing and storage must
be considered since they can be manipulated to control the oxidative susceptibility inherent
in some food systems. In muscle tissue lipids, oxidative flavor deterioration can arise
either rapidly as in refiigerated, cooked meats, apparent within 48 hours at 4°C.

Oxidative flavor deterioration can also arise slowly as in raw meat and fatty tissue, usually
requiring months of freezer storage (Spanier et al., 1992). It is now generally accepted
that the former process, producing a condition referred to as warmed-over flavor (WOF),
is not limited to cooked meat. Rapid off-flavor development is promoted in raw meat by
any process involving disruption of the muscle structure in the presence of air, including
cooking, grinding, or restructuring (St. Angelo, 1996) causing release of free iron and
catalyzing oxidation (Drum and Spanier, 1991). While cooked meat is more susceptible
to lipid oxidation than raw meat, when raw meat is subjected to size reduction,
temperature abuse and/or prolonged storage, oxidation can become a serious problem
(Rhee, 1988). Disruption of tissue is an important mechanism for the formation of

aldehydes, ketones, alcohols, and oxoacids in many plant as well as animal tissues (Hsieh

14

and Kinsella, 1989), since increased oxygen availability accelerates oxidation of
unsaturated fatty acids (Muller and Gautier, 1994). In low moisture foods, the drying
process creates channels as water is removed, allowing rapid migration of oxygen, and
also ruptures some lipid globules increasing the lipid surface area (F ritsch, 1994). A low
storage temperature, protection from light and oxygen, and a favorable water activity all
improve the oxidative stability of a food system containing lipid.

Lipoxygenases are present in many plant, animal, poultry, and fish tissues
producing flavor compounds desirable in some foods and offensive in others (Hsieh,
1994). Hexanal is a major product of the activity of a lipoxygenase-hydroperoxide lyase
enzyme system in a variety of plants, including apple, banana, corn, cucumber, green bean
and tomato (Hsieh, 1994; Muller and Gautier, 1994). Hexanal contributes to the
acceptable, characteristic “green” or “fresh” flavor in some foods, such as in green beans
(Hsieh, 1994). Hexanal also yields undesirable “green” or “beany” flavor in others, as in
soy products (MacLeod and Ames, 1988). Both plant and animal lipoxygenases can use
linoleic acid as a substrate; however, there is no reported direct correlation between
lipoxygenase activity and flavor generation in muscle foods (Hsieh, 1994). In raw muscle
foods, lipid oxidation may be catalyzed by activated lipoxygenases; while in cooked,
stored meat products, lipid oxidation is completely dependent on nonenzymatic catalysis

(St. Angelo, 1996).

2.3 Flavor Significance of Oxidative Products
To evaluate the organoleptic importance of a volatile compound arising from lipid

oxidation, it is necessary to know its threshold value and concentration in the sample

15

(Frankel, 1982; Gray and Monahan, 1992). Certain compounds present in trace
quantities have such an intensive odor they are significant to the overall flavor profile
(Frankel, 1982). Hexanal is one of the compounds contributing to the undesirable odor
and flavor associated with rancidity (Pearson et al., 1977). Characterized as having a
“green” odor as a pure compound, hexanal has a threshold in oil of 80 ppb (Mottram,
1987). Hexanal and other aldehydes have been used successfully to follow lipid oxidation
in meat products (Gray and Monahan, 1992). An increase in lipid-derived volatile
compounds, including hexanal, closely follows changes in sensory descriptors, such as

painty, used by panelists (Spanier et al., 1992).

2.4 Hexanal as an Indicator of Lipid Oxidation

Hexanal content appears to be a sensitive and reliable indicator for evaluation of
the oxidative state and flavor quality of meat and meat products (Shahidi, 1994). Hexanal
content has been monitored during refiigerated storage of cooked meats including:
ground pork (Shahidi et al., 1987); ground beef (St. Angelo et al., 1988); beef (St. Angelo
et al., 1987); roast turkey, chicken, and beef (Dupuy et al., 1987); turkey rolls (Wu and
Sheldon, 1988); roast pork (Robson et al., 1989); broiler breast, thigh, and/or skin (Ang
and Young, 1989; Ang and Lyon, 1990; Ajuyah et al., 1993); chicken patties (Su et al.,
1991; Ang and Huang, 1993); chevon (Lamikanra and Dupuy, 1990); and restructured
beef steaks (Stoick et al., 1991). The concentration of hexanal has also been monitored
during fiozen storage of raw meats, including ground pork (Brewer et al., 1992),
restructured beef steaks (Stoick et al., 1991), and alligator meat (Cadwallader et al.,

1994), and of cooked meats, including beef slices (Hwang et al., 1990), ground turkey

16

patties (Craig et al., 1991), and restructured chicken nuggets (Lai et al., 1995).
Measurement of hexanal has been used to assess the extent of lipid oxidation in canned
tuna (Przybylski et al., 1991) and to compare cured and uncured meats, including ham
(Cross and Ziegler, 1965), pork (Ramarathnam et al., 1991b), and beef and chicken
(Ramarathnam et al., 1991a). Larick et a1. (1992) found that pork from pigs fed diets
containing 4 or 6% linoleic acid had higher concentrations of hexanal than pigs fed diets of
2% linoleic acid. The authors suggest that while a more unsaturated fatty acid profile may
be desirable, the impact of fatty acid modification on the flavor and oxidative stability of
fresh pork needs to be evaluated.

The use of hexanal content as an indication of the extent of lipid oxidation is not
limited to meat and meat products. Measurement of hexanal content has been used to
monitor the development of lipid oxidation during storage or processing in a range of
foods (Table 2.2). In some cases, hexanal was used in combination with a second volatile
compound to monitor lipid oxidation and resultant flavor quality. Pentanal and hexanal
were used to assess flavor quality of vegetable oils (Warner et al., 1978) and potato chips
(J eon and Bassette, 1984). Both studies concluded that hexanal alone was as useful an
indicator as the pair. Octanal and hexanal were used to assess oxidation of hazelnuts
(Kinderlerer and Johnson, 1992). Octanal was selected as an indicator due to the high
amount of Oleic acid in hazelnut oil. While the content of both compounds increased
during storage, hexanal content increased more rapidly. Bengtsson et a1. (1967) used
measurement of ethanol and hexanal as well as sensory evaluation to assess oxidation in
peas subjected to different harvesting conditions, then blanched and frozen. Ethanol was

selected as a indicator compound because its presence indicates enzyme activity while

17

Table 2.2 Studies using hexanal measurement to assess lipid oxidation in foods during
processing or storage

 

 

Food Stage Monitored Reference
Cookies Storage Lingnert (1980)
Brown rice Storage Shin et al. (1986)
Hazelnuts Storage Kinderlerer and Johnson
(1992)
Milk base, spray-dried Storage Roozen and Linssen (1992)
Milk powder, whole Storage Hall and Andersson (1985)
Oat breakfast cereal Storage Fritsch and Gale (1977)
Oat products Processing Ekstrand et al. (1993)
Oatmeal Storage Guth and Grosch (1994)
Peanut paste Processing Muego-Gnanasekharan and
Resurreccion (1993)
Peanuts Storage Bett and Boylston (1992)
Peas Processing Bengtsson et al. (1967)
Pecans Storage Erickson (1993)
Pinto bean, dehydrated Storage Hartman et al. (1994)
Potato chip Storage .leon and Bassette (1984)
Potato flakes Storage Sapers et al. (1972)
Potato granules Processing Hallberg and Lingnert (1991)
Potato granules Storage Boggs et al. (1963)
Potatoes, puffed Storage Konstance et al. (1978)
Soybeans Processing Moreira et al. (1993)
Vegetable oil Storage Warner et al. (1978)

 

18

hexanal is formed either enzymatically or via autoxidation (Bengtsson et al., 1967). Both
ethanol and hexanal content corresponded with sensory results during post-harvest storage
at 20°C. However, only hexanal content reflected sensory scores during frozen storage
since hexanal continued to form while ethanol did not.

The ultimate criterion of suitability of a compound as an index of lipid oxidation
for foods is adequate correlation with sensory data (Shahidi, 1994). Hexanal has been
shown to relate well to off-flavor development in various foods during storage.
Correlation coefficients have been reported between sensory scores and hexanal content
for cooked and stored broiler breast and thigh (Ang and Lyon, 1990), ground beef (St.
Angelo et al., 1987), beef slices (Hwang et al., 1990), ground pork (Shahidi et al., 1987),
as well as oat breakfast cereal (Fritsch and Gale, 1977) with values of 0.92 and 0.84, 0.80,

0.89, 0.98, and 0.99, respectively.

2.5 Measurement of Lipid Oxidation 1

A uniform and satisfactory method to evaluate flavor quality and stability of lipid
foods needs to be developed (Min, 1981). When evaluating the usefulness of an analytical
procedure, three criteria should be considered: 1) the representativeness of the parameter
to be measured--does it consistently occur in oxidizing systems and does the degree to
which it occurs correspond to the extent of oxidation; 2) the specificity of the parameter
for lipid oxidation; and, 3) the specificity of the method for that parameter (Gray, 1978).

The methods used to measure the degree of lipid oxidation in foods can be
grouped into two categories based on the changes occurring in the oxidizing system.

Changes due to the primary reaction of lipid oxidation include hydroperoxide formation,

19

measured and expressed as the peroxide value (PV) and oxygen absorption. Changes
due to the secondary reactions of lipid oxidation include malonaldehyde formation,
traditionally measured by the thiobarituric acid (TBA) test, and volatile compound
formation, measured by gas chromatography (GC). These methods can be ranked by their
respective predictive values for stability, shelf-life, and consumer acceptability of the
product, in decreasing order, as sensory evaluation, headspace volatiles, oxygen
absorption, PV, and TBA-reactive substances (RS) (Frankel, 1993).

In general, measurement of the primary changes is useful only in the initial stages
of lipid oxidation (Pearson et al., 1977; Melton, 1983; Ajuyah et al., 1993; Rossell, 1994).
Hydroperoxides are tasteless and odorless and, therefore, do not contribute to rancid
flavor (Prior and Loliger, 1994). As intermediates, the peroxide concentration will
decrease as oxidative deterioration worsens. This may result in an underestimation of the
degree of oxidation (Gray and Monahan, 1992), particularly during storage (Shahidi,
1994). Another drawback is that the fat extraction step, necessary in the analysis of foods,
may cause the formation of additional peroxides or the decomposition of the existing
peroxides prior to PV determination (Rossell, 1994). Oxygen absorption is measured
indirectly as the loss of oxygen in the headspace above a sample in a sealed container.
Oxygen absorption methods have limited sensitivity and therefore require a high degree of
oxidation to detect a change (Frankel, 1993; Prior and Loliger, 1994). In addition, oxygen
absorption may occur during protein oxidation, as well as during lipid oxidation, and is
therefore not specific (Melton, 1983).

The secondary reactions result in formation of compounds which will affect the

flavor profile of the oxidizing food system. Sensory evaluation is the most important and

20

common way to determine flavor quality (Min, 1981) since odor and flavor evaluations
relate directly to consumer acceptance (Frankel, 1993). However, sensory evaluation is
time-consuming, expensive (Min, 1981) and the quality of the data is highly dependent on
the quality of training received by the panelists (Frankel, 1993) since recognizing and
quantifying lipid oxidation off-flavors requires both good taste acuity and considerable
experience (Fritsch, 1994).

The TBA test is the oldest and most frequently used test for assessing lipid
oxidation in muscle foods (Melton, 1983; Shahidi, 1994; Prior and Loliger, 1994) and
other biological systems (Gray and Monahan, 1992). The test is simple and sensitive,
resulting in the formation of a red chromophore measurable spectrophotometrically at
532-535 nm or by high-pressure liquid chromatography (HPLC) or GC. It has several
limitations. First, malonaldehyde is a relatively minor lipid oxidation product arising only
from fatty acids with three or more double bonds (Gray and Monahan, 1992; Prior and
Loliger, 1994). Therefore, it is not suitable for the measurement of oxidation products of
foods containing mainly Oleic acid with one double bond and linoleic acid with two double
bonds (Frankel, 1993). Second, the formation of malonaldehyde does not always result in
increased TBA values in stored muscle foods since the compound itself may react firrther
(Igene et al., 1985; St. Angelo, 1996). Third, it is not specific for lipid oxidation since
other food components also react with TBA (Prior and Loliger, 1994) and may result in
an overestimation of the extent of oxidation (Frankel, 1993). This limitation is reflected in
the common use of the term “thiobarbituric acid-reactive substances” suggesting that
other materials are measured in addition to malonaldehyde. Fourth, the amount of

chromophore formed and detected varies depending on the protocol used (St. Angelo,

21

1996). Therefore, the TBA test should be used in combination with a complementary
procedure, such as hexanal measurement, and/or fiequently compared to trained sensory
panel results (Igene et al., 1985; Gray and Monahan, 1992; Prior and Loliger, 1994).
Analysis of volatile compounds by GC is closely related to flavor evaluation and
therefore is the most suitable method for comparison with the results of sensory panel
tests (Frankel, 1993). In fact, GC is the most commonly used technique for flavor analysis
(Reineccius, 1996). Due to its high resolution, GC is the most powerfial tool available for
detemrining volatile oxidation products (Prior and Loliger, 1994). However, sample
preparation procedures used to obtain sufiicient amounts of the compounds to be
instrumentally detectable usually cause qualitative and quantitative compositional changes
in the sample (Jennings and Filsoof, 197 7). Isolation of the volatile compounds prior to
GC analysis is typically achieved by one of three techniques, or variation of these:
simultaneous distillation-extraction (SDE), static headspace, and dynamic headspace (also
called purge-and-trap). The SDE technique involves distillation of an aqueous solution of
the food and simultaneous extraction of the distillate with solvent. It is used extensively
for flavor extraction. It is used less widely for oxidation determination of foods because
of possible destruction of volatiles during distillation, loss of volatiles during
concentration, and coelution of some compounds with the solvent (Prior and Loliger,
1994). Static headspace sampling involves direct injection of the gases above a food
sample sealed in an airtight container onto a GC for analysis and requires virtually no
sample preparation. This technique is relatively simple, reproducible, and rapid
(Reineccius, 1996) and most accurately represents the flavor/odor above a food; however,

sensitivity is a problem a’rior and Loliger, 1994) since insufficient quantities of

22

compounds may not permit accurate and precise quantitation (Reineccius, 1996).
Dynamic headspace involves purging the volatiles from a sample by a stream of inert gas
and trapping them with an adsorbent material. Because volatiles from large amounts of
sample can be concentrated, the sensitivity is better than that of the static headspace
technique. However, several parameters will affect the performance of a purge-and-trap
system: sample temperature; length of time of volatile collection; nitrogen flow rate;
sample vessel geometry; use of vacuum; adsorbent material type; amount of adsorbent;
and sample sparging versus headspace purging (Vercellotti et al., 1992). Difi‘erences in
the relative affinities for an adsorbent and the vapor pressures of volatile compounds can
result in either breakthrough losses on trapping or poor recoveries (Vercellotti et al.,
1992). Recovery of volatile compounds from the adsorbent can be accomplished by either
thermal desorption or solvent elution. While thermal desorption is faster than solvent
elution, the heat may alter or destroy some volatile compounds (Prior and Loliger, 1994).
Thermal desorption and associated equipment, in general, is complex and costly compared
to materials needed for solvent elution (Olafsdottir et al., 1985). However, depending on
the chromatograph conditions and the solvent used, early-eluting compounds may be lost
in the solvent peak (Jennings and Filsoof, 1977). While clearly no single sample
preparation technique is uniformly satisfactory, a given procedure may be better for a
particular compound in a particular sample (Jennings and Filsoof, 1977).

Another drawback of GC headspace analysis is its restriction to measurement of
volatile compounds ignoring the contributions of non-volatile compounds to flavor
(Jennings, 1977). A survey of the literature indicates that hexanal binding has been

observed in numerous protein systems including: soy protein (Arai et al., 1970; Franzen

23

and Kinsella, 1974), soy protein isolate (Franzen and Kinsella, 1974; Gremli, 1974), soy
glycinin and beta-conglycinin (O’Keefe et al., 1991a, b), lysozyme (Funes et al., 1980;
Tashiro et al., 1985; Okitani et al., 1986), whey-based and egg-based fat replacers
(Schirle-Keller et al., 1992; Schirle-Keller et al., 1994), alpha-lactalbumin, bovine serum
albumin (BSA), leaf protein concentrate, single-cell protein, textured vegetable protein
(Franzen and Kinsella, 1974), myosin and actin (Gutheil and Bailey, 1992). Gremli (1974)
developed an analytical method using a high vacuum-shell fi'eezing system to determine
the amount of a flavor compound present in and “retained” by a protein solution .
preventing volatilization. By comparing measurements taken using this method to those
obtained using an ambient headspace procedure and using a solution devoid of protein as
the control, the percentages of flavor compounds “reversibly bound” and “irreversibly
bound” by the protein were calculated. For a solution containing 5% soy protein, the
amounts of hexanal reversibly and irreversibly retained were 37-44% and <5%,
respectively. To measure the concentration of protein—bound carbonyl compounds present
in a solution, Franzen and Kinsella (1974) used a variety of proteins to survey the extent
of nonvolatilization of various carbonyl compounds. The concentration of volatile
carbonyl compound over a carbonyl-protein solution was measured by headspace GC and
compared to control solutions devoid of protein. The concentration ratios of hexanal in
various protein solutions suggested that 10-21% of the hexanal added to the solution was
retained by the protein. Protein-bound hexanal has been measured directly by enzyme
assay (Chiba et al., 1979). Aldehyde dehydrogenase purified from bovine liver
mitochondria was found to convert soy protein-bound aldehyde and fi'ee aldehyde to

alcohol irrespective of carbon chain length and at comparable rates. Using a second

24

enzyme, alcohol dehydrogenase, which does not act on protein-bound aldehyde, total and
free aldehyde content can be measured and bound aldehyde content determined by
difference.

While it is extremely doubtful that any instrumental method will replace human
sensory evaluation of foods in the foreseeable future, various techniques can be used as
screening procedures and to supplement sensory analysis (Reineccius, 1996). Despite the
limitations, the purge-and-trap technique has been used for many quality control methods
(Vercellotti et al., 1992) and solvent extraction from a Tenax adsorbent has become
popular in research studies (Reineccius, 1996). However, fiirther use of GC outside the
research laboratory is hindered by time limitations and method complexities (Reineccius,

1996)

2.6 Immunoassays

The basis of all irnmunoassays is the highly specific immunological interaction
between an antibody and a corresponding antigen (Hefle, 1995). Currently, the most
common type of immunoassay is the enzyme-linked immunosorbent assay (ELISA)
(F ukal, 1991). Amplification is achieved by using a detector system consisting of an
enzyme label plus substrate. Detection is indicated by color development and quantitation
correlates to color intensity. Use of a competitive format allows quantitation of a
particular analyte in an unknown concentration in a sample (Figure 2.3).

Immunoassays are a rapid, cost-effective, and easy-to-use alternative to
conventional analytical techniques achieving sensitivity and specificity without requiring

highly trained analysts or sophisticated equipment (Samarajeewa et al., 1991). In addition,

25

' O'<: O-<<O

>88 Eofioggfig magmas—mace 8265 ozfioqfioo md 253m

 

 

om 3

35.5me was.
>00mfiz<
Egamzfizm

 

 

   

<u oEooaméomzc<

zm0_._.z< 5:5
0.953

new >DOm_._.Z<

if

 

sz_._.z< 5:5
O 0 3:80 ommca 2.0m

26

immunoassays permit simultaneous testing of many samples. Commercially available
immunoassay kits are portable and can be used for routine surveillance in a processing
facility or retail outlet. As an analytical tool, immunoassays have already proven to be
very useful for the assessment of a variety of substances in meat products, including
detection of pesticides, drugs, microorganisms, and microbial toxins (Fukal, 1991). To
improve the quality and safety of meat and poultry products, ELISAs have been developed
to monitor thermal processing, to detect species adulteration, non-meat protein addition,
various contaminants and residues, and to diagnose animal diseases (Fukal, 1991; Smith,
1995). Immunoassays developed to monitor endpoint cooking temperature replace assays
based on protein solubility, enzyme activity, color, and electrophoretic patterns which are
often imprecise and difiicult to interpret (Smith, 1995). ELISA kits using antibodies to
species-specific serum proteins have been commercialized for detection of species
adulteration of as many as 11 species by one assay (Smith, 1995). Meat sample
preparation may consist of as little as a saline extraction and dilution of the extract (F ukal,
1991).

Antibody production requires an irnmunogenic molecule to elicit an immune
response in the animal system. Haptens, substances possessing a molecular weight of less
than 1,000 daltons, are not irnmunogenic and must be chemically bound to a carrier
protein to fiinction as irmnunogens (Hefle, 1995). Protein modifications as simple as
ethylation and methylation have resulted in production of specific antibodies against the
modified protein and not to epitopes of the native protein (Steinbrecher et al., 1984).
Anti-aldehyde antibodies have already been developed to formaldehyde (Steinbrecher et

al., 1984), acetaldehyde (Steinbrecher et al., 1984; Israel et al., 1986; Klassen et al., 1990;

27

Niemela et al., 1991; Perata et al, 1992; Worrall et al., 1991; Israel et al., 1992; Klassen et
al., 1994; Thiele et al., 1994), malonaldehyde (Haberland et al., 1988; Palinski et al., 1989;
Palinski et al., 1995; Preobrazhensky et al., 1995), and 4-hydroxynonenal (Palinski et al.,
1989; Petit et al., 1995, Uchida et al., 1995). Petit et al. (1995) proposed the use of an
immunoassay incorporating anti-malonaldehyde (MA) antibodies for measuring MA-
induced cell membrane alterations resulting from lipid oxidation. Niemela (1993)
developed an immunoassay incorporating anti-acetaldehyde (AA) antibodies for detection
of AA-adducts and proposed its use for diagnosis of excessive alcohol consumption prior

to the appearance of clinical signs of chronic alcoholism.

2.7 Epitope Structure

Numerous studies have demonstrated covalent binding between aldehydes and
proteins occurs in vivo and that the adducts can be measured by immunological methods
(Curtiss and Witztum, 1983; Israel et al., 1986; Niemela et al., 1991; Worrall et al., 1991;
Palinski et al., 1995; Preobrazhensky et al., 1995; Yoritaka et al., 1996). Modifications
altering protein structure as little as methylation can render endogenous proteins
immunogenic (Steinbrecher et al., 1984).

The products of aldehyde-protein reactions differ in mechanism of formation and
structure depending on the reaction conditions and can be grouped into three categories:
unstable, stable-reduced, and stable-nonreduced (Figure 2.4). The “unstable” adducts are
Schiff bases, resulting from reaction between carbonyl carbons and free amino groups.
These products form relatively quickly (San George and Hobennan, 1986) and readily

dissociate when exposed to dialysis, gel filtration, or treatment with weak acid or base

28

Baggage—fin new .3268
-oESm A83 biomv 05883 acetone £895-35on Sow was—82 $260K mo womb ooh: «c 85835 vN eBmE

$coér£u=o~m£§ $508-3 FODOD< OmODOmmZOZ w._m<._.w

.5 Eu £0 £0 £0

HODDD< OmODOmm w..m<._.w

.5 Eu £0 £0 Eu -mo .22 Ifommummmrful

Ew<m aaiomv 5:02 mam/3,52:

,5 £0 £0 £0 Eu .5 H z Iazomzouzoumol

29

(Donohue et al., 1983; Tuma et al., 1984). These Schifi‘ bases can be stabilized by
reduction to secondary amines using a reducing agent. The resultant “stable-reduced”
products are relatively irreversible and generally withstand exhaustive dialysis and gel
filtration (Klassen et al., 1994). Because the reduced products are also stable to
denaturation and enzymatic and acid hydrolysis, the binding sites can be identified by
amino acid analysis (San George and Hoberrnan, 1986). “Stable-nonreduced” adducts are
also stable to prolonged dialysis but are distinctly difi‘erent from those produced under
reducing conditions (Klassen et al., 1994). Binding occurs initially between an aldehydic
carbonyl and an amino group of a protein forming a Schiff base as before. According to
San George and Hobennan (1986), the amino group in the protein terminus and the Schiff
base intermediate is stabilized by binding of the carbonyl carbon with the nitrogen of the
peptide bond between the N-terminal and adjoining amino acid residue (San George and
Hoberrnan, 1986). The resultant structure is a stable 5-membered ring, or imidazolidinone
derivative.

The structural difi‘erences of protein-aldehyde adducts formed under differing
reaction conditions is supported by several immunological studies. Antibodies, based on
epitope specificities, are able to distinguish aldehyde-modified proteins prepared under
reducing condition from those prepared under non-reducing conditions (Klassen et al.,
1990; Klassen et al., 1994; Thiele et al., 1994). The effect of reaction conditions on the
type and amounts of adducts formed was clearly demonstrated in a study of nonenzymatic
glucosylation. Curtiss and Witztum (1983) developed an immunoassay to measure
glucitollysine, the reduced hexose alcohol form of glucose conjugated to the e-amino

group of lysine. Glucose and low-density lipoprotein (LDL) were reacted under differing

30

conditions and the amounts of glucitollysine formed were compared. In the absence of a
reducing agent, glucose formed a labile intermediate Schifi' base with the e-amino group of
lysine which underwent an Amadori rearrangement to form a relatively stable ketoamine
which equilibrated with its hemiketal or ring form. When glucosylation is carried out in
the presence of sodium cyanoborohydride (NaCNBH3), the labile Schifl‘ base is
immediately and quantitatively reduced to glucitollysine. When sodium borohydride
(Nam-14) was used as the reducing agent, Schifi‘ base, ketoamine, and hemiketal products
are all reduced, resulting in the formation of glucositollysine and mannositollysine, its
epimer. The application of these findings was as follows. Glucosylated proteins and
lipoproteins have been implicated in diabetes-related atherosclerosis. The concentrations
of glucitollysine residues were found to be threefold higher in diabetic total plasma
proteins and isolated lipoproteins compared to nondiabetic proteins. Since the adducts
formed in vivo were mostly Amadori forms and, therefore, not recognized by the
antibody, pretreatment of the plasma with each reducing agent was required and the
difference in glucitollysine contents taken. An immunoassay for the quantitative
assessment of glycitollysine content should prove useful.

The type of the reducing agent also influences the adduct structure. Tuma et al.
(1987) attempted to characterize, by high-performance liquid chromatographic analysis,
the stable adducts formed through reactions of BSA and radiolabelled acetaldehyde using
three different reducing agents: NaBH4, N3CNBH3, and ascorbate. Each reaction
generated a different elution pattern. When NaBI-L was used as the reducing agent, four
different types of adducts were found, including ethyllysine. When NaCNBH; was used

instead, the ethyllysine structure predominated with only minor amounts of other adducts.

31

These findings can be explained by the fact that NaCNBH3 readily and selectively reduces
Schifi‘ bases giving a higher yield of modified lysines than NaBI-L which, in addition to
reducing Schifi‘ bases, also reduces aldehydes and ketones to alcohols (Jentoft and
Dearborn, 1979). Ascorbate, a physiological reducing agent, was found to stabilize
acetaldehyde-BSA adducts at concentrations comparable to those found in the liver (Tuma
et al., 1984). However, none of the four difi‘erent adduct structures formed in the
presence of ascorbate were ethyllysine (Tuma et al., 1987). Tuma et al. (1987) also
studied non-reducing conditions and the adducts formed using polylysine instead of BSA
under each of the scenarios discussed. Overall, of the six different adduct structures
detected, only one could not positively be identified as a lysine derivative.

Other studies have also indicated that lysine is the major amino acid participating in
the binding of aldehyde to protein (Table 2.3). Jentoft and Dearborn (1979) reported that
arginine, histidine, methionine, and tryptophan do not form stable derivatives with

aldehyde under conditions of reductive alkylation or do so very slowly.

$585me $388386 co 33m B
£82m 05 @808 2 3.0.3 a

 

32

 

33: a a 2.2 BEEEEEBEO 28 £996 8:232
63: :55; Ea ago 052386 98:6 5339.6
333 .130 Baum—em comma—-Ecocoixeutnmé _m:o=o§xo€»:-v ooze—32
3%: .a a 28:8 05355838.“: $202203: 8332
$8: .a a 205
.Ea: .a a made .33 c .a a .32 2335mm incense/V. caesium
$3: .a a EoEEam 052:5”: 82er8< 8:235
is: a a sfiefiem 05335055 8383,88 SEED:
ooaobmom Ben—8:5 “Endowed eomuomom

 

Emma memento. we ongxnoconofixo 832% main anoa
can £9302“ «o 8:82 ,3 vogue meowocsfig cmcmmwm mammogram mo 39E... 28583 05 $358830 863m .m.~ DEE.

CHAPTER3

PRODUCTION AND CHARACTERIZATION OF POLYCLONAL

ANTIBODIES TO HEXANAL—LYSINE ADDUCT S

3.1 ABSTRACT

Hexanal content is a popular index of lipid oxidation in foods, including muscle tissue.
Methods currently available to monitor oxidation of unsaturated fatty acids, including the
thiobarbituric acid-reactive substances (TBA-RS) assay, are less than optimal, particularly
in a complex matrix such as meat. The purpose of this study was to develop and
characterize hexanal-specific polyclonal antibodies and incorporate these into an enzyme-
linked immunosorbent assay (ELISA). Because free hexanal is nonimmunogenic, it was
covalently attached to a carrier-protein prior to immunization. Using sodium
cyanoborohydride to form stable conjugates, hexanal was reacted with bovine serum
albumin (BSA), for use as the imrnunogen, and to chicken serum albumin (C SA) for use as
ELISA solid phase. The degree of modification of BSA with hexanal was determined by
measuring fi'ee amino groups by trinitrobenzene sulfonic acid (TNBS) assay and complete
loss of lysine residues was observed following reductive alkylation. Polyclonal antibodies
produced in rabbits immunized with hexanal-modified BSA had antiserum titers as high as

6.15 x 10'5 . A competitive indirect ELISA was developed to evaluate antisera reactivity

33

34
by CI-ELISA. The immunoassay was sensitive with a detection limit of 1.5 ng/mL

hexanal and precise with a mean coefficient of variation of 9.3% (n = 281). The
antibodies were found to be carrier-independent, reacting with hexanal-B SA, hexanal-
CSA and hexanal-KLH conjugates but not the corresponding unmodified proteins.
Aldehyde specificity was evaluated using CSA-conjugated alkanals of 3-9 carbons. Cross-
reactivity with n-heptanal, n-pentanal, n-octanal and 2-methylbutanal was determined to be
86.3%, 11.8%, 2.2%, and 1.1%, respectively. The antibodies reacted strongly with
hexanal-modified lysine (a-, e-diaminocaproic acid) and very strongly with hexanal-
modified e-aminocaproic acid. The antibodies did not recognize fiee amino acids or free

hexanal. The immunodominant epitope is proposed to be hexanal-modified lysine.

3.2 INTRODUCTION

Lipid oxidation leads to the destruction of lipids in biological and food systems
(Kappus, 1991). A number of diseases have been related to the induction of lipid
peroxidation, including cancer (Kappus, 1991), rheumatoid arthritis (Lopez-Bote et al.,
1993) and Parkinson’s disease (Yoritaka et al., 1996). Evidence indicating that oxidized
low density lipoprotein (LDL) contributes to atherogenesis is substantial (Palinski et al.,
1995). Rancidity arising from lipid oxidation is often the decisive factor in determining the
usefiil storage life of food products (Frankel, 1993). Unfortunately, methodological
problems have severely limited efl‘orts to achieve a clearer understanding of lipid oxidation
and its effects (Gutteridge and Halliwell, 1990); however, development of improved

analytical procedures will lead to great advancement in the knowledge of lipid oxidation

(Kappus, 1991).

35
The most widely used assay for measuring lipid peroxidation is the 2—thiobarbituric

acid (TBA) test (Gutteridge and Halliwell, 1990; Gray and Monahan, 1992). While this
test is sensitive for oxidation products of fatty acids containing three or more double
bonds, it is not suitable for oxidation products of Oleic and linoleic acids (Frankel, 1993).
Another limitation of the assay is its lack of specificity since many other compounds are
also TBA-reactive, including amino acids and carbohydrates (Gutteridge and Halliwell,
1990). Many techniques have been used to measure lipid oxidation and evaluate oxidative
stability, each with limitations which have been previously reviewed (Gray, 1978; Melton,
1983; Prior and Loliger, 1994; Rossell, 1994; Shahidi, 1994). One approach that has been
successfully used to follow lipid oxidation in both biological specimens and food systems
is hexanal measurement by headspace gas chromatography (Shahidi, 1994). Hexanal is an
important oxidative decomposition product of omega-6 polyunsaturated fatty acids
(Frankel and Tappel, 1991) and is one of the compounds contributing to the undesirable
odor and flavor associated with rancidity in food products (Pearson etal., 1977). Static
headspace sampling involves direct injection of the gases above a sealed sample onto a gas
chromatograph (GC) for analysis and requires virtually no sample preparation. This
technique has been used successfully in oxidative susceptibility studies to measure hexanal
in the headspace of red blood cell membranes (Frankel and Tappel, 1991), of liver
homogenate (Frankel et al., 1989) and of plasma lipoproteins (Walzem et al., 1995).
While this technique is relatively simple, rapid, and reproducible, it lacks the sensitivity
necessary to detect compounds present in small quantitites (Reineccius, 1996). Dynamic
headspace methods, which involve purging volatile compounds from a sample by a stream

of inert gas and trapping them with an adsorbent material, have been developed to

36

improve sensitivity. However, several parameters will afi’ect the performance of a
dynamic headspace or purge-and trap system, including sample temperature, length of
time of volatile collection, carrier gas flow rate, sample vessel geometry, and type and
amount of adsorbent (Vercellotti et al., 1992). Currently, neither a standard method for
the quantitation of hexanal nor a convention for the reporting of hexanal data exist,
making comparison of results difficult. Hexanal-protein binding (Franzen and Kinsella,
1974; Gremli, 1974; Esterbauer et al., 1987; O’Keefe, 1991a,b; Gutheil and Bailey, 1992;
Schirle-Keller et al., 1992; Holley et al., 1993; Schirle-Keller et al., 1994) further
complicates accurate quantitation by headspace gas chromatographic methods.

Immunoassays are increasing in popularity because they ofi‘er both analytical and
cost-savings benefits over conventional techniques, including sensitivity, specificity, speed,
and simultaneous testing of many samples (Fukal, 1991; Sarnarajeewa et al., 1991;
Deshpande, 1994; Hefle, 1995). Protein modifications as simple as ethylation and
methylation have resulted in production of specific antibodies against the modified protein
and not to epitopes of the native protein (Steinbrecher et al., 1984). Anti-aldehyde
antibodies have already been developed to formaldehyde (Steinbrecher et al., 1984),
acetaldehyde (AA) (Steinbrecher et al., 1984; Israel et al., 1986; Klassen et al., 1990;
Niemela et al., 1991; Perata et al., 1991; Worrall et al., 1991; Israel et al., 1992; Klassen
et al., 1994; Thiele et al., 1994), malonaldehyde (MA) (Haberland et al., 1988; Palinski et
al., 1989; Chancerelle et al., 1991; Palinski et al., 1995; Preobrazhensky et al., 1995), and
4-hydroxynonenal (Palinski et al., 1989; Petit et al., 1995; Yoritaka et al., 1996). Petit et
al. (1995) proposed the use of an immunoassay incorporating anti-MA antibodies for

measuring MA-induced cell membrane alterations resulting fi'om lipid oxidation. Niemela

37
(1993) developed an immunoassay incorporating anti-AA antibodies for detection of AA-

adducts and proposed its use for diagnosis of excessive alcohol consumption prior to the
appearance of clinical signs of chronic alcoholism. In food, Jourdan et al. (1984)
developed an ELISA to measure limonin, a bitter triterpenelactone, in citrus juice.

The primary goal of this study was to produce and characterize sensitive
polyclonal antibodies specific to hexanal to be incorporated into an enzyme immunoassay.
Our hypothesis was that antibodies to hexanal could be produced and would recognize a
derivatized lysine epitope. The specific objectives were three-fold. First, because hexanal
is a hapten and non-immunogenic itself, it had to be covalently bound to a carrier protein.
It was critical to attach as many hexanal molecules as possible to the carrier protein,
presumably via the e-amino group of the lysine side chains, to ensure all protein fragments
appearing on the surface of antigen-presenting cells contained hexanal. Second, following
immunization of rabbits with hexanal-modified BSA, it was necessary to develop an
ELISA to determine sensitivity and specificity of the resultant polyclonal antibodies.
Third, it was important to characterize the antibodies for protein, aldehyde and amino acid
specificities by assessing antibody cross-reactivity with compounds of related and/or

relevant structure to that of the antigen.

3.3 MATERIALS & METHODS

3.3.1 Protein-Hapten Conjugation by Reductive Alkylation. Alkylation of proteins
and amino acids was performed by the method of Steinbrecher et al. (1984) with these
modifications: 1) aldehyde was added in excess to obtain maximum modification, 2)

aldehyde was added immediately prior to reducing agent, and 3) reaction was run at

38

ambient temperature and a basic pH to accelerate the reaction (J entoft and Dearbom,
1979). Specifically, an aliquot of 246 pL of hexanal (II-9008, Sigma Chemical Co., St.
Louis, MO) was added to 9 mL of 16.7 mg/mL bovine serum albumin (BSA; A-7511,
Sigma) diluted in phosphate-bufi‘ered saline (PBS; 0.1M sodium chloride, 0.01M sodium
phosphate, pH 7.4) and the solution vortexed. Reducing conditions at pH 8.5-9.0 were
established by adding 1 mL of 800 mM sodium cyanoborohydride (NaCNBH3; S-8628,
Sigma) dissolved in 0.1 N NaOH to the aldehyde-protein mixture. The resultant reaction
concentrations were: 200 mM aldehyde, 15 mg/mL protein, and 80 mM
cyanoborohydride. The solution was held at ambient temperature for 12 hr before
dialyzing overnight against two changes of PBS. The same protocol was used to modify
keyhole limpet hemocyanin (KLH; H-7017, Sigma), chicken semm albumin (CSA; A-
3014, Sigma), and a high molecular weight standard protein mixture (SDS-6H, Sigma)
with hexanal. The standard contained 6 proteins with molecular weights ranging from 29-
205 kD. The molecular weight standard was reconstituted by adding 1.8 mL PBS (pH
7.4), 100 11L of 10% sodium dodecyl sulfate (SDS; 28364, Pierce, Rockford, IL) and 100
uL of 0.1 N HCl and standing at ambient temperature overnight.

CSA was also modified with other aldehydes, ketones and alcohols: propanal
(propionaldehyde; P-6889, Sigma), butanal (butyraldehyde; 810,328-4, Aldrich Chemical
Co., Milwaukee, WI), pentanal (valeraldehyde; V-23 78, Sigma), heptanal (heptaldehyde;
H212-0, Aldrich), octanal (0-7128, Sigma), nonanal (nonyl aldehyde; N3,080—3, Aldrich),
trans-2-hexenal (I-I-73 83, Sigma), 2-methylpentanal (25,856-3, Aldrich), 2-methylbutanal
(M3,347-6, Aldrich), 3-methylbutanal (isovaleraldehyde; 14,645-5, Aldrich), and 2-

heptanone (12,336-6, Aldrich). CSA was carried through the reductive alkylation

39
procedure with the alcohols, 2-hexanol (12,857-0, Aldrich) and n-hexanol (hexyl alcohol;

H1,330-3, Aldrich) substituted for an aldehyde, as described above.

Hexanal-modified amino acids were prepared as described above, except that the
reaction concentrations were 200 mM hexanal, 20 mM sodium cyanoborohydride and 200
mM amino acid, and dialysis was not performed due to the small size of the molecules.
The amino acids were L-lysine monohydrochloride (L-5626, Sigma), D-lysine
monohydrochloride (L-5876), L-arginine (A-5006, Sigma), glycine (15527-013, Life
Technologies, Inc., Gaithersburg, MD), and e-amino-n-caproic acid (A-7824, Sigma).
Octanal was substituted for hexanal for reductive alkylation with y-arnino-n-butanoic acid
(GABA; A-2129, Sigma).

The degree of modification achieved in each reaction was determined by measuring
the loss of reactive amino groups by the trinitrobenzenesulfonic (TNBS) acid assay of
Habeeb (1966). Briefly, 1 mL of protein solution, containing approximately 1 mg/mL
protein, was added to 1 mL of 4% sodium bicarbonate (NaHCO3; 7412, Mallinckrodt,
Paris, KY) bufl‘er, pH 8.5, and 1 mL of0. 1% TNBS (picrylsulfonic acid; P-2297, Sigma).
The solution was vortexed and incubated for 2 hr in a 40°C water bath. Following
incubation, 1 mL of 10% SDS was added and the solution vortexed to solubilize the
protein. Finally, 0.5 mL of 1 N HCl was added and the solution vortexed. The
absorbance was read at 335 nm. The results were determined from a standard curve of
leucine solutions ranging in concentration from 10-3 00 M and expressed as moles
reactive amino groups per mole protein (Adler-Nissen, 1979).

Amino acid analysis was performed by the method of Heinrikson and Meredith

(1984) using high performance liquid chromatography following sample derivatization of

4O
native and hexanal-modified CSA with phenylisothiocyanate. The results were expressed

as a molar percentage for each amino acid.

3.3.2 Protein Labeling by Reductive Alkylation. The number of adducts formed
between hexanal and CSA was determined by measuring the incorporation of tritium from
radiolabeled sodium cyanoborohydride ([3H]NaCNBH3; EC5911, Amersham Corp,
Arlington Heights, IL) (Klassen et al., 1990). CSA (1.5 mg) and hexanal (2.5 uL) were
combined with 87.5 uL of PBS. The protein solution and 10 [1L of 0.1 N NaOH were
added directly to the vial containing the isotope to prevent unnecessary handling and loss.
The reactant concentrations were 15 mg/mL CSA, 200 mM hexanal, and 4 mM
[3H]NaCNBH3. The reaction was allowed to continue for 12 hr at ambient temperature.
The solution was then dialyzed using a Slide-A-Lyzer cassette (66415, Pierce Chemical
Co., Rockford, IL) at ambient temperature against 3 changes of 150 mL of PBS over 24
hr. The solution was diluted 1:100 in PBS and a 10 uL aliquot was combined with 4 mL
of scintillation fluid (Safety-Solve, Research Products International Corp, Mt. Prospect,
IL). Radioactivity was measured using a scintillation counter (Packard 4430, United
Technologies, Downers Grove, IL). The number of adducts formed was expressed as
moles tritium incorporated per mole protein.

3.3.3 Polyclonal Antibody Production. Three 3-month old, female White New Zealand
rabbits (Hazelton Laboratories, Kalamazoo, MI), identified as #45903, 45904, 45911,
were housed by the Michigan State University Laboratory Animal Resources and handled
according to the Rabbit Antibody Production Service (RAPS) program. Each rabbit was
humanized subcutaneously with a total of 1 mL of antigen-adjuvant emulsion administered

in 8-10 sites on its back. The initial injection consisted of 500 pg of hexanal-modified

41
BSA in 0.5 mL of 0.85% sterile saline emulsified with 0.5 mL of Freund’s Complete

adjuvant (F -588 1, Sigma) per rabbit. Emulsification was carried out according to the
procedure of Desrocher (1994). At total volume of 4 mL of antigen solution was
prepared to allow for losses during emulsification and injection. Booster injections were
administered at 4, 8, and 18 weeks and consisted of 500 pg of hexanal-modified BSA
emulsified with F reund’s Incomplete adjuvant (F -5506, Sigma). Each rabbit was bled
from the marginal ear vein prior to the initial injection to obtain control (preirnmune)
serum and 10 days following each injection to check serum titers.

The sera were fiactionated using saturated ammonium sulfate (SAS; pH 7.6)
according to a modification of the method of Harlow and Lane (1988). Briefly, the blood
was allowed to coagulate at ambient temperature for 2 hr and the clot to contract at 4°C
overnight. Next, the blood was decanted into a centrifuge tube and spun at 10,000 x g for
20 min at 4°C. Saturated ammonium sulfate (SAS; pH 7.6) was added dropwise to the
supernatant to achieve 30% saturation (Hebert et al., 1973). The solution was centrifuged
at 10,000 x g for 20 min at 4°C, the pellet dissolved to the original volume in PBS, diluted
2:1 with SAS, and the solution stirred at room temperature for 30 min; this was performed
twice. The solution was centrifuged a final time and diluted in PBS to half the original
volume before dialyzing at 4°C against 4 L PBS for 3 days, changing the buffer every 24
hr. The serum was diluted to the original serum volume and frozen for storage.

3.3.4 Indirect ELISA for serum titer determination. Serum titers for each bleeding
were determined by indirect ELISA. The solid phase support was a 96-well microtiter
plate consisting of disposable Irnmunlon 2 Removawell strips (011-010-6302, Dynatech

Laboratories, Inc., Chantilly, VA) fitted into a Removawell strip holder (011-010-6604,

42
Dynatech). Each plate was coated with 100 uL per well of 2 ug/mL hexanal-modified

CSA in 0.1M carbonate buffer 03H 9.6) and held at 4°C for 16 hr to ensure binding, then
washed four times with PBS-Tween (0.1 M sodium chloride, 0.015 M sodium phosphate,
0.05 % polyoxyethylenesorbitan monolaurate, Tween 20; P-1379, Sigma) using an
automatic plate washer (Model ELP-40, Bio-Tek Instruments, Inc., Winooski, VT). Each
plate was pretreated with 300 11L per well of PBS-casein (0.5% casein; C-7078, Sigma) by
incubation at 37°C for 30 min to prevent nonspecific protein binding (V ogt et al., 1987)
and again washed four times. Each sera was diluted in casein-PBS to concentrations
ranging from 103 to 10° before applying 50 ILL per well to triplicate wells and incubating
for 60 min at 37°C. Following four washes, 100 pL per well of horseradish peroxidase-
conjugated goat anti-rabbit IgG (GAR IgG-HRP; 55679, Organon Teknika Corp, West
Chester, PA) diluted 1:500 in casein-PBS was applied and the plate incubated for 30 min
at 37°C. Following eight washes, a substrate solution consisting of l mL‘of 300 ug/mL
2,2’-azino-bis[3-ethylbenzthiazo-line-6-sulfonic acid] (ABTS; A-1888, Sigma), 11 mL of
citrate buffer (pH 4.0), and 8 [LL of concentrated hydrogen peroxide was prepared and

100 ILL of the solution applied to each well. Bound peroxidase activity was measured
spectrophotometrically at 405 nm using a microplate reader (THERMOmax, Molecular
Devices Corp, Menlo Park, CA) following 30 min of color development at ambient
temperature. Titer was arbitrarily defined as the reciprocal of the highest dilution of serum
to give absorbance twice that of preirnmune serum at the same dilution. Serum fi'om the

second boost of rabbit #45903 was used for all further testing because of its high titer.

43
3.3.5 Antibody Specificity Determination by CI-ELISA. Cross-reactivity testing was

done by competitive indirect (CI) ELISA to determine the protein specificity, aldehyde
specificity, and amino acid specificity of the antibodies. To establish that the antibody
reactivity was not solely with carrier protein, three different proteins were tested in
unmodified and hexanal-modified form. In addition, free hexanal, not attached to a
carrier, was tested for reactivity. Once the protein carrier requirement for recognition by
the antibody was established, further testing was performed using conjugates. To
characterize the “hapten-end” of the epitope, CSA conjugates of various aldehydes,
ketones, or alcohols were tested for cross-reactivity. To characterize the “carrier-end” of
the epitope, hexanal conjugated to various amino acids was tested. Compounds which
were not conjugates, i.e. caproic acid (C-2250, Sigma), tridecanoic acid (T-0502, Sigma)
and free amino acids, were diluted to a concentration of 200 mM in PBS and checked for
reactivity by CI-ELISA. Cross-reactivity was defined as the protein concentration of
hexanal-modified CSA required for 50% inhibition divided by protein concentration of
aldehyde-modified protein for 50% inhibition (Deshpande, 1996).

The competitive assay was the same as the noncompetitive ELISA with three
exceptions: 1) 50 ILL per well of standard or test solution was incubated simultaneously
with 50 1.1L per well of antiserum, 2) the antiserum dilution was 123160, and 3) a standard
curve of hexanal-modified CSA was used. To improve the solubility of hydrophobic free
hexanal, methanol in PBS at 0, 10, 20 and 30% was used for sample dilutions of hexanal
and hexanal-CSA conjugate.

For the protein, aldehyde and carbonyl compound specificity experiments,

conjugate concentrations are given as protein concentration, expressed in micrograms per

44
milliliter. For the amino acid specificity experiments, conjugate concentrations are given

as amino acid concentration, expressed in millimoles per liter. For evaluation of antiserum
reactivity with unmodified and hexanal-modified CSA, hexanal concentration, expressed in
nanograms per milliliter, is used. The hexanal concentration of CSA-hexanal conjugate
was calculated from the protein concentration based on tritium-labeling of available
lysines. All experiments were run in triplicate.

3.3.6 Assay Sensitivity and Precision. To establish the working range of the CI-
ELISA, a standard curve of maximally modified hexanal-CSA adduct was produced by
diluting a stock solution of 15 mg/mL protein to protein concentrations of 0.01, 0.03,
0.10, 0.32, 1.0, 3.2, 10 and 32 ug/mL. Intraplate and interplate assay precisions of the
CI-ELISA were determined within one microtiter plate and among eight plates,
respectively, using hexanal-modified CSA solutions of 0.01, 0.1 and 1 ug/mL protein.
3.3.7 Western Blot. The antibody specificities to unmodified and hexanal-modified
proteins were evaluated by the Western blot method of Wang et al. (1994). Unmodified
and hexanal-modified BSA, CSA and high molecular weight standard (HMW S) protein
mixture were diluted in PBS to protein concentrations of 100, 100 and 250 ug/mL,
respectively. A 0.5 mL aliquot of protein solution was combined with 0.5 mL of SDS-
reducing buffer (0.5M Tris-HCl, pH 6.8, 10% w/v SDS, Z-B-mercaptoethanol), and the
mixture boiled for 5 min prior to loading 10 uL of each onto the gel. The resolving gel
was 12% abrylamide and the stacking gel was 4% acrylamide (Laemmli, 1970).
Electrophoresis was performed using a Mini-Protein Gel assembly (Bio-Rad Laboratories,
Hercules, CA) run at a current of 55 mA and a voltage of 200 V for 1 hr. Transfer of the

proteins from the gel to nitrocellulose membrane (BA-S 85, Schleicher and Schuell,

45
Keene, NH) was performed by electrophoresis for 1 hr at 350 mA and 100 V. The

protein-containing membrane was washed thoroughly with PB S-Tween and blocked for
30 min at ambient temperature with PB S-casein. The membrane was washed thoroughly
and incubated with anti-hexanal antiserum diluted in 0.5% PBS-casein 1:2000 for 30 min.
The membrane was washed thoroughly and incubated with GAR IgG-HRP conjugate
diluted in 0.5% PBS-casein 1:2000 for 10 min. The membrane was washed thoroughly,
incubated with catalyzed substrate solution, and stopped afier 7 min with deionized water.
The substrate solution was prepared by dissolving 24 mg 3,3’,5,5’-tetramethyl benzidine
(TMB; T-2885, Sigma) and 80 mg dioctyl sulfosuccinate (D-0885, Sigma) in 10 mL
ethanol and diluting to 40 mL with 0.1 M citrate bufi‘er (pH 5.0). A 20 u]. aliquot of 30%
hydrogen peroxide was added just prior to incubation to catalyze the peroxidase-substrate

reaction.

3.4 RESULTS AND DISCUSSION

3.4.1 Preparation of Hexanal Conjugates. Following reductive alkylation of CSA with
hexanal, amino acid analysis of the native and hexanal-modified PITC derivatives was
performed. The lysine contents were 11.6 and 0.2 mole % for native CSA and hexanal-
modified CSA, respectively (Table 3.1). No change was observed in the content of any
other amino acid. Similarly, Weisgraber et al. (1978) subjected LDL to reductive
methylation using formaldehyde and compared the amino acid compositions of the native
and modified LDL solutions. They reported modification of 18 of 20 lysine residues and

no changes in any other residues. These findings were expected since, the reducing agent,

46

Table 3.1 Amino acid composition (mole percent) of native and hexanal-modified

 

 

chicken serum albumin (CSA)

Amino Acid Native CSA Modified CSA
ASx 6.3 9.4
GLx 13.1 13.9
SER 6.4 7.2
GLY 5.6 5.1
HIS 2.0 2.7
ARG 4 7 5 5
THR 3.6 4.1
ALA 8 0 8 7
PRO 7.0 7.3
TYR 3.6 4.2
VAL 6.3 7.0
MET 2.1 4.0
ILE 6.0 6.3
LEU 8.0 8.4
PHE 5.7 6.3

LYS 1 1.6 0.2

 

47
NaCNBH3, is highly specific and reduces only Schifi‘bases formed via s-amino groups of

lysine residues and or-amino termini and not aldehydes (Jentofi and Dearbom, 1979).

The amount of tritium incorporated into CSA during protein labeling via reductive
alkylation was 32.6 moles per mole CSA Stoichiometrically, this corresponds to 32.6
moles of lysine or hexanal per mole CSA. Using the molar ratio of 32:1 hexanal to CSA,
the hexanal concentration of CSA-hexanal conjugate was calculated fi'om the protein
concentration. The calculation, given molecular masses of 66,300 and 100 daltons for

CSA and hexanal, respectively, was as follows:

hexanal ng= pg CSAx [1 pm_ol_eCSAl x I32 goles hexanal] x 100 hexanal x|1000 ngl
mL mL [66,300 pg CSA] [1 pmole CSA] [1 pmole hexanal] [1 pg]

Since the values determined by tritium-labeling concurred with values of 28-31
moles of modified lysine per mole of CSA as determined by TNBS assay, the TNBS assay
was used in all subsequent experiments to determine modification. The degrees of
modification for the hexanal-BSA conjugates used in antibody production were 97.2%,
97.1%, 97.6%, and 98.4% for the primary injection and first, second, and third boosts,
respectively, as determined by TNBS assay.

3.4.2 Antiserum Titers. Polyclonal antisera which recognize protein-bound hexanal
were successfiilly produced in rabbits. The highest titers achieved, 6.15 x 10'5 and 6.28 x
105, were from the antisera of rabbits 45903 and 45904 following the second boost (Table
3.2). Rabbit 45911 produced its highest serum titer, 1.64 x 10", following the third
boost. Palinski et al. (1990) reported titers of greater than 10" for polyclonal antisera to

MA-modified LDL after a primary and two booster injections. The antisemm collected

Table 3.2

48

Serum titers of polyclonal antibodies against hexanal-modified bovine
serum albumin (BSA) in rabbits1

 

 

Week2 Titer x 10'5
Rabbit Rabbit Rabbit
45903 45904 4591 1
5 2.38 3.07 0.30
9 6.15 6.28 0.49
19 1.00 0.90 " 1.64
76 1.09 0.28 --

 

I The titer of each serum was arbitrarily designated as the maximum
dilution that yielded twice the absorbance of the same dilution of
preimmune serum in an indirect enzyme-linked immunosorbent assay.

2 Subcutaneous booster injections were performed at weeks 4, 8, 18 and
75 using 500 pg of hexanal-modified BSA

49
following the second boost of rabbit 45903 was used for all subsequent testing since it

had one of the highest titers.

3.4.3 Assay Sensitivity and Precision. The assay sensitivity, or limit of detection,
defined as two times the standard deviation of the mean response of the zero standard
(Deshpande, 1996), was determined to be 1.5 nanograms hexanal per milliliter using
hexanal-modified CSA in the CI-ELISA. The working range, or linear portion, of the
assay for hexanal-modified CSA was 0.03-3 pg/mL protein. A standard curve of hexanal
concentration versus absorbance was generated by substituting hexanal concentration for
protein concentration based on the amount of hexanal bound during reductive alkylation
as determined by tritium incorporation (Figure 3.1). The intraplate coefficient of
variation (CV) values for hexanal-CSA adduct were 4.0, 8.4 and 13.9% (n = 21) for
protein concentrations of 0.01 , 0.1 and 1 pg/mL, respectively. The interplate CV values
for hexanal-CSA adduct among 8 plates were 6.6% (n = 94), 8.4% (n = 93) and 13.0% (n
= 94) for protein concentrations of 0.01, 0.1 and 1 pg/mL, respectively.

3.4.4 Protein Specificity. CSA, KLH and BSA were substantially modified by reductive
alkylation with hexanal as indicated by loss of reactive amino groups as measured by
TNBS assay. The degrees of modification of hexanal-modified CSA, KLH, and BSA
were 95.3%, 86.8%, and 94.8%, respectively. By competitive ELISA, the antiserum did
not show recognition of the unmodified forms of CSA, KLH or BSA, yet did bind to each
of the hexanal-modified proteins (Figure 3.2). Lack of recognition of unmodified BSA,
the immunization carrier protein, suggests that the protein was modified sufficiently to
present primarily hexanal moieties as the antigenic determinants. Hexanal-modified KLH

required a concentration of approximately 30 times that of modified CSA or modified

50

 

 
   

 

 

 

2.00
y = -0.337Ln(x) + 2.0254

g 1.50 . R2=0.9956
V)
O
S;
Q)
2
m 1.00 r
.D
I—
O
U)
.D
<

0.50 .

0.00 L...qu I L....... , ,.,,L,

1 10 100 1000

Hexanal Concentration (ng/mL)

Figure 3.1 Standard curve of hexanal-modified chicken serum albumin (CSA). Hexanal
concentration determined from protein concentration, based on amount of
tritium incorporated into hexanal-CSA during reductive alkylation.

51

 

 

 

 

 

 

 

 

 

 

 

 

e A. "
3‘— .‘ ‘4- .4 V i‘
‘:——————"’ -
1.20 - " \
E 1.00 .
V)
o
it
8 0.80 -
5
.o
S
g 0.60 - CS A
+CSA-l—Iexanal
0-40 - —a—BSA
+BSA-Hexanal
0.20 » -°-KLH \
+KLH-Hexanal
0.00 4 - - 4
0.01 0.1 1 10

100

Protein Concentration (ug/mL)

Figure 3.2 Protein specificity of antibodies to hexanal-modified bovine serum albumin
(BSA) was determined by CI—ELISA. Microtiter wells were coated with 2
pg/mL hexanal-modified chicken serum albumin (CSA) and incubated with
antiserum diluted 1:3160. Unmodified and modified CSA, BSA and keyhole
limpet hemocyanin (KLH) competed at concentrations of 0.01, 0.1, 1, 10 and
100 pg/mL. Determinations were performed in triplicate.

52
BSA to achieve 50% binding inhibition, likely due in part to the lower degree of amino

group modification and in part to fewer lysine derivatives per mole protein.

A Western blot of unmodified and hexanal-modified BSA, CSA and a HMWS
protein mixture showed antiserum recognition of the hexanal-modified proteins but not the
unmodified forms (Figure 3 .3), confirming the results of the ELISA. Hexanal-modified
proteins reacted with antiserum regardless of molecular weight. For the modified HMWS,
bands are visible for myosin (205 kD), B-galactosidase (116 kD), phosphorylase b (97.4
kD), BSA (66 kD) and 0A (45 kD) suggesting the immuno-dominant epitope is a
structure common to all hexanal-modified proteins.

Free hexanal, at concentrations of 7, 70, and 700 ng/mL, did not inhibit antibody
binding regardless of the methanol concentration (Figure 3.4). Methanol at concentrations
of 0, 10, 20 and 30% was added to improve the solubility of unbound hexanal since
hexanal-protein conjugate in 10, 20 and 30% methanol showed little difference in binding
from the control (0% methanol). While methanol did not appear to interfere with the
assay, it did not increase free hexanal reactivity. Since antibodies can only be made to
haptens by first coupling them to a protein carrier, in many cases the antibody elicited will
not react with the free, native molecule, but only with the conjugated version (Morris,
1995). Our finding suggests the antiserum recognizes an epitope larger that the six-carbon
hexanal molecule.

3.4.5 Aldehyde Specificity. CSA was modified by reductive alkylation with various
aldehydes (Table 3.3). Modification, expressed as percent loss in reactive amino groups,
was at least 90.8%. This finding eliminates the possibility that difi‘erences in epitope

density caused by varying degrees of modification are responsible for differences in

53

' v"
V V *
S’ :3 (r
’1? a? "’
Io lo 3‘
v viz. v vi. 3’;

6” 6° a? a? 11’ ~z~
ii“
.' - I

H I“ «4.00m
' 1

W

6

 

 

 

Figure 3.3 Western blot of unmodified and hexanal—modified chicken serum albumin
(CSA; Lanes 1 and 2), bovine serum albumin (BSA; Lanes 3 and 4) and high
molecular weight standard (HMW S; Lanes 5 and 6) containing myosin (205
kD), B-galactosidase (116 kD), phosphorylase b (97.4 kD), BSA (66 kD) and
ovalbumin (45 kD).

54

 

 

 

 

 

 

 

 

 

2.20
2.00 -
1.80 - Q
1.60 -
E 1.40 . \
V3
S?
v 1.20 . ‘~
Q)
a \
e 1.00 - .
g +Control
< 030 " +10% MeOH
+ 20% MeOH
0'60 ' + 30% MeOI-l
0 40 -1:1— 0% MeOH
' ' -o— 10% MeOH
0 20 F —a— 20% MeOH
' -o— 30% MeOH
0.00 444...;x1 n A a IAHAI 1 A n TARA;
l 10 100 1000
Hexanal Concentration (ng/mL)

Figure 3.4 Specificity of antibodies to hexanal-modified bovine serum albumin was
determined by CI-ELISA. Microtiter wells were coated with 2 pg/mL
hexanal-modified chicken serum albumin (CSA) and incubated with antiserum
diluted 1:3160. Free hexanal (open symbols) or CSA-conjugated hexanal
(closed symbols) competed in 0, 10, 20 or 30% methanol. Hexanal
concentrations of conjugates were based on tritium labeling results.
Determinations were performed in triplicate.

55

Table 3.3 Concentration of reactive amino groups in native and modified chicken
serum albumin (CSA) solutions as measured by trinitrobenzene sulfonic

 

 

acid (TNBS) assay
Solution Reactive amino groups Lossl
(mole NHz/mole CSA) (%)

Native CSA 31.0 --
Propanal-CSA 2.2 92.9
Butanal-CSA 2.3 92.5
Pentanal-CSA 2.1 93.1
Hexanal-CSA 2.3 92.6
Heptanal-CSA 2.0 93.4
Octanal-CSA 2.3 92.7
Nonanal-CSA 2.9 90.8
2-t-hexenal-CSA 0.6 98. 1
2-Methylpentanal 0.9 97.0
2-Methylbutanal 0.9 97.0
3-Methylpentanal 1 .2 96.2

 

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

56
binding inhibition among the conjugates. Antiserum cross-reactivity occurred to some

extent with all of the aliphatic aldehydes tested, except propanal, and appeared to increase
as the number of carbons approached six (Figure 3.5). Heptanal-modified CSA showed
the highest cross-reactivity at 86.3% followed by pentanal- and octanal-modified CSA at
11.8% and 1.9%, respectively (Table 3.4). The differences in antiserum reactivity suggest
that antibodies can discriminate between differences in the carbon number of the aldehyde.
Previous reports support the same conclusion. Klassen et al. (1990) reported that
antiserum to acetaldehyde-modified tubulin showed decreasing recognition of aldehyde-
modified BSA with increasing aldehyde carbon number with relative absorbances of 1.000,
0.802, and 0.531 for acetaldehyde (2C), propanal (3C), and butanal (4C) adducts,
respectively. Perata et al. (1991) reported that antiserum to acetaldehyde-modified KLH
exhibited the following cross-reactivities with aldehyde-modified BSA: 6.6, 100, 106, 0,
and 0% for formaldehyde, acetaldehyde, propanal, butanal, and pentanal, respectively. In
another study, an antiserum to carbamylated-LDL was produced (Steinbrecher et al.,
1984) which recognized homocitrulline (e-carbamyl-diaminocaproic acid), a lysine
derivative. Homocitrulline was highly competitive with carbamylated-LDL while citrulline
(O-carbamyl-diaminopentanoic acid) containing one fewer carbon was a very weak
competitor. Cross-reactivity of our antibodies with heptanal and pentanal should not
compromise the usefulness of an ELISA for monitoring lipid oxidation for two reasons.
First, the concentrations of either heptanal and pentanal in chicken may be 30 times less
than that of hexanal in oxidized lipid-containing samples (Ajuyah et al., 1993). Second,
both aldehydes are lipid oxidation products, so cross-reactivity with these aldehydes will

not make the assay nonspecific for lipid oxidation.

57

 

 

 

 

 

 

 

 

 

100
+ 6C-CSA
—c1— 7C-CSA
1 -o— 5C-CSA
+ 8C-CSA
A -e— 4C-CSA
23 —a— 9C-CSA
E —o— 3C-CSA
:—§
.5 50
2'0
.5
"U
.S
m
f
i
0
0.01 0.1 1 10 100 1000

Protein Concentration (ug/mL)

Figure 3.5 Aliphatic aldehyde specificity of antibodies to hexanal-modified bovine serum
albumin was determined by CI-ELISA. Microtiter wells were coated with 2
pg/mL hexanal-modified chicken serum albumin (CSA) and incubated with
antiserum diluted 1:3160. Aldehyde-modified CSA competed at protein
concentrations of 0.01, 0.1, 1, 10, 100 and 1,000 pg/mL. Aldehydes were:
propanal (3C), butanal (4C), pentanal (5C), hexanal (6C), heptanal (7C),
octanal (8C) and nonanal (9C). Determinations were performed in triplicate.

58

Table 3.4 Cross-reactivity of polyclonal antibodies against hexanal-modified bovine
serum albumin with aldehyde-modified chicken serum albumin (CSA) by
indirect competitive ELISA

Aldehyde Chemical Cross-
Component Structure Reactivityl
(%)
Hexanal CSA-NH(CH2)5CH3 100.0
2-t-Hexenal CSA-NHCHCH=CH(CH2)2CH3 100.0
Heptanal CSA-NH(CH2)6CH3 86.3
Pentanal CSA-NH(CH2)4 CH3 1 1.8
3-Methylbutanal CSA-NH(CH2)ZCH(CH3)CH3 2.2
Octanal CSA-NH(CH2)7CH3 1.9
2-Methylpentanal CSA-NHCHZCH(CH3)(CH2)2CH3 1.1
Butanal CSA-NH(CH2)3CH3 0.3
Nonanal CSA-NH(CH2),,CH3 0.1
Propanal CSA-NH(CH2)2CH3 <0. 1
2-Methylbutana1 CSA-NHCH2CH(CH3)CH2CH3 <0.1

 

 

 

 

I Defined as concentration of hexanal-CSA required for 50% inhibition
divided by concentration of aldehyde-CSA required for 50% inhibition
multiplied by 100

59
The competitive abilities of other carbonyl compounds were considered as well.

CSA conjugates of the Strecker degradation products, 2-methylbutanal, 2-methylpentanal,
and 3-methylbutanal, were evaluated for cross-reactivity. These branched aldehyde
conjugates were considerably less efi‘ective competitors than the aliphatic ones (Figure

3 .6). The cross-reactivities of 3-methylbutanal- and 2-methylpentanal-modified CSA were

.fl

2.2% and 1.1%, respectively (Table 3.4). In addition, conjugates of the ketone, 2-

heptanone, and the alcohols, n- and 2-hexanol, were evaluated. CSA modified with 2-

.F.rl 1.1..“
1‘ .

heptanone proved to be a weak competitor, in part, due to modification of only 83.6% as
measured by TNBS assay. The alcohols were unreactive with the antibodies, likely due to
their lack of a carbonyl group preventing reaction with CSA during reductive alkylation.
The monounsaturated six-carbon aldehyde, 2-trans-hexenal, showed cross-reactivity of
100%, likely due its length and to reduction of its double bond during reductive alkylation.
3.4.6 Amino Acid Specificity. Hexanal'conjugates of the amino acids, glycine, lysine
(or-,e-diaminocaproic acid) and arginine, and the lysine derivative, e-aminocaproic acid
(ACA), were evaluated for competitiveness by CI-ELISA. In addition, y-aminobutanoic
acid (GABA) was modified with octanal to achieve a molecule of length equal to that of
hexanal-modified e-ACA. The difference in structures was in the position of the nitrogen
atom (Figure 3.7). While neither the free amino acids nor free a-ACA were not
recognized by the antiserum, hexanal-modified e—ACA was a highly effective competitor
and, in fact, a better competitor than hexanal-modified lysine (Figure 3.8). The
concentration required for 50% antibody binding inhibition was 1091 mM for lysine-
hexanal conjugate compared to 234 mM for e-ACA-hexanal conjugate (Table 3.5).

Palinski et a1. (1989) reported that MA-modified e-ACA and MA-modified t-

60

 

 

 

 

 

 

 

 

 

100
+ 2-t-HEX-CSA
+ 3-MB-CSA
—e— 2-MP-CSA
+ 2-MB-CSA
e\: —x— 2-K-CSA
r:
.2
:5.
TE" 50
”50
.E
1::
.S
m
t t
b /—
0 " _
0.001 0.1 10 1000

Protein Concentration (ug/mL)

Figure 3.6 Carbonyl compound specificity of antibodies to hexanal-modified bovine
serum albumin was determined by CI-ELISA. Microtiter wells were coated
with 2 pg/mL hexanal-modified chicken serum albumin (CSA) and incubated
with antiserum diluted 1:3160. Modified CSA competed at protein
concentrations of 0.001, 0.01, 0.1, 1, 10, 100 and 1,000 pg/mL. Modifying
compounds were: 2-irans-hexenal (2-t-I-IEX-C SA), 3-methylbutanal (3 -MB-
CSA), 2-methy1pentanal (2-MP-CSA), 2-methylbutanal (2-MB-CSA) and 2-
hexanone (2-K-CSA). Determinations performed in triplicate.

61

.28 20:802.: 28 How
205392.:5-» Locum—608-3580 .23 omen—80:68-» “636084386: ”3:309:00 88a 58:5 .«o $5885 Em 8&5

Boa 20:833-:

IOOO-~IO-~IO-~IONIOAIORIOAIORION...ORIORIOAIO

Eon BogSEEBAEEmw «0360649880

IOOOAIUAIOkIOiZRI0-~IO-~IO-~I0-~IO-~IORIOAIO

Eon 3230058920:on BEEEEBGSBI

IOOOAIO-~IO-~IO-~IO-~IO-IZRIORIUAIONIOAI05:0

62

 

lOO

 

-I— ACA-Hexanal
—O— GABA-Octanal
+ LYS-Hexanal

+ GLY-Hexanal

 

 

 

 

Binding Inhibition (%)

 

 

 

0.001 0.01 0.1 l 10

Amino Acid Concentration (mM)

Figure 3.8 Acid specificity of antibodies to hexanal-modified bovine serum albumin was
determined by CI-ELISA. Microtiter wells were coated with 2 ug/mL
hexanal-modified chicken serum albumin and incubated with antiserum diluted
123160. Conjugates competed at concentrations of 0.001, 0.01, 0.1, l and 10
mM. Conjugates were: e-aminocaproic acid (ACAyhexanal, y-aminobutan-
oic acid (GABA)-octanal, lysine (LYS)-hexanal, glycine (GLY)-hexanal.
Determinations performed in triplicate.

63

Table 3.5 Reactivity of polyclonal antibodies against hexanal-modified bovine serum
albumin with aldehyde-amino acid conjugates by competitive indirect ELISA

 

 

Concentration Required
Conjugate for 50% Inhibition
(mM)
s-Aminocaproic acid-Hexanal 234
y-Aminobutanoic acid-Octanal 260
Lysine-Hexanal l 090

Glycine-Hexanal 3090

 

64
butoxycarbonyllysine could compete, binding antibodies to MA-modified LDL. The lower

reactivity of modified lysine may be due to steric hindrance caused by the proximity of the
a- and e-amino groups. Hexanal-modified glycine and arginine were considerably less
competitive. Their reactivity is likely attributable to modification of the a—amino groups
which were not blocked. Jentofi and Dearborn (1979) used nine Na-acetyl amino acids to
determine which amino acids had side chains capable of forming stable derivatives with
formaldehyde under reductive methylation conditions and concluded that arginine,
histidine, methionine and tryptophan did not. They found that tyrosine, serine, asparagine
and cysteine could form stable derivatives but at a reaction rate at least 1000-fold slower
than Nu-acetyllysine.

Our experimental findings indicate that our antiserum recognizes protein-bound
hexanal regardless of the protein carrier and can discriminate difi'erences as small as one
carbon in the aldehyde chain-length. We propose that the immunodominant epitope is a
hexanal-modified lysine residue. Tuma et al. (1987) examined the effect of reaction
conditions on the types and relative amounts of products formed between acetaldehyde
and BSA. When sodium cyanoborohydride was used, the predominant product was the
acetaldehyde-derivatized lysine, N—ethyllysine. Two additional compounds, formed in
much higher amounts when a reducing agent was not used, were found in minimal
quantities when cyanoborohydride was used. While the molecular structures of these two
adducts were not established, they were determined to be modified lysine residues. Our
results suggest that the modification to the lysine residues occurring during reductive
alkylation is the addition of a hexanal molecule producing a hexanal-modified lysine

residue. While the length of the epitope (number of atoms) seems to significantly

65
influence the ability of the antibodies to recognize the structure, the presence of a nitrogen

atom may also be important. In a comparison of hexanal-modified a-ACA and octanal-
modified GABA, structures with the same chain length but with the nitrogen atom at a
different position, the two compounds compete equally well (Figure 3.8). However, the
l3-carbon carboxylic acid, tridecanoic acid, was not reactive suggesting the requirement
of a nitrogen atom for reactivity with these antibodies.

The results of this study demonstrate that sensitive and specific polyclonal
antibodies that react primarily to hexanal-modified lysine adducts can be produced.
Development of an ELISA incorporating these antibodies to monitor lipid oxidation in

biological systems by measuring the hexanal content, therefore, seems feasible.

 

I.”

CHAPTER 4
ELISA TO MONITOR LIPID OXIDATION IN A MEAT MODEL SYSTEM

BY MEASUREMENT OF HEXANAL CONTENT

4.1 ABSTRACT

Lipid oxidation in fresh, frozen or cooked meat is generally associated with the
development of objectionable flavors and odors and, therefore, can be responsible for
limiting shelf-life or loss of consumer acceptability. Hexanal content has been used as an
indicator of the extent of lipid oxidation in meats and other foods. The purpose of this
study was to demonstrate the potential use of a hexanal-specific antibody-based ELISA to
measure hexanal concentration in muscle, by correlating ELISA results to. results
determined by gas chromatography.

Two competitive indirect (CI) ELISAs were designed using anti-hexanal-BSA
antibodies: one using polyclonal antibodies, developed in a previous study, and one using
a monoclonal antibody, developed in a separate study. First, each assay was tested for its
ability to detect differences in the hexanal content of solutions of ovalbumin and chicken
muscle homogenate spiked with hexanal to difl‘erent concentrations. Second, both assays
were used to monitor lipid oxidation in a meat model system by measuring the hexanal
content of chicken thigh homogenate subjected to an accelerated oxidation at 50°C for 36

hr. Because the antibodies were shown in previous studies to recognize hexanal-lysine

66

67
adducts and not free hexanal, a reductive alkylation procedure was incorporated into

sample preparation prior to testing by ELISA. Third, the hexanal contents in the meat
model system as determined by each ELISA were compared with the results of a dynamic
headspace gas chromatography (HS-GC) method using Tenax traps and a thiobarbituric
acid-reactive substances (TBA-RS) assay. The polyclonal antibody-based CI-ELISA
showed strong correlations with both the HS-GC and TBA-RS methods with coefiicients
of 0.89 and 0.85, respectively. The monoclonal-based ELISA gave similar results. The
results of our study indicate that use of our ELISAs to monitor lipid oxidation in chicken

thigh homogenate is a faster and simpler alternative to GC.

4.2 INTRODUCTION

Lipid oxidation occurring in fi'esh, frozen or cooked meat is generally associated
with the development of rancid flavors and odors and a concomitant reduction in the
acceptability of meat (Gray et al., 1994). Therefore, the potential for lipid oxidation and
the development of rancid flavor is one of the single greatest constraints in determining the
processing and shelf-life characteristics of muscle foods (Allen and Foegeding, 1981).
While hydroperoxides, the primary products of lipid oxidation, are odorless and tasteless,
their degradation leads to the formation of complex mixtures of low-molecular weight
compounds with distinctive odor and flavor characteristics (Shahidi and Peg, 1994). The
flavor significance of a particular volatile compound depends on the combination of its
individual threshold value and its concentration in the sample (Gray et al., 1994).
Aliphatic breakdown products are significant to meat flavor due to their low odor

threshold values and increasing concentrations during oxidation (Shahidi and Peg, 1994).

68
One of the most important precursors of aldehyde compounds is linoleic acid due to its

high susceptibility to oxidation, abundance in foods (Kochhar, 1993) and ubiquity in
muscle tissue (Shahidi, 1994). Hexanal is the only aldehyde to autoxidatively arise fiom
both linoleic acid hydroperoxides and some of the unsaturated aldehyde products
(Schieberle and Grosch, 1981). Hexanal has an odor threshold of 5 ng/mL in water
(Buttery et al., 1988) and 6 ug/mL in cooked, lean ground beef (Brewer and Vega, 1995).
The ultimate criterion for suitability of any compound as an index of lipid oxidation
is its adequate correlation with sensory data (Shahidi, 1994). Hexanal content has been
shown to correlate well with off-flavor development during storage of cooked ground beef
(St. Angelo et al., 1987), cooked ground pork (Shahidi et al., 1987), cooked chicken
breast and thigh (Ang and Lyon, 1990) with coeflicients of 0.80, 0.98, 0.92 and 0.84,
respectively. Hexanal has been used successfully to follow lipid oxidation in meat
products (Gray et al., 1994); however, the bulk of the reported studies evaluated cooked
meats. Hexanal concentrations may also be used for evaluating frozen raw and cured-
meat products where oxidation proceeds slowly (Shahidi and Pegg, 1994). Stoick et al.
(199]) compared changes in hexanal concentrations of restructured beef steaks during 6
months of frozen storage to evaluate the antioxidant efficacy of several different
substances. Hexanal was one of several volatile compounds used to compare the extent of
lipid oxidation among frozen ground pork packaged in different materials (Brewer et
al., 1992) and between fresh farmed-raised versus wild alligator tail meat (Cadwalladar et
al., 1994). Analysis of the volatile compounds of fresh irradiated chicken showed
increased hexanal content with increased radiation dose (Hansen et al., 1987). Any degree

of oxidation occurring in raw tissue can accelerate the development of oxidized ofilflavors

69
in cooked products due to the free radical chain reaction mechanism of autoxidation

(Rhee, 1988).

Hexanal content has potential for use as an indicator for quality control purposes
during processing and storage of meat products (Shahidi and Pegg, 1994). Unfortunately,
static headspace techniques, while simple, rapid and reproducible, often lack sensitivity
since insuflicient quantities of indicator compounds are obtained (Reineccius, 1996).
Improved sensitivity is achieved by use of dynamic, or purge-and-trap, techniques which
use an adsorbent material, such as Tenax, to collect volatiles over a period of time but are,
therefore, not rapid methods. As an analytical tool, the immunoassay has already proven
to be very usefirl for the assessment of a variety of substances in meat products, including
various meat species, non-meat proteins, pesticides, and microorganisms (Fukal, 1991).
Immunoassays are a rapid, cost-effective, and easy-to-use alternative to conventional
analytical techniques achieving sensitivity and specificity without requiring highly trained
analysts or sophisticated equipment (Samarajeewa et al., 1991). In addition, immunoassay
kits are portable and can be used for routine surveillance in a processing facility or retail
outlet.

This study was undertaken to develop a novel immunoassay for monitoring lipid
oxidation in meat and other foods and, in particular, to replace more labor-intensive and
time-consuming headspace gas chromatography (HS-GC) methods. We reported the
production and characterization of polyclonal (Chapter 3) and monoclonal antibodies
(Zielinski, 1997) to hexanal-modified bovine serum albumin (BSA). Both were found to
be sensitive and specific for protein-bound hexanal independent of the carrier. This study

consisted of three specific objectives. The first objective was to show that both the

70
polyclonal-based and monoclonal-based assays could detect differences in the hexanal

content of solutions of ovalbumin and chicken muscle homogenate spiked with hexanal to
different concentrations. The second objective was to monitor lipid oxidation in a meat
model system using both ELISAs to measure the hexanal content. Because the antibodies
recognize hexanal-lysine adducts and not fiee hexanal, a reductive alkylation procedure
was incorporated into sample preparation prior to testing by ELISA. Sample preparation
of muscle tissue, as part of the meat model system, consisted of three steps: 1) tissue
homogenization and protein extraction, 2) accelerated oxidation, and 3) reductive
alkylation. The third objective of this study was to compare the hexanal contents in the
meat model system determined by each ELISA with the results of a dynamic headspace
gas chromatography (HS-GC) method using Tenax traps and a thiobarbituric acid-reactive

substances (TBA-RS) assay.

4.3 MATERIAL AND METHODS

4.3.1 Preparation of Muscle Homogenate. The first step in the preparation of muscle
tissue samples was to produce a protein extract homogenate as part of a meat model
system. Homogenates were prepared from chicken thigh muscle and from boneless,
skinless chicken breast purchased fi‘om a local retail store. For the thigh muscle
homogenate, the muscle was deboned and frozen at -80°C with skin and fat retained. The
frozen muscle was shattered with a hammer to pieces of a 2 cm diameter or less and
combined with equally sized pieces of dry ice in a Waring blender (Model 1120, Winsted,
CT). The pieces were pulverized to a fine powder with 30 sec of mixing on high speed.

Three volumes of 0.3 M sodium chloride solution were added to extract both myofibrillar

71
and sacroplasmic proteins. The solution was mixed at high speed for 60 sec and strained

through 2 layers of cheese cloth. For the preparation of breast muscle homogenate, the
muscle was trimmed of visible fat and connective tissue and ground twice through a
Hobart Kitchen-Aid food grinder (Model KF-A, Troy, OH) using a 4 mm plate. The
ground muscle was combined with three volumes of 0.3 M sodium chloride solution,
mixed and strained as described above.
4.3.2 Hexanal Modification of Protein Solutions. Solutions of chicken serum albumin
(CSA; A-3014, Sigma), ovalbumin (OA, chicken egg albumin; A-5253, Sigma) and
chicken muscle homogenate were modified by reductive alkylation with excess hexanal to
achieve maximal modification. Maximal modification was defined as greater than 90%
loss of reactive amino groups from the carrier protein. Alkylation of protein was
performed by the method of Steinbrecher et al. (1984) with modifications described in
Chapter 3. Briefly, hexanal (II-9008, Sigma Chemical Co., St. Louis, MO) was added to
the carrier protein diluted in phosphate-buffered saline (PBS; 0.1M sodium chloride,
0.01M sodium phosphate, pH 7.4) and the solution vortexed. Reducing conditions were
created by adding 1 part 800 mM sodium cyanoborohydride (NaCNBH; S-8628, Sigma)
in 0.1N NaOH to 9 parts of the hexanal-protein mixture. The resultant reaction
concentrations were: 200 mM hexanal, 15 mg/mL protein, and 80 mM cyanoborohydride
at pH 8.5-9.0. Each solution was vortexed, held for 12 hr and dialyzed overnight at
ambient temperature against two changes of PBS.

OA and chicken thigh homogenate were also differentially modified with hexanal
as described above with the following exceptions. To prepare differentially-modified

proteins, aliquots ofl.23, 2.46, 4.92, 7.38, 9.84, 12.3, 18.5, 24.6, 30.8 and 246 LIL of

72
hexanal were added to solutions of 10 mg/mL protein to yield solutions of 1, 2, 4, 6, 8,

10, 15, 20, 25 and 200 mM hexanal and alkylation performed as described above. The
hexanal concentrations were those prior to dialysis since unbound hexanal would have
been lost in the dialysate. The degree of modification of each protein was calculated as a
percent loss based on reactive amino group concentrations before and after reductive
alkylation, as determined by trinitrobenzenesulfonic acid (TNBS) assay (Habeeb, 1966).
Prior to analysis by ELISA, the maximally-modified solutions of CSA, OA and
chicken breast muscle homogenate were diluted to achieve protein concentrations of
0.003, 0.01, 0.03, 0.1, 0.3, 1, 3, 10, 30 and 100 ug/mL. The differentially-modified
solutions of OA and chicken thigh muscle homogenate were diluted 1:500 to a protein
concentration of 20 ug/mL. Each sample set was subjected to CI-ELISA as described in
Chapter 3 and the results were calculated as micrograms hexanal per gram of meat based
on a standard curve of hexanal-modified CSA. All determinations were made in triplicate.
4.3.3 Accelerated Lipid Oxidation Model System. The second step in the sample
preparation of muscle, as part of the meat model system, was to combine 4.8 mL chicken
thigh muscle homogenate with 16.8 mL of PBS in 50 mL polypropylene centrifuge tubes.
Lipid oxidation of chicken thigh homogenate was induced by heating in a 50°C water bath
over a 36 hr period. Sets of three solutions were immersed in the water bath every 6 hr
and held until the first set had heated 36 hr. The last set, time 0, was not heated. Aliquots
of homogenate from each solution were withdrawn for analysis by TBA-RS assay and by
ELISA following reductive alkylation. The remaining solutions were frozen to -20°C in

the tubes in which they were heated until prepared for analysis by gas chromatography.

73
4.3.4 Hexanal Measurement by CI-ELISA. The third and final step of sample

preparation, as part of the meat model system, was reductive alkylation. One milliliter of
the oxidized thigh muscle homogenate solution was combined with 1 mL of 160 mM
cyanoborohydride to reduce covalent bonds formed between the muscle proteins and
carbonyl compounds produced during oxidation of the chicken lipids and phospholipids.
Incubation and 6 hr of dialysis were performed as described previously.

The protein-bound hexanal concentration of each reduced sample was determined
by CI-ELISA. In addition to the polyclonal-based assay, a monoclonal antibody-based
assay was also used. The monoclonal-based CI-ELISA incorporated supernatant from a
hybridoma, designated 45P1-D9, produced by Zielinski (1997) fi'om mice immunized with
hexanal-modified BSA. The monoclonal-based assay was performed in a manner identical
to that of the polyclonal assay except that hybridoma supernatant diluted 1:500 was
substituted for polyclonal antiserum diluted 1:3160 and goat-anti-mouse (GAM)-
horseradish peroxidase (HRP) was used instead of goat-anti-rabbit (GAR)-HRP.
Determinations were made in triplicate. The intraplate precision of the polyclonal-based
assay was determined using hexanal-chicken muscle protein adduct at protein
concentrations of 0. 1, 1 and 10 ug/mL.

4.3.5 TBA-RS Assay. The method of Buege and Aust (197 8) was used to quantify
thiobarbituric acid reactive substances (TBA-RS) in the lipid oxidation model system.
Following accelerated oxidation, a 2 mL aliquot of homogenate was transferred to a 15
mL centrifuge tube containing 4 mL of TBA reagent. The composition of the TBA
reagent was 15% w/v TBA (T-5500, Sigma), 0.375% w/v trichloroacetic acid (TCA;

0414-01, J.T. Baker, Inc, Phillipsburg, NJ) and 0.25 N hydrochloric acid. The samples

74
were capped loosely, boiled for 15 min, and centrifuged at 1000 x g for 15 min at ambient

temperature. The absorbance of the supernatant of each solution was determined
spectrophotometrically at 535 nm. The results Were calculated as milligrams
malonaldehyde per kilogram of meat using a molar extinction coemcient of 1.56 x 10" L x
mole'l x cm". Determinations were made in triplicate.

4.3.6 Isolation and Concentration of Volatile Compounds. Dynamic headspace gas
chromatography (HS-GC) methods of Liu et al. (1992) and Koelsch et al. (1991) were
modified to measure hexanal concentration by adsorption onto Tenax and desorption
using 2-methylbutane. Prior to use, each 10.2 cm metal trap (MSU Biochemistry Shop)
was packed with 200 mg of 60/80 mesh Tenax-TA (04916, Alltech Associates, Inc.,
Deerfield, IL) and each end plugged with silane-treated glass wool (2-0411, Supelco Inc.,
Bellefonte, PA) (Liu et al., 1992). The packed traps were conditioned with nitrogen gas
at a flow rate 25 mL/min at 150°C for at least 12 hr. The conditioned traps were cooled
in a dessicator before being attached via a stainless steel Swagelok reducing union to an
outlet joint of a purge and trap (P&T) apparatus for volatile collection. To volatilize the
flavor compounds, 21 mL of model system homogenate was thawed and combined with
25 mL of PBS in a 500 mL round-bottom flask (605020-1024, Kontes, Vineland, NJ)
connected via an L-shaped glass joint to the P&T apparatus. Each solution was sparged
with a 30 mL/min nitrogen stream while heating to 60°C on a heating pocket for 90 min.
Vercellotti et al. (1992) reported that sparging alone onto Tenax was more efficient at
recovering hexanal than purging or sparging under vacuum. To extract the volatile
compounds fi'om the Tenax, each trap was pushed into a single-hole rubber stopper and

placed in the end of a 15 mL glass graduated centrifuge tube. Using a Pasteur pipet, 2-

75
methylbutane (27,034-2, I-IPLC grade, Aldrich Chemical Co., Inc., Milwaukee, WI) was

added dropwise to the end of the trap until a total of 1.5-2.0 mL was eluted by
centifirgation at 80 x g at 4°C. The extracts were transferred into 2 mL concentrator
tubes (569812-0002, Kontes), concentrated under nitrogen to 1.0 mL, transferred into 2
mL vials with teflon-lined caps (Supelco, Inc., Bellefonte, PA), and stored in a -20°C
freezer. To determine recovery, hexanal was added to 20 mL deionized water and
subjected to the isolation and concentration procedure. The recovery was determined to
be 81.2% (n = 4; c.v. = 4.77%).

4.3.7 Hexanal Measurement by Gas Chromatography. A Hewlett-Packard 5890A
gas chromatograph equipped with a flame ionization detector was used with a DB-225
firsed silica capillary column (30 m x 0.25 mm id, 0.25 pm film thickness; J&W Scientific
Inc., Rancho Cordova, CA) to separate hexanal. The column was held at an initial
temperature of 3 5°C for 10 min, then the column temperature was increased at the rates of
3°C per min to 50°C, 6°C per min to 90°C, and 40°C per min to a final temperature of
200°C. The injection port and detector temperatures were 225°C and 275°C,
respectively. Helium was used as the carrier gas at a flow rate of 40 mL/min with a 3:]
split ratio. Peak integration was performed by a personal computer running Hewlett-
Packard 3365 Series H Chem Station software (HP, Minneapolis, MN). A 10 uL syringe
(Hamilton Co., Reno, NV) was used to inject 2 uL aliquots. The syringe and all solutions
were held at -20°C to limit volatilization prior to injection. Peak retention times of
hexanal reference (Fl-9008, Sigma) and sample solutions were compared to identify the

hexanal elution peaks. The results were calculated as micrograms hexanal per gram of

76
meat using a standard curve of 3, 6, 20, 60 and 200 ug/mL hexanal in 2-methylbutane.

Determinations were made in triplicate.

4.3.8 Statistical Analysis. Statistical analysis was performed using SAS (SAS Institute,
Inc., Cary, NC). The general linear model (GLM) procedure was used to perform analysis
of variance since the data was unbalanced due to laboratory error. Mean comparisons
were performed for the hexanal content data collected. Interactions and contrasts were
examined as well. The data collection methods were GC-HS, polyclonal-based ELISA
(pELISA) and monoclonal-based ELISA (mELISA). Measurements were taken, in
triplicate, every six hours for 36h beginning at time zero. Correlation coefficients were
determined between pairs of methods among GC-HS, pELISA, mELISA and TBA-RS

considering all seven sampling times.

4.4 RESULTS AND DISCUSSION

4.4.1 Hexanal Measurement by Polyclonal Antibody-based CI-ELISA. ReductiVe
alkylation of CSA and CA with hexanal resulted in 98.1% and 97.7% loss of free amino
groups from their native fonns, respectively. Degree of modification for the muscle
homogenate could not be determined due to incomplete solubility. Maximally-modified
CSA, OA and chicken breast homogenate conjugates of 10 mg/mL protein diluted to
protein concentrations ranging from 0.003 to 100 ug/mL showed increasing inhibition of
binding between antibody and plate-bound antigen, as indicated by decreasing
absorbances, with increasing protein concentration (Figure 4.1). An increase in hexanal
concentration was concomitant with an increase in protein concentration because the

hexanal was covalently bound to the protein. The results take the form of a typical

77

 

1.80

1.60 -

1.40 -

1.20 -

l.00 -

0.80 ~

Absorbance (405 nm)

0.60 .

0.40 -

0.20 -

 

0.00

A l

 

\ -l— CBH-Hexanal

 

+ CSA-Hexanal

—o— OA-Hexanal

 

 

Annnnl A A LAILLLI

 

 

0.001

Figure 4.1 Effect of protein (hexanal) concentration on polyclonal antibody binding as
measured by CI-ELISA. Solutions of 10 mg/mL protein of chicken serum
albumin (CSA), ovalbumin (OA) and chicken breast homogenate (CBH) were
maximally-modified with excess hexanal, diluted to protein concentrations of
0.003, 0.01, 0.03, 0.1, 0.3, 1, 3, 10, 30, 100 ug/mL and assayed.

0.01

0.] l 10

Protein Concentration (ug/mL)

Determinations were made in triplicate.

100

1000

 

78
sigrnoidal dose-response curve generated over a range of “doses” with fixed antibody

concentration (Brady, 1995). The protein concentrations required for 50% inhibition
varied among the conjugates, ranging from 0.1 ug/mL for CSA to l ug/mL for the
chicken breast homogenate. While both OA and CSA showed a 98% loss of reactive
amino groups, the absolute number of reactive amino groups likely difiers among the
proteins. The working range was 0.1-10 ug/mL protein for the chicken breast
homogenate versus 0.02-0.5 ug protein per mL for CSA. While the working range of the
ELISA was broader for the chicken breast homogenate, the assay was more sensitive for
CSA-bound hexanal. The intraplate precision for hexanal-chicken protein adduct added at
protein concentrations of 0. 1, 1 and 10 g/mL was 4.6, 4.9 and 7.4%, respectively.
Reductive alkylation with different amounts of hexanal at a constant protein
concentration of 10 mg/mL yielded proteins with different degrees of arrrino group
modification (Table 4.1). The concentration of reactive amino groups decreased as the
amount of added hexanal increased, as measured by TNBS assay. The concentration of
reactive amino groups decreased more than 10 times from an initial concentration of
49,000 M in unmodified 0A to 4,090 M in a 1 uM hexanal-0A solution, resulting in
91.7% modification. As hexanal concentration increased from 1 mM to 25 mM, the
degree of modification of amino groups increased only modestly from 91.7 to 98.2%.
Over the same range of hexanal concentrations, antibody binding inhibition increased fiom
11.4% to 91.7%, showing much greater sensitivity to changes in hexanal concentration.
This may be due to increasing modification of the immunodominant lysine residues with
increasing hexanal addition. Chicken breast homogenate showed the same trend as

indicated by decreasing absorbance (Figure 4.2). Therefore, for samples of similar

 

79

Table 4.1 Comparison of loss of reactive amino groups and antibody binding inhibition
of difi‘erentially-modified ovalbumin-hexanal conjugates

 

 

Hexanal Reactive Amino Amino Group Binding
Concentration‘ Group Concentration2 Modification3 Inhibition4

(mM) (11M) (%) (%)

0 49,000 -- 0

1 4,090 91.7 1 1.4

2 3,720 92.4 17.6

4 3 ,83 0 92.2 36.0

6 3, 1 80 93 .5 66.0

8 2,900 94. 1 79. 1

10 2,340 95.2 84.4

15 1,700 96.5 88.2

20 1,030 97.9 90.6

25 901 98.2 91 .7

 

1 Concentration of hexanal added to solution prior to reduction and dialysis.

2 Concentration based on leucine standard curve detemrined by trinitro-
benzenesulfonic acid (TNBS) assay.

3 Defined as loss of reactive amino groups, determined by TNBS assay.

4 Determined by polyclonal antibody-based competitive indirect-ELISA.

80

 

2.00

1.80 -

 

1.60 - -<>- OA-Hexanal

 

+ CBH-Hexanal
1.40 - 7

 

1.20 -

1.00 -

0.80 -

Absorbance (405 nm)

0.60 l-

0.40 .

0.20 P

 

 

 

0.00 . -4--n...
1 10 100

Added Hexanal Concentration (mM)

Figure 4.2 Effect of hexanal concentration on polyclonal antibody binding as measured
by CI-ELISA. Solutions of 10 mg/mL protein of ovalbumin (OA) and
chicken breast homogenate (CBH) were differentially-modified with varying
amounts of hexanal. The added hexanal concentrations (prior to reduction,
dialysis and dilution) were 2, 4, 6, 8, 10, 15, 20 and 25 mM. The solutions
were diluted to 20 ug/mL protein and assayed. Determinations were made in
triplicate.

8 1
protein concentration modified with different concentrations of hexanal, difi‘erences in

antibody binding inhibition should be detected.

The relationship between the concentrations of hexanal added and of hexanal
bound to the carrier protein and quantified by CI-ELISA was examined. Measured
hexanal concentration, based on a standard curve of CSA-hexanal for which the amount of
hexanal bound was determined previously by tritium-labeling, was substituted for
absorbance (Figure 4.3). Hexanal recoveries, for solutions in the linear portion of the
curve, ranged from 3.1-15.4% and increased with increasing hexanal concentration and
constant protein concentration. Therefore, not all the added hexanal was detected. This
was due, in part, to the loss of some hexanal during dialysis, evident from the odor of the
dialysate. However, given that all ovalbunrin solutions were extensively modified, the
assay may underestimate the amount of hexanal present, at least when hexanal is added at
very high (i.e. millimolar) concentrations. Gutheil and Bailey (1992) studied binding of
hexanal to muscle myofibrillar proteins in aqueous solutions using an static headspace
method and found a similar occurrence. They obtained recoveries of approximately 0.4,
14 and 27% for solutions containing 4.4 mg/mL myosin to which 500, 600 and 700 ug/mL
(5, 6 and 7 mM) hexanal had been added. They also considered the effect of differing
protein concentrations on hexanal binding and concluded that higher myosin
concentrations disproportionately bound more hexanal than lower myosin concentrations,
emphasizing the importance of protein concentration when estimating the actual hexanal
concentration in a sample.

4.4.2 Comparison of a Monoclonal versus Polyclonal Antibody-based CI-ELISA.

Difl‘erentially-modified solutions of chicken thigh homogenate were diluted 1:5000, 1:500

82

 

400

OA-Hexanal

300 -

200 -

100

Measured Hexanal Concentration (ng/mL)

 

 

 

 

0 500 1000 1500 2000

Added Hexanal Concentration (ng/mL)

Figure 4.3 Comparison of added and measured hexanal concentrations for hexanal-
modified ovalbumin solutions. Solutions of 10 mg/mL protein were
combined with varying amounts of hexanal, reduced, dialyzed and diluted
1:500 to a protein concentration of 20 ug/mL. Determinations were made in
triplicate by polyclonal antibody-based CI-ELISA.

 

83
and 1:50 to achieve solutions of low (2 ug/mL), medium (20 ug/mL) and high (200

ug/mL) protein concentrations. In general, the binding curves of both the polyclonal and
monoclonal antibodies were linear from 1 to 10 mMs (100-1000 ug/mL) at low, medium
and high protein concentrations (Figure 4.4) . However, the sensitivities of the assays
varied. At medium concentration, the range of absorbance units was about 1.3 for both
assays. However, at low protein concentration, the polyclonal-based assay spanned 2.0
units, at high concentration, it only covered 0.3 units. The opposite was true of the
monoclonal-based assay which spanned 1.2 units at the high protein. This finding
suggested that the monoclonal-based assay would provide better sensitivity to detect
hexanal in oxidized muscle homogenates, requiring 2000 ug/mL protein to obtain a high
enough concentration of hexanal to detect.

4.4.3 Detection of hexanal in oxidized chicken thigh homogenate. Chicken muscle
homogenate solutions showed increasing hexanal and malonaldehyde equivalent
concentrations, as measured by HS-GC and TBA-RS, respectively, over the first 30 hr
followed by a decline (Figure 4.5). The two ELISAs showed increasing hexanal content
only to 24 hr. Shahidi and Pegg (1994) observed a linear increase in the hexanal
concentration of cooked pork over 6 days of frozen storage followed by a decrease,
possibly due to further oxidation of hexanal to hexanoic acid. Both our CI-ELISAs
successfully provided quantitative results. The polyclonal-based assay gave values ranging
from 0.45 to 2.3 ng/g hexanal; the monoclonal-based assay gave values of 0.97 to 6.4
ng/g. The HS-GC results ranged from 2.0 to 10.4 ng/g. Most of the published works
measuring hexanal in meat report GC peak areas or integrator counts limiting comparisons

of hexanal contents to within a study. However, several authors reported hexanal

84

 

 

  
  
 
  

 

 

 

 

3.00
-O—MAb-Low
2‘50 ' -o—MAb-Med
-O—MAb—High
+PAb—Low
’2‘ 2.00 _ +PAb-Med
m +PAb-High
O
:5
8 1.50 .
5
.D
1—
O
U)
.0
< 1.00 .
0.50 .
0_()0 . . ....... . . ....... . . ...J..

 

 

10 100 l 000 l 0000

Added Hexanal Concentration (ug/mL)

Figure 4.4 Binding curves of differentially-modified chicken thigh homogenate solutions
diluted to low (2 ug/mL), medium (20 ug/mL) and high (200 ug/mL) protein
concentrations. Solutions had added hexanal concentrations (prior to
reduction and dialysis) of 50, 100, 200, 500, 1000, 2000 and 5000 ug/mL.
Solutions were subjected to monoclonal antibody (MAb; open symbols)-based
and polyclonal antibody (PAb; filled symbols)-based CI-ELISA.
Determinations were made in triplicate.

 

85

 

 

   
  
 

 

 

 

 

 

 

 

12 24

L +HS—GC 1
10 - —D—mELISA - 20 ’53
.53 +pELISA E,
E? g -1r-IILAJES 16 I;
= .2
o H
§ 6 - 12 g
t: O
o L)
“—i 3 i

4 . .
55 8
g; 2
I 2
O
2 4 Te‘
2

 

0 12 24 36

Time (hr)

Figure 4.5 Effect of incubation time on hexanal and malonaldehyde concentrations of
chicken muscle homogenate as measured by dynamic headspace gas
chromatography (HS-GC), monoclonal antibody-based competitive indirect
ELISA (mELISA), polyclonal antibody-based competitive indirect ELISA
(pELISA) and thiobarbituric acid-reactive substances (TBA-RS) assay. TBA-
RS results expressed as milligrams malonaldehyde per kilogram meat.
Determinations were made in triplicate.

 

86
concentrations based on grams of meat ranging from 30 ng/g to 30 ug/g suggesting our

results of 400-1 ,000 ng/g are in line. Statistically, each method showed a linear rate of
change in hexanal content over time. The mELISA also showed a quadratic rate of
change in hexanal content over time. The linear rate of change for GC-HS was greater
than that for pELISA. Comparisons were made of hexanal contents determined by
pELISA and mELISA with GC-HS data at each time point. Hexanal contents determined
by mELISA were not significantly different from values determined by HS-GC except at
t = 12h ((1 = 0.05). Hexanal contents determined by pELISA were significantly different
from values determined by HS-GC at all time points (or = 0.05). Therefore, based on the
data collected in a meat model system study, the monoclonal-based ELISA may be a
suitable alternative to HS-GC analysis for hexanal measurement in a meat system.

Each method showed strong correlation with each of the other methods (Table
4.2). The pair of tests showing the highest correlation was HS-GC and TBA-RS (r =
0.97). Ang and Lyon (1990) reported coefiicients of 0.88, 0.90, and 0.92 for cooked and
stored broiler thigh, skin, and breast, respectively, comparing values obtained by HS-GC
(hexanal content) and TBA methods. Other investigators have also found strong
correlations between GC and TBA methods for cooked meats including chicken breast
patties (Ang and Young, 1989; Su et al., 1991), ground pork (Shahidi et al., 1987), beef
loin slices (Hwang et al., 1990) and roast beef (St. Angelo et al., 1987) with coefficients
of 0.86, 0.95, 0.99, 0.81 and 0.92, respectively. Despite the many limitations of TBA
assays, they can be used effectively to monitor and evaluate lipid oxidation in meat and

other biological tissues when used to assess the extent in general, rather than to quantify

87

Table 4.2 Correlation coefficients (r) of measurements of the extent of lipid oxidation
of chicken thigh homogenate. Methods were headspace gas chromatography
(HS-GC), thiobarbituric acid reactive substances (TBA-RS) assay and
competitive indirect (CI) ELISA.

 

 

HS-GC TBA-RS
CI-ELISA (monoclonal) 0.81 0.77
CI-ELISA (polyclonal) 0.89 0.85

HS-GC -- 0.97

 

88
malonaldehyde, or when used in combination with a measurement such as hexanal content

(Gray et al., 1994).

The dynamic HS-GC method used in this study, while sensitive, was very tedious
and time-consuming. The time required for one sample was approximately 2.5 hr, about
the same for both the ELISA and the HS-GC methods. However, the GC method was
limited by the size of P&T set-up available and by the length of the chromatography run,
i.e. the elution time of hexanal, since only one sample can be run at one time. We had a
P&T set-up for 3 samples and a run-time of 25 min. Also, we used external standards to
calibrate requiring additional time. By ELISA, we were able to run 48 samples including a
standard curve, in triplicate, in 2.5 hr, using 2 microtiter plates. Our GC method also
required fastidious laboratory practices, including careful cleaning of the glassware and
extraction in a cold room to prevent evaporative loss due to the low boiling point of the
solvent. Despite the fact that many quality control methods use the dynamic headspace
technique (Vercellotti et al., 1992), these methods may not be practical for routine
procedures in food-processing plants (Ang et al., 1994). In addition, the occurrence of
hexanal-protein binding (Franzen and Kinsella, 1974; Gremli, 1974; Holley et al.,l993;
Gutheil and Bailey, 1992; O’Keefe, 199la,b; Schirle-Keller et al., 1992; Schirle-Keller et
al., 1994; Esterbauer et al., 1987) complicates accurate quantitation by HS-GC methods
which only measure volatile compounds (Jennings, 1977) neglecting nonvolatiles or
volatiles physically bound in the sample (Westendorf, 1985).

The development of objectionable flavors and odors by oxidation has obvious
detrimental consequences on food quality and consumer acceptability (Frankel, 1982).

Hexanal content has been reported to be a sensitive and reliable indicator for evaluation of

89
the oxidative state and flavor quality of meat and meat products (Shahidi, 1994). We have

shown that use of a hexanal-specific antibody-based ELISA to monitor lipid oxidation in

chicken thigh homogenate is a faster and simpler alternative to dynamic HS-GC.

1.

CHAPTERS

CONCLUSIONS

We successfully produced highly sensitive and specific polyclonal antibodies which

react primarily with hexanal-lysine adducts. The irnmunogen was bovine serum ,

 

albumin modified extensively with hexanal by reductive alkylation using sodium
cyanoborohydride as the reducing agent. By selecting a reducing agent specific for
Schiff bases, we were able to limit the products of the protein-hexanal reaction to
lysine derivatives. By using an excess of hexanal, we were able to promote

predominantly hexanal-modified lysine residues as the antigenic determinant.

We found only two compounds to cause cross-reactivity with our polyclonal antiserum
against hexanal-modified bovine serum albumin at a level above 2%. The compounds
were heptanal and pentanal, the aliphatic aldehydes with one greater and one fewer
carbons than hexanal, respectively. Because both compounds are also lipid oxidation
breakdown products, our assay remains specific for monitoring lipid oxidation. In
addition, neither heptanal nor pentanal is generally found in foods at concentrations as

high as the concentrations at which hexanal may occur.

90

91
3. We established that our polyclonal antibody-based immunoassay can measure hexanal

concentrations more rapidly and simply than a dynamic headspace gas chromatography
method. In addition, the ELISA correlated well with both GC and TBA methods.
While not as sensitive as the monoclonal antibody-based assay, the polyclonal-based
assay was able to detect differences in the hexanal concentrations of chicken thigh
muscle homogenate incubated over 36 hr and is a viable alternative method to dynamic

headspace gas chromatography.

85 91

CHAPTER 6

FUTURE RESEARCH

The findings of the preceding study suggest that further research is warranted.
The sensitivity of the polyclonal antiserum may be increased by performing immunoaffinity
purification on an antigen column to select for the antibodies with more affinity for a
particular antigen, such as oxidized chicken protein. Another approach to improve assay
sensitivity is to use a direct competitive format instead of the indirect format. This
requires either: 1) conjugation of hexanal to horseradish peroxidase (HRP) to compete
with antigen-containing extract for plate-bound antibody or 2) conjugation of polyclonal
antibody and HRP for which antigen-containing extract and plate-bound antigen would
compete. Also, firrther optimization of the assay may be attempted, particularly the use
of another substrate, such as 3,3’,5,5’-tetramethylbenzidine (TMB; Deshpande, 1996) or
o-phenylene diamine (OPD; Kemeny, 1991), or another enzyme-substrate pair.

Additional experimentation is necessary to develop a clearer understanding of
hexanal—muscle protein binding. Once a model is established, factors corresponding to
protein concentration can be determined to adjust for hexanal present but not detected.

To continue pursuing application of our ELISA within the meat industry, the next
step would be to move fi'om a model system to an actual muscle or meat product.
Appropriate extraction procedures must be developed for different products, such as raw

or cooked meats, and whole muscle, formed or comminuted products. Use of the assay

92

 

93
should not be limited to the product itself. Raw materials to be used in formed and

comminuted products could also be evaluated for lipid oxidation. Extenders, such as soy
protein, are known to contain hexanal imparting ofllflavor (MacLeod and Ames, 1988).
Other sources of ofllflavor may be in the production process. Chiller water of commercial
poultry processors, used to decrease carcass temperature, has been found to contain
aldehydes, including hexanal (Tsai et al., 1986). However, the latter application would
require a “carrier” for the hexanal since the antibodies do not recognize fi'ee hexanal.

Results of a preliminary study suggest that tris(hydroxymethyl)aminomethane
(Tris) buffer used in place of protein may function as the necessary carrier for hexanal.
Use of Tris buffer could be specified in the protocol as a sample preparation reagent for
low protein or nonprotein samples. This approach may pemrit application of our ELISA
to vegetable oils (Reineccius, 1996) and potato products (Salinas et al., 1994) for which
hexanal has long been an indicator of lipid oxidation.

While several published studies have reported strong correlations between the
development of hexanal in meat and sensory data, the researchers measured volatile
hexanal using gas chromatography. In addition to the fact that the use of an ELISA to
measure hexanal is a novel approach, both the polyclonal-based assay and the monoclonal-
based assay measure total hexanal (volatile and nonvolatile). Therefore, sensory
evaluation should be included in future studies of meat or other foods.

A broader reaching, but nonetheless reasonable application of our assay within the
poultry industry is for assessment of turkey spermatozoa quality. Most turkeys are bred
by artificial insemination because the size of the turkey male attained through genetic

selection makes them unable to perform natural mating (Sexton, 1979). Formation of lipid

 

94
peroxides limits survival of spermatozoa during in vitro storage (Wishart, 1989) and TBA-

RS assays are used to monitor extent of lipid peroxidation (Cecil and Bakst, 1993).
Spermatozoa, particularly fowl and turkey, contain high concentrations of arachidonic
acid, a hexanal precursor (Darin-Bennett et al., 1974). Hexanal measurement as an
alternative to TBA-RS assay to assess extent of lipid oxidation in biological samples has
already been successful (Frankel et al., 1992). In addition, Jones et al. (1979) stated that
measurement of the extent of lipid peroxidation may provide a usefirl basis for biochemical

appraisal of human sperm quality.

 

LIST OF REFERENCES

Adler-Nissen, J. 1979. Determination of the degree of hydrolysis of food protein
hydrolysates by trinitrobenzenesulfonic acid. J. Agric. Food Chem. 27(6): 1256-1262.

Ajuyah, A.O., Fenton, T.W., Hardin, R.T., and Sim, IS. 1993. Measuring lipid oxidation
volatiles in meats. J. Food Sci. 58(2): 270-273, 277.

Allen, JG 1988. Problems associated with developing food immunoassays. Ch. 15 in
Immunoassay; for Veterinmnd Food Analysis—1. BA Morris, M.N. Clifl'ord and R.
Jackman (Ed), p. 183-192. Elsevier Applied Science, London, England.

Allen, CE. and EA. Foegeding. 1981. Some lipid characteristics and interaction in muscle
food—A review. Food Technol. 35(5): 253-257.

Ang, C.Y.W. and Young, LL. 1989. Rapid headspace gas chromatographic methods for
assessment of oxidative stability of cooked chicken meat. J. Assoc. Ofi‘. Anal. Chem.
72(2): 277-281.

Ang, C.Y.W. and Lyon, B.G. 1990. Evaluations of warrned-over flavor during chill
storage of cooked broiler breast, thigh and skin by chemical, instrumental and sensory
methods. J Food Sci. 55(3): 644-648, 673.

Ang, C.Y.W. and Huang, Y. W. 1993. Internal temperature and packaging system afl‘ect
stability of cooked chicken leg patties during refrigerated storage. J. Food Sci. 58(2): 265-
269, 277.

Ang, C.Y.W., Liu, F ., and Sun, T. 1994. Development of a dynamic headspace GC
method for assessing the influence of heating end-point temperature on volatiles of
chicken breast meat. J. Agric. Food Chem. 42(11): 2493-2498.

Arai, S., Noguchi, M., Yarnashita, M. Kato, H., and Fujimaki, M. 1970. Studies on flavor
components in soybean Part VI. Some evidence for occurence of protein-flavor binding.

Agric. Biol. Chem. 34(10): 1569-1573.

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

Badings, 1970. Cold storage deflects in butter and their relation to the autooxidation of
unsaturated fatty acids. Neth. Milk Dairy J. 24: 147-256. [In O’Brien, RI. 1987.

95

96

Biologically active products of lipid peroxidation. Ch. 6 in Autoxidation at Unsaturated
Lipids, H.W.-S. Chan (Ed), p. 264-268. Academic Press, Orlando, FL.]

Bengtsson, B.L., Bosund, I., and Rasmussen, I. 1967. Hexanal and ethanol formation in
peas in relation to ofllflavor development. Food Technol. 21(3): 478-482.

Bett, KL. and Boylston, TD. 1992. Effect of storage on roasted peanut quality:
Descriptive sensory analysis and gas chromatographic techniques. Ch. 19 in @211
Oxidation in Food. A]. St. Angelo (Ed), p. 322-343. ACS Symposium Series 500.
American Chemical Society, Washington, DC.

Boggs, M.M., Buttery, R.G., Venstrom, D.W., and Belote, ML. 1963. Relation of
hexanal in vapor above stored potato granules to subjective flavor estimates. J. Food Sci.
29: 487-489.

Brady, J.F. 1995. Interpretation of immunoassay data. Ch. 19 in Immunoanalxsis ot

Agrochemicals: Emerging T echnologz’es, J .0. Nelson, AE. Karu and RB. Wong (Ed),
p.266-286. American Chemical Society, Washington, DC.

Brewer, M.S., Ikins, W.G., and Harbers, C.AZ. 1992. TBA values, sensory
characteristics, and volatiles in ground pork during long-term frozen storage: Efl‘ects of
packaging. J. Food Sci 57(3): 558-563, 580.

Brewer, MS. and Vega, JD. 1995. Detectable odor thresholds of selected lipid oxidation
compounds in a meat model system. J. Food Sci. 60(3): 592-595.

Buege, J .A. and Aust, SD. 197 8. Microsomal lipid peroxidation: The thiobarbituric acid
assay. Methods Enzymol. 52: 302.

Buttery, R.G., Tumbaugh, J.G., and Ling, LC. 1988. Contribution of volatiles to rice
aroma. J. Agric. Food Chem. 36(5): 1006-1009.

Cadwallader, K.R., Baek, H..,H Chung, H.Y., and Moody, M.W. 1994. Contribution of
lipid-derived components to the flavor of alligator meat. Ch. 13 in Lipids in Food Flavors.
C.-T. Ho and T.G. Hartman (Ed), p. 186-195. American Chemical Society, Washington,
DC.

Cecil, HO and Bakst, MR. 1993. In vitro lipid peroxidation of turkey spermatozoa.
Poultry Sci. 72: 1370-1378.

Chancerelle, Y., Alban, C., Viret, R., Tosetti, F. and Kergonou, J.-F. 1991.
Immunological relevance of malonic dialdehyde (MDA): IV. Further evidences about the
epitope recognized by antibodies obtained from rabbits immunized with MDA-modified
lysozyme. Biochem. Intern. 24(1): 157-163.

97

Chiba, H., Takahashi, N., Kitabatake, N. and Sasaki, K 1979. Enzymatic improvement of
food flavor III. Oxidation of the soybeacn protein-bound aldehyde by aldehyde
dehydrogenase. Agric. Biol. Chem. 43(9): 1891-1897.

Craig, J., Bowers, J .A., and Seib, P. 1991. Sodium tripolyphosphate and sodium ascorbate
monophosphate as inhibitors of off-flavor development in cooked vacuum-packaged,
fi'ozen turkey. J. Food Sci. 56(6): 1529-1531, 1561.

Cross, OK. and Ziegler, J. 1965. A comparison of the volatile fractions from cured and
uncured meat. J. Food Sci. 30(4): 610-614.

Cuppett, S. and Warner, K. 1996. Antioxidant common interest group update. Lipid
Oxidation and Quality Division Newsletter, December. American Oil Chemists’ Society,
Champaign, IL.

Curtiss, L.K. and Witztum, IL 1983. A novel method for generating region-specific
monoclonal antibodies to modified proteins. J. Clin. Invest. 72: 1427-1438.

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

Darin-Bennett, A. 1974. Phospholipids of dog and fowl spermatozoa. J. Reprod. Fertil.
41: 471-474.

Decker, EA. and Hultin, HO. 1992. Lipid oxidation in muscle foods via redox iron. Ch. 3
in Lipid Oxidation in Food. A]. St. Angelo (Ed), p. 33-54. ACS Symposium Series 500.
American Chemical Society Symposium, Washington, DC.

Deshpande, SS. 1996. Assay development, evaluation, and validation. Ch. 9 in En_zyme
Immunoassays: From Concept to Product Development. p. 275-359. Chapman & Hall,
New York, NY.

Desrocher, L. 1994. ELISA for rapid endpoint temperature determination; and effects of
delayed bleeding on residual serum proteins in turkey muscle. MS. Thesis. Michigan
State University, East Lansing, MI.

Donohue, T.M., Tuma, D.J., and Sorrell, MP. 1983. Acetaldehyde adducts with proteins:
Binding of [”C]acetaldehyde to serum albumin. Arch. Biochem. Biophysics 220(1): 23 9-
246.

Drumm, TD. and Spanier, AM. 1991. Changes in the content of lipid autoxidation and
sulfirr-containing compounds in cooked beef during storage. J. Agric. Food Chem. 39(2):
336-343.

98

Dupuy, H.P., Bailey, M.E., St. Angelo, Al, Vercellotti, JR and Legendre, MG. 1987.
Instrumental analyses of volatiles related to warmed-over flavor of cooked meats. In
Med-Over Flavor of Meat. A]. St. Angelo and ME. Bailey (Ed), p. 165-191.
Academic Press, Orlando, FL.

Ekstrand, B., Gangby, I., Akesson, G., Stollman, U., Lingnert, H. and Dahl, S. 1993.
Lipase activity and development of rancidity in oats and oat products related to heat
treatment during processing. J. Cereal Sci. 17: 247-254.

Erickson, MC. 1993. Contribution of phospholipids to headspace volatiles during storage
of pecans. J. Food Quality 16(1): 13-24.

Esterbauer. H., 1982. Aldehydic products of lipid peroxidation. In Free R_adicals. Lipid
Peroxidation and Cancer. D.C.H. McBrien and TE Slater (Ed), p. 101-122. Academic
Press, New York, NY.

Esterbauer, H., Jurgens, G, Quehenberger, O. and Koller, E. 1987. Autoxidation of human
low density lipoprotein: Loss of polyunsaturated fatty acids and vitamin E and generation
of aldehydes. J. Lipid Research 28: 495-509.

Frankel, EN. 1982. Volatile lipid oxidation products. Prog. Lipid Res. 22: 1-33.

Frankel, EN. 1984. Lipid oxidation: Mechanisms, products and biological significance. J.
Amer. Oil Chem. Soc. 61(12): 1908-1914.

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.

Frankel, EN. and AL. Tappel. 1991. Headspace gas chromatography of volatile lipid
peroxidation products fi'om human red blood cell membranes. Lipids 26(6): 479-484.

Frankel, E.N., Hu, M.-L., and Tappel, AL. 1989. Rapid headspace gas chromatography
of hexanal as a measure of lipid peroxidation in biological samples. Lipids 24(11): 976-
981.

Franzen, K.L. and Kinsella, J .E. 1974. Parameters affecting the binding of volatile flavor
compounds in model food systems. 1. Proteins. J. Agr. Food Chem. 22(4): 675-678.

Fritsch, CW. 1994. Lipid oxidation—The other dimensions. Inform 5: 423-436.

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(5): 43 l-
448.

99

Funes, J., Yong, S., and Karel, M. 1980. Changes in lysozyme due to reactions with
volatile products of peroxiding methyl linoeate. J. Agric. Food Chem. 28: 794-798.

Gray. J .1. 1978. Measurement of lipid oxidation: A review. J. Amer. Oil Chem Soc. 55:
539-546.

Gray, 11. and Monahan, F.J. 1992. Measurement of lipid oxidation in meat and meat
products. Trends Food Sci. Technol. 3: 315-319.

Gray, J .1, Pearson, AM. and Monahan, F .J . 1994. Flavor and aroma problems and their
measurement in meat, poultry and fish products. Ch. 10 in Advances in Meat Reseazgg
Volume 9: Quality Attributes and Their Measurement in Meat. Poultry and Fish
Products, A.M. Pearson and TR. Dutson (Ed), p. 250-288. Blackie Academic &
Professional, London, England.

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

Guth, H. and Grosch, W. 1994. Flavor changes of oat meal extrusion products during
storage. In Trends in Flavour Research, H. Maarse and D.G. van der Heij (Ed), p. 395-
399. Elsevier Science, Amsterdam, Netherlands.

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

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.

Haberland, M.E., Fong, D. and Cheng, L. 1988. Malondialdehyde-altered protein occurs
in atheroma of watanabe heritable hyperlipidemic rabbits. Science 241: 215-218.

Hall, G. and Andersson, J. 1985. Flavor changes in whole milk powder during storage. III.
Relationships between flavor properties and volatile compounds. J. Food Quality 7: 23 7-
253.

Halliwell, B. 1995. How to characterize an antioxidant: an update. Biochem. Soc. Syrnp.
61 :73-101.

 

100

Hallberg, ML and Lingnert, H. 1991. Lipid oxidation in potato slices under conditions
simulating the production of potato granules. J. Amer. Oil Chem. Soc. 68(3): 167-170.

Hansen, T.J., Chen, G.-C. and Shieh, J .J . 1987. Volatiles in skin of low dose irradiated
fresh chicken. J. Food Sci. 52(5): 1180-1182.

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

Hartman, T. G., Karmas, K, Salinas, J. P., Ruiz, R., Lech, J. and Rosen, R. T. 1994. Effect
of packaging on the lipid oxidation storage stability of dehydrated pinto beans. Ch. 11 in

W, C. -T. Ho and T. G. Hartman (5d,) p. 158- 167. American Chemical
Society, Washington, DC.

Hebert, G.A., Pelham, PL. and Pittman, B. 1973. Determination of the optimal ,
ammonium sulfate concentration for the fractionation of rabbit, sheep, horse, and goat
antisera. Appl. Microbiol. 25(1): 26-36.

Hefle, S.L. 1995. Immunoassay fundamentals. Food Technol. 49(2): 102-107.

Heinrikson, KL. and Meredith, SC. 1984. Amino acid analysis by reverse-phase high-
perforrnance liquid chromatography: Precolumn derivatization with phenylisothiocynate.
Anal. Biochem. 136: 65-74.

Heizmann, C.W., Muller, G., Jenny, E., Wilson, K.J., Landon, F. and Olomucki, A. 1981.
Muscle B-actinin and serum albumin of the chicken are indistinguishable by
physiochemical and immunological criteria. Proc. Natl. Acad. Sci. USA 78(1): 74-77.

Holley, A.E., Walker, M.K., Cheeseman, K.H., and Slater, T.F. 1993. Measurement of n-
alkanals and hydroxyalkenals in biological samples. Free Radical Biology & Medicine
15: 281-289.

Hsieh, RJ. 1994. Lipoxygenase pathways and food flavors. Ch. 3 in Lipid in £di
EMILE, C.-T. Ho and T.G. Hartman (Ed), p. 30-47. American Chemical Society,
Washington, DC.

Hsieh, R.J. and Kinsella, J.E. 1989. Oxidation of polyunsaturated fatty acids:
Mechanisms, products, and inhibition with emphasis on fish. Adv. Food Nutr. Res. 33:
233-341.

Hwang, S.-Y., Bowers, J .A. and Kropf, DH. 1990. Flavor, texture, color, and hexanal
and TBA values of frozen cooked beef packaged in modified atmosphere. J. Food Sci.
55(1): 26-29.

101

Igene, J. O., Yamauchi, K., Pearson, A. M., Gray, J. 1., and Aust, S. D. 1985. Evaluation of
2-thiobarbituric acid reactive substances (TBRS) in relation to warmed-over flavor
(WOF) development in cooked chicken. J. Agric. Food Chem. 33. 362-367.

Ikeda, K., Higashi, T., Sano, H., Jinnouchi, Y., Yoshida, M., Araki, T., Ueda, S., and
Horiuchi, S. 1996. N‘-(carboxymethyl)lysine protein adduct is a major immunological
epitope in proteins modified with advanced glycation end products of the Maillard
reaction. Biochem. 35(24): 8075-8083.

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, O., Zamel, D., Shami, E., Zywulko, M., Klajner, F.,
and Borgono, C. 1992. Hypersensitivity to acetaldehyde-protein adducts. Molec.
Pharmacol. 42: 711-717.

Jennings, W. G. 1977. Objective measurements of flavor quality: General approaches,

problems, pitfalls, and accomplishments. Ch. 1 in W
W. R. A. Scanlan (Ed. ), p. 1 -7. American Chemical Society, Washington, DC.

Jennings, W.G. and Filsoof, M. 1977. Comparison of sample preparation techniques for
gas chromatographic analysis. J. Agric. Food Chem. 25(3): 440-445.

J entoft, N. and Dearborn, D.G. 1979. Labeling of proteins by reductive methylation using
sodium cyanoborohydride. J. Biol. Chem. 254(11): 4359-4365.

Jeon, 1.]. and Bassette, R. 1984. Analysis of n-pentanal and n-hexanal as indices of potato
chip shelf-life. J. Food Quality 7(2): 97-106.

Jones, R., Mann, T. and Sherins, R. 1979. Peroxidative breakdown of phospholipids in
human spermatozoa, spermicidal properties of fatty acid peroxides, and protective action
of seminal plasma. Fertility and Sterility 1(5): 531-537.

Jourdan, P.S., Mansell, R.L., Oliver, D.G. and Weiler, E.W. 1984. Competitive solid
phase enzyme-linked immunoassay for the quantification of lirnonin in citrus. Anal.
Biochem. 138: 19-24.

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

Kanner, J. 1992. Mechanism of nonenzymic lipid peroxidation in muscle foods. Ch. 4 in

W, A]. St. Angelo (Ed), p. 55-73. ACS Symposium Series 500.
American Chemical Society Symposium, Washington, DC.

102

Kappus, H. 1991. Lipid peroxidation: mechanism and biological relevance. Ch. 4 in Free
Radicals and Food Additives, O.I. Aruoma and B. Halliwell (Ed), p. 59-75. Taylor &
Francis, London, England.

Kemeny, D.M. 1991. W. Pergamon Press, Inc., Elmsford, NY.

Kinderlerer, J .L. and Johnson, S. 1992. Rancidity in hazelnuts due to volatile aliphatic
aldehydes. J. Sci. Food Agric. 58: 89-93.

Kinsman, D.M. 1994. Historical perspective and current status. Ch. 1 in W;

W. D.M. Kinsman. A.W. Kotula and BC.
Breidenstein (Ed), p. 1-24. Chapman & Hall, New York, NY.

Klassen, L. W. ,Xu, D. S., Tuma, D. J. 1990. Immune response to acetaldehyde adducts. In

AlcahaLlnrmrmnmndulatiQanndAIDS Seminara, D Watson. R R., and Pawlowski A
(Ed), p. 333- 340. Alan R. Liss, Inc., New York, NY.

Klassen, L.W., Tuma, D.J., Sorrell, M.F., McDonald, T.L., Devasure, J .M., and Thiele,
GM. 1994. Detection of reduced acetaldehyde protein adducts using a turique
monoclonal antibody. Alcoh.-Clin. Exper. Research 18: 1.

Kochhar, S. P. 1993. Oxidative pathways to the formation of off-flavors. Ch. 6 in M
Manama-flavors, M. J. Saxby (Ed. ) Blackie Academic & Professional, New York,
NY.

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(3): 816-820, 834.

Konstance, R.P., Sullivan, J .F ., Talley, F .B., Calhoun, M.J., and Craig, J .C. 1978. Flavor
and storage stability of explosion-puffed potatoes: Autoxidation. J. Food Sci. 43: 411-
414.

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

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

Lai, S.-M., Gray, J.I., Booren, A.M., Crackel, R.L., and Gill, IL. 1995. Assessment of
off-flavor development in restructured chicken nuggets using hexanal and TBARS
measurements and sensory evaluation. J. Sci. Food Agric. 67(4): 447-452.

103

Larnikanra, VT. and Dupuy, HP. 1990. Analysis of volatiles related to warmed over
flavor of cooked chevon. J. Food Sci. 55(3): 861-862.

Larick, D.K., Turner, B.E., Schoenberr, W.D., Coffey, MT. and Pilkington, DH. 1992.
Volatile compound content and fatty acid composition of pork as influenced by linoleic
acid content of the diet. J. Anim. Sci. 70:1397-1403.

Lillard, DA. 1987. Oxidative deterioration in meat, poultry, and fish. In W
W A.J. St. Angelo and ME. Bailey (Ed) p. 41-67. Academic Press, Inc.,
Orlando, FL.

Lingnert, H. 1980. Antioxidative maillard reaction products 111. Application in cookies.
J. Food Proces. and Preserv. 4: 219-233.

Lingnert, H. 1992. Influence of food processing on lipid oxidation and flavor stability.

Ch. 16 111W, A..J St. Angelo (Ed.,) p. 292- 301. ACS Symposium
Series 500. American Chemical Society Symposium, Washington, DC.

Liu, H.F., Booren, A.M., Gray, J.I., and Crackel, KL. 1992. Antioxidant efficacy of
oleoresin rosemary and sodium trypolyphosphate in restructured pork steaks. J. Food Sci.
57(4): 803-806.

Lopez-Bote, J .P., Langa, C., Lastres, P., Rius, C., Marquet, A., Ramos-Ruiz, R. and
Bemabeu, C. 1993. Aggregated human immunoglobulins bind to modified proteins.
Scand. J. Immunol. 37: 593-601.

Love, J .D. and Pearson, A.M. 1971. Lipid oxidation in meat and meat products. J. Am.
Oil Chem. Soc. 48: 547-549.

MacLeod, G. and Ames, J. 1988. Soy flavor and its improvement. Crit. Rev. Food Sci.
Nut. 27(4): 219-398.

Matsuda, T., Kato, Y., Watanabe, K. and Nakamura, R. 1986. Immunochemical
properties of proteins glycosylated through Maillard reaction: B-lactoglobulin-lactose
and ovalbumin-glucose systems. J. Food Sci. 50: 618-621.

Means, GE. and F eeney, RE. 1995. Reductive alkylation of proteins. Anal. Biochem.
224: 1-16.

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

Min, DB. 1981. Correlation of sensory evaluation and instrumental gas chromatographic
analysis of edible oils. J. Food Sci. 46(5): 1453-1456.

104

Moreira, M.A., Tavares, S.R., Ramos, V., and de Barros, E.G. 1993. Hexanal production
and TBA number are reduced in soybean [Glycine max (L.) Merr.] seeds lacking
lipoxygenase isozymes 2 and 3. J. Agric. Food Chem. 41(1): 103-108.

Monis, R. J. 1995. Antigen-antibody interactions: how affinity and kinetics affect assay
design and selection procedures. Ch. 3 1n WWW

WWW M..A RitterandHHM Ladyman(Ed.,) p 34- 59
Cambridge University Press, Cambridge, England.

Mottram, 1987. Lipid oxidation and flavour in meat and meat products. Food Sci.
Technol. Today 1(3): 159-162.

Muego-Gnanasekharan, K.F. and Resurreccion, A.V.A. 1993. Physiochemical and
sensory characteristics of peanut paste as affected by processing conditions. J. Food
Process Preserv 17(5): 321-336.

Muller, BL and Gautier, A.E. 1994. Green note production: A challenge for

biotechnology. In W, H. Maarse and D.G. van der Heij (Ed),
p. 475-480. Elsevier Applied Science, Amsterdam, Netherlands.

Niemela, O. 1993. Acetaldehyde adducts of proteins: Diagnostic and pathogenic
implications in diseases caused by excessive alcohol consumption. Scand. J. Clin. Lab.
Invest. 53, Suppl. 213: 45-54.

Niemela, O., Juvonen, T. and Parkkila, S. 1991. Immunohistochemical demonstration of
acetaldehyde-modified epitopes in human liver after alcohol consumption. J. Clin. Invest.
87: 1367-1374.

Nonaka, M. and Pippen, EL. 1966. Volatiles and oxidative flavor deterioration in fiied
chicken. J. Agric. Food Chem. 14(1): 2-4.

O'Keefe, S.F., Resurreccion, A.P., Wilson, L.A., and Murphy, P.A. 1991a. Temperature
effect on binding of volatile flavor compounds to soy protein in aqueous model systems.
J. Food Sci. 56(3): 802-806.

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

Okitani, A., Kaneko, S., Tashiro, Y., Hayase, F. and Kato, H. 1986. Polymerization of
proteins and impairment of their amino acid residues due to vaporized hexanal. In Amino:

Wm M- FUJimaki, M. Namiki and H.
Kato (Ed), p. 125-134. Elsevier Science Publishing Co., Inc., New York, NY.

 

105

Olafsdottir, G., Steinke, J .A., and Lindsay, R.C. 1985. Quantitative performance of a
simple Tenax-GC adsorption method for use in the analysis of aroma volatiles. J. Food

Sci. 50(5): 1431-1436.

Palinski, W., Miller, E. and Witztum, J .L. 1995. Immunization of low density lipoprotein
(LDL) receptor-deficient rabbits with homologous malondialdehyde-modified LDL
reduces atherogenesis. Proc. Natl. Acad. Sci. USA 92: 821-825.

Palinski, W. Rosenfeld, M.E., Yla-Herttuala, S., Gurtner, G.C., Socher, S.S., Butler,
S.W., Parthasarathy, S., Carew, T.E., Steinberg, D. and Witztum, J.L. 1989. Low density
lipoprotein undergoes oxidative modification in vivo. Proc. Natl. Acad. Sci. USA 86:
1372-1376.

 

Pearson, A. M, Love, J. D., and Shorland, F. B. 1977. Warmed-over flavor in meat, J

poultry, and fish. In Wank, C. O. Chichester, E. M. Mrak and G. F.
Stewart (Ed.,) Vol. 23, pp. 1-7.4 Academic Press, New York, NY.

Perata, P., Vemieri, P., Annellini, D., Bugnoli, M., Tognoni, F., and Alpi, A. 1992.
Immunological detection of acetaldehyde-protein adducts in ethanol-treated carrot cells.
Plant Physiol. 98: 913-918.

Petit, E., Divoux, D., Chancerelle, Y., Kergonou, J .F ., and Nouvelot, A. 1995.
Immunological approach to investigating membrane cell damages induced by
lipoperoxidative stress: Application to far UV-irradiated erythrocytes. Biolog. Trace
Elem. Res. 47: 17-27.

Preobrazhensky, S., Trakht, 1., Chestkov, V., and Wentz, M. 1995. Monoclonal antibody-
based immunoassay for evaluation of lipoprotein oxidation. Anal. Biochem. 227: 225-
234.

Prior, E. and Loliger, J. 1994. Spectrophotometric and chromatographic assays. Ch. 6 in
W, J. C. Allen and R. J. Hamilton (Ed ) p. 104-127. Blackie Academic &
Professional, London, England.

Przybylski, R., Eskin, N.A.M., and Malcolmson, L.J. 1991. Development of a gas
chromatographic system for trapping and analyzing volatiles from canned tuna. Can. Inst.
Sci. Technol. J. 24: 129-135.

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

Ramarathnam, N., Rubin, LJ. and Diosady, L.L. 1991b. Studies on meat flavor I.
Qualitative and quantitation differences in uncured and cured pork. J. Agric. Food Chem.
39: 344-350.

106

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

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

22 in Warsaw, T.-C- Lee and H-J. Kim (Ed),
p. 241-251. American Chemical Society, Washington, DC.

Rhee, 1988. Enzymic and nonenzymic catalysis of lipid oxidation in muscle foods. Food
Technol. 42(6): 127-132.

Robson, C.P., Collison, R., and MacFie, H.J.H. 1989. Factors affecting the shelf-life of
precooked chilled roast pork. Int. J. Food Sci. Technol. 24: 59-67.

Roozen, J. P. and Linssen, J. P. H. 1992. Factors affecting lipid autoxidation of a spray-

dried milk base for baby food. Ch. 17' in W, A. J. St. Angelo (Ed),
p. 302- 309. ACS Symposium Series 500. American Chemical Society, Washington, DC.

Rossell, J .B. 1994. Measurement of rancidity. Ch. 2 in W, J .C. Allen and
R.J. Hamilton (Ed), p. 22-53. Blackie Academic & Professional, London, England.

St. Angelo, A.J. 1996. Lipid oxidation in foods. Crit. Rev. Food Sci. Nutr. 36(3): 175-
224.

St. Angelo, A.J., Vercellotti, J.R., Dupuy, HP, and Spanier, A.M. 1988. Assessment of
beef flavor quality: A multidisciplinary approach. Food Technol. 42(6): 133-138.

St. Angelo, A.J., Vercellotti, J .R., Legendre, M.G., Vinnett, C.H., Kuan, J .W., James, C.,
and Dupuy, HP. 1987. Chemical and instrumental analyses of warmed-over flavor in
beef. J. Food Sci. 52(5): 1163-1168.

Salinas, J. P., Hartman, T. G, Karmas, K., Lech, J. and Rosen, R. T. 1994. Lipid-derived
aroma compounds' in cooked potatoes and reconstituted dehydrated potato granules. Ch. 8
ianLdLmEQQdflmz, C..-T Ho and T..G Hartman (Ed.,) p. 108- 129. American
Chemical Society, Washington, DC.

Sarnarajeewa, U., Wei, C.I., Huang, TS, and Marshall, MR. 1991. Application of
immunoassay in the food industry. Crit. Rev. Food Sci. Nutr. 29(6): 403-434.

San George, R. C. and Hoberrnan, H. D. 1986. Reaction of acetaldehyde with hemoglobin.
J. Biol. Chem. 261(15): 6811-6821.

Sapers, G.M., Panasiuk, 0., Talley, F .B. and Osman, SF. 1972. Flavor quality and
stability of potato flakes: Volatile components associated with storage changes. J. Food
Sci. 37: 579-583.

 

107

Schieberle, P. and Grosch, W. 1981. Model experiments about the formation of volatile
carbonyl compounds. J. Amer. Oil Chem. Soc. 58: 602-607.

Schirle-Keller, J .-P., Chang, H.H., and Reineccius, GA. 1992. Interaction of flavor
compounds with microparticulated proteins. J. Food Sci. 57(6): 1448-1451.

Schirle-Keller, J .P., Reineccius, G.A., and Hatchwell, LC. 1994. Flavor interactions with
fat replacers: Effect of oil level. J. Food Sci. 59(4): 813-815, 875.

Sexton, T.J. 1979. Preservation of poultry semen—a review. Ch. 12 in Animal
Reproduction, J. Hawk (Ed), p. 159-169. Allanheld, Osmun & Co. Publishers, Inc.,
Montclair, NJ.

Shahidi, F. 1994. Assessment of lipid oxidation and off-flavour development in meat and

meat products. Ch. 14 in W F. Shahidi (Ed), p. 247-
266. Blackie Academic & Professional, London, England.

Shahidi, F. and Pegg, 1994. Hexanal as an Indicator of the Flavor Deterioration of Meat

and Meat Products. Ch. 18 in W, A.J. St. Angelo (Ed), p. 105-1 19.
ACS Symposium Series 500. American Chemical Society, Washington, DC.

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

Shi, H. and Ho, C.-T. 1994. The flavor of poultry meat. Ch. 4 in W
Em F. Shahidi (Ed), p. 52-70. Blackie Academic & Professional, London,

England.

Shibamoto, T. 1990. A new analytical method for low-molecular-weight aldehydes. In

W, G. Charalambous, (Ed), p. 471-484. Elsevier Science
Publishers, Amsterdam, Netherlands.

Shin, M.G., Yoon, S.H., Rhee, J .S. and Kwon, T.-W. 1986. Correlation between
oxidative deterioration of unsaturated lipid and n-hexanal during storage of brown rice. J.
Food Sci. 51(2): 460-463.

Smith, D.M. 1995. Immunoassays in process control and speciation of meats. Food
Technol. 49(2): 116-119.

Spanier, A.M., Miller, J .A. and Bland, J .M. 1992. Lipid oxidation: Effect on meat

proteins. Ch. 7 in W, A.J. St. Angelo (Ed), p.105-119. ACS
Symposium Series 500. American Chemical Society, Washington, DC.

 

108

Steinbrecher, U. P., Milton, F., Witztum, J. L., and Curtiss, L. K. 1984.1mmunogenicity of
homologous low density lipoprotein after methylation, ethylation, acetylation, or
carbamylation: Generation of antibodies specific for derivized lysine. J. Lipid Res. 25:

1109-1116.

Stoick, S.M., Gray, J .I., Booren, A.M. and Buckley, DJ. 1991. Oxidative stability of
restructured beef steaks processed with oleoresin rosemary, tertiary butylhydroqumone,
and sodium tripolyphosphate. J. Food Sci. 56(3): 597-600.

Su, Y., Ang, C.Y.W. and Lillard, DA. 1991. Precooking method affects warmed-over
flavor of broiler breast patties. J. Food Sci. 56(4): 881-883, 898.

Tashiro, Y., Okitani, A., Utsunomiya, N., Kaneko, S. and Kato, H. 1985. Changes in
lysozyme due to interaction with vaporized hexanal. Agric. Biol. Chem. 49(6): 1739-
1747.

Thiele, G.M., Wegter, K.M., Sorrell, M.F., Tuma, D.J., McDonald, T.L., and Klassen,
L.W. 1994. Specificity to n-ethyl lysine of a monoclonal antibody to acetaldehyde-
modified proteins prepared under reducing conditions. Biochemical Pharmacology 48(1):
183-189.

Thornton, L. 1996. US. exports hit $1.7 billion. Broiler Ind. April: 22-27.

Tsai, L- S., Mapes, C. J. and Huxsoll, C. C. 1987. Aldehydes' in poultry chiller water.
Poultry Sci. 66: 983-989.

Tuma, D.J., Donohue, T.M., Jr., Medina, V.A. and Sorrell, M.F. 1984. Enhancement of
acetaldehyde-protein adduct formation by L-ascorbate. Arch. Biochem. Biophys. 234(2):
377-381.

Tuma, D.J., Newman, M.R., Donohue, T.M., and Sorrell, M.F. 1987. Covalent binding of
acetaldehyde to proteins: Participation of lysine residues. Alcoholism: Clinical
Experimental Res. 1 1(6): 579-584.

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

Vercellotti, J. R., Mills, 0. E., Bett, K. L. and Sullen, D. L. 1992. Gas chromatographic
analyses of lipid oxidation volatiles 1n foods. Ch. 13' in W, A. J. St.
Angelo (Ed.,) p. 232- 265. ACS Symposium Series 500. American Chemical Society,
Washington, DC.

 

109
Vogt, R. F., Phillips, D. L., Henderson, L. O., Whitfield W., and Spierto, F. W. 1987.

Quantitative differences among various proteins as blocking agents for ELISA microtiter
plates. J. Immunol. Met. 101: 43- 50.

Walzem, ILL, Watkins, S., Frankel, E. N., Hansen, R. J., and German, J. B. 1995. Older
plasma lipoproteins are more susceptible to oxidation: a linking mechanism for the lipid
and oxidation theories of atherosclerotic cardiovascular disease. Proc. Natl. Acad. Sci.

USA 92(16): 7460- 7464.

Wang, C.-H. 1994. Enzyme-linked immunosorbent assay for determination of endpoint
processing temperatures in uncured poultry products. Ph.D. Dissertation. Michigan State
University, East Lansing, MI.

Wang, S.-F. 1990. Effect of pH and NaCl on the thermal-physical behavior of chicken
breast salt-soluble proteins. MS. Thesis, Michigan State University, East Lansing, MI.

Warner, K., Evans, CD, List, G.R., Dupuy, H.P., Wadsworth, 1.1. and Goheen, GE.
1978. Flavor score correlation with pentanal and hexanal contents of vegetable oil. J.
Amer. Oil Chem. Soc. 55: 252-256.

Weisgraber, K.H., Innerarity, TL. and Mahley, R.W. 1978. Role of the lysine residues of
plasma lipoproteins in high affinity binding to cell surface receptors on human
fibroblasts. J. Biol. Chem. 253(24): 9053-9062.

Westendorf, R. G. 1985. Automated analysis of volatile flavor compounds. Ch. 10 1n

W D D Bills and C J Mussinan
(Ed.,) p. 138- 153. American Chemical Society, Washington, DC.

Wiskhart, G.J. 1989. Physiological changes in fowl and turkey spermatozoa during in
vitro storage. Brit. Poultry Sci. 30: 443-454.

Worrall, S., DeJersey, J ., Shanley, BC. and Wilce, PA. 1991. Antibodies against
acetaldehyde-modified epitopes: An elevated lgA response in alcoholics. Eur. J. Clin.
Invest. 21: 90-95.

Wu, TC. and Sheldon, B.W. 1988. Flavor components and factors associated with the
development of off-flavors in cooked turkey rolls. J. Food Sci. 53(1): 49-54.

Yoritaka, A., Hattori, N., Uchida, K., Tanaka, M., Stadtrnan, ER, and Mizuno, Y. 1996.
Immunohistochemical detection of 4-hydroxynonenal protein adducts in Parkinson
disease. Proc. Natl. Acad. Sci. USA 93: 2696-2701.

Zhang, H., Yang, Y. and Steinbrecher, UP. 1993. Structural requirements for the binding
of modified proteins to the scavenger receptor of macrophages. J. Biol. Chem. 268(8):
5535-5542.

 

 

MICHIGAN sran UNIV. LIBRARIES
1111111111111111”llll1111”111111111111111111
31293015639143