HutVHﬁxwﬁevwﬁnhﬁwgﬂaﬂﬂ‘J-quﬁuﬂgﬁgﬁ‘ < , A”; ‘

.v u} I???

 

 

.4
V41 331 rkﬂﬂﬂisﬁuuqtaii . s. A‘ .1,
h. . 3K . {in .450. :égiia 53“ ‘53; \‘s.
“115;;xgszwh.§ . . g": 9!
. . 3115:1433.1.)\2‘ 3.. $§J§l¢1§2 {\‘Vq ,
5341113111.z\$\!«‘11‘§nﬁti«z§4\z§§;\hwi: 331
H, . - . a .5). & .u n 1331!. .

 

\txﬁi‘i» ‘iigiﬂmﬂilfnﬂﬂ‘tiﬁw

Y’aiau‘ﬁzcvwi , “1413111.. iikdniagiﬁai . 33.31 3:59-

3:11; a? 5E2. 4"}? 1141111411. §§1§?\1 17.5.1.3! g

‘ H ‘ ‘ .h\..\rﬁnﬂu.oﬂﬂwn§3lti§t1\¥x.312x1131‘13danceaﬂilxi‘ 1.33..“ .E A

ginﬁlnllaﬁ. Jﬁ‘glwl‘ﬂ‘s.n.§.:u.: ~ .. 1 {313.111 \ . 313‘ ‘

‘ v I ‘ V \I t . . .- g; in?!“ ,
n . . . ‘ , z _ i§.i.§.u . .. ML?

J
)1?
in; v 3:. 332.1331

 

.1<.u;atu¢..n\111..ﬁ.‘4.§‘a1u..‘.m| whammy? . 5 , ‘

‘ , u ‘ $31 3.
‘ sag. . ... . ‘ , .

. .. 391.34..“ ii. . §_Vi§n_11 .
15%;; 3‘1!!! ‘ . v . 1
, sags. 33113:;

‘ .31 .

, ‘ \

.iJﬂqu; 5114:.) .
' I. ‘ V \

 

 

 

 

LIBRARY
Michigan State
University

 

 

 

This is to certify that the

dissertation entitled

THE CHEMICAL MODIFICATION OF BIOLOGICALLY

ACTIVE ALLYLIC HYDROXY FATTY ACIDS FOR

THEIR SELECTIVE DETERMINATION BY
ELECTRON CAPTURE

presented by

LINDA CAROL DOHERTY

has been accepted towards fulﬁllment
ofthe requirements for

Ph.D.

Chemistry

degree in

    
 

Major professor

Datel?) (EL/(ﬁg 9-:

MSU IS an Afﬁrmamw Atlinn/Equul Oppurlum'lr Inxlllulinn 0-12771

 

 

MSU

LlBRARlES

V

 

 

 

RETURNING MATERIALS:
Place in book drop to
remove this checkout from
your record. FINES will

 

be charged if book is
returned after the date
stamped below.

 

 

 

 

THE CHEMICAL MODIFICATION OF BIOLOGICALLY ACTIVE ALLYLIC
HYDROXY FATTY ACIDS FOR THEIR SELECTIVE DETERMINATION
BY ELECTRON CAPTURE

By

Linda Carol Doherty

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Chemistry

1987

ABSTRACT
THE CHEMICAL MODIFICATION OF BIOLOGICALLY ACTIVE ALLYLIC

HYDROXY FATTY ACIDS FOR THEIR SELECTIVE DETERMINATION
BY ELECTRON CAPTURE

By

Linda Carol Doherty

Increases in certain allylic hydroxy fatty acids have been associated with diseases
such as psoriasis, asthma and cancer. These acids are present in very low concentrations
in biological samples so that sensitive analytical methods are required for detection.
Some of the most important allylic hydroxy fatty acids are the metabolites of the
lipoxygenase enzyme oxidation of eicosatetraenoic (arachidonic) acid. The hydroxy
acids formed are called hydroxyeicosatetraenoic acids or HETEs. The HETEs have a
characteristic allylic hydroxy moiety; it is this structural feature that has been exploited in

this research project.

The objective of this work was to investigate whether these lipoxygenase
metabolites of arachidonic acid could be selectively modified, by chemical means, to a
compounds with improved chromatographic properties over those of the parent
compound. The allylic alcohols of the lipoxygenase pathway were readily oxidized to
unsaturated ketones, a distinguishing structural feature that was exploited for analytical
advantage. Also, the chemical modiﬁcation was designed to add electrophilic character
to the analyte so the highly sensitive technique of electron capture negative ionization
mass spectrometry could be used. The modification strategy selectively converted the
analyte, by virtue of its structural features, to a compound with greatly improved stability
thereby avoiding some of the problems of degradation and isomeiization associated with

the analysis of lipoxygenase metabolites.

The allylic hydroxyl was oxidized with a mild oxidizing agent to the on, B—
unsaturated ketone. Reagents such as manganese dioxide, dichlorodicyanobenzoquinone
and pyridinium dichromate favored the oxidation of allylic alcohols over compounds
containing isolated hydroxyl groups. This oxidative speciﬁcity discriminated against the
chemical background which could not be volatilized or transformed into an electrophilic
species. This methodology was developed as an alternative to procedures involving the

general derivatization scheme typically employed for hydroxy acids.

The production of the allylic alcohols and the refinement of the oxidation are
discussed in detail. The improved oxidation procedure was applied to a sample of islet
cells and the results compared with a method that employed conventional derivatization
techniques to illustrate the potential usefulness of the proposed chemical modiﬁcation
procedure. Finally, other lipoxygenase metabolites are introduced and their
determinations by conventional gas chromatographic-mass spectrometric and high

performance liquid chromatographic techniques demonstrated.

To my family:
Dave and Doris, my parents;

Lesley, my sister and Tom, my loving husband

iv

ACKNOWLEDGEMENTS

I would like to thank my preceptor, Dr. J. T. Watson for his encouragements and
suggestions throughout my studies. His positive outlook for the success of this project

gave me the conﬁdence to make progress no matter how slowly.

I would also like to thank the staff of the NIH/MSU Mass Spectrometry Facility.
Without the help of Mike Davenport, Mel Micke, Kathy Berwald, Jo Dutson and Vicki
McPharlin, I would still have years of work ahead of me. The technical knowledge I
have gained from each of these people has deﬁnitely made me a more experienced

scientist.

I acknowledge the ﬁnancial support I received from the National Institutes of
Health and the emotional support my fellow Watson group members have always
extended. The parties, picnics, camping trips and facility social hours helped me to form
some of my fondest memories of graduate school especially those special memories of

the fun times with Jack and Tonya Leary and with Kathleen Kayganich.

Finally, I want to extend my deepest appreciation to those people that made it
possible for me to attain this goal. To my parents Dave and Doris, thank you for letting
me determine my own career and for always showing your love and approval when it was
most important to me. To my sister, Lesley, thank you for helping me to grow into a
responsible person and keeping me in touch with the latest fashions. To my husband
Tom, thank you for helping me realize that I am a capable scientist and for showing your

love and patience when it was most needed.

V

 

TABLE OF CONTENTS

List of Figures ...............................
List of Tables ...............................
Key to Abbreviations ............................

Chapter I: Introduction ..........................

References Chapter I ....................
Chapter II: Conventional Methods of Analysis ............

Bioassay ............................
RIA ..............................
HPLC .............................
Mass Spectrometry .....................
Sample Introduction .....................
References Chapter II ....................

Chapter III: Autoxidation, Photooxidation and
Lipoxygenation of Fatty Acids ...........

Acid ..............................

vi

13.

14.
15.
16.
19.
26.
35.

37.

40.

42.
43.
46.

Chapter IV:

Chapter V:

Lipoxygenase ......................... 53.

Production of lS-I-IETE by Graff Method ........ 62.
Lands Preparation of 15-I-IETE .............. 69.
Other Important Isomeric HETEs ............. 76.
References Chapter HI ................... 78.
Development of Methodology .............. 80.
The Selective Oxidation of the HODDs ......... 82.
The Selective Oxidation of the HETEs .......... 91.
DDQ and PDC ........................ 94.

Dichlorodicyanobenzoquinone and

Prostaglandin B1 ................... 100.

Pyridinium dichromate and

Prostaglandin B1 ................... 105.

Optimization of pyridinium dichromate
and 15-HETE ..................... 106.

Comparison of Oxidation to Conventional

Derivatization ........................ 1 1 1.
References Chapter IV ................... 118
Collaborative Projects .................... 119
Project I ............................ 120.
Project H ........................... 121.
Project III ........................... 122.
Experimental Protocol ................... 131.
Silanization of Glassware .............. 131.
Project I ........................ 131.
Project II ........................ 136.
Project 111- The Birch Reduction .......... 141.

vii

References Chapter V

viii

LIST OF FIGURES

Figure 1-1. The various lipoxygenases that metabolize arachidonic
acid. (source: The Leukotrienes; chemistry and
biology 1984, L. W. Chakrin and D. M. Bailey eds,
Academic Press: Orlando, p. 198). ................

Figure 1-2. The arachidonic acid cascade. The large arrows are the
enzymes cyclooxygenase (right) and 5-1ipoxygenase
(left). (source: The Biochemistry of Arachidonic Acid
Metabolism 1985, W. E. M. Lands ed., Martinus
Nijhoff Publishing: Boston). ....................

Figure 1-3. The 5-1ipoxygenase pathway showing the formation of
the various leukotrienes. source: The Leukotrienes;
chemistry and biology 1984, L. W. Chakrin and D. M.
Bailey eds., Academic Press: Orlando, p. 199). .........

Figure 1-4. The effects of assorted arachidonic acid lipoxygenase
products on a polymorphonuclear (PMN) leukocyte (a
white blood cell). ..........................
Figure 1-5. Inhibitors of the arachidonic acid cascade. The

horizontal lines are where the inhibitor works. (20:4:
arachidonic acid, ETYA= eicosatetraynoic acid,
BW775= a Burroughs-Welcome competitive inhibitor,
NSAIDS= nonsteroidal antiinﬂammatory drugs). .......

Figure 2-1. , A typical 70 eV PEI mass spectrum of a hydroxy
unsaturated fatty acid, 13- OTMS 18: 2 ME (MW: 382).
The fragments at 225 and 311 are conﬁrmation ions for
this molecule. m/z= 367 and 292 are (M-15)+ and
(M-90) respectively. ........................

Figure 2-2. The fragmentation of a PFB ester under ECNI mass
spectrometric conditions. The stable carboxylate anion
is formed. ...............................

ix

Figure 2-3. Selected ion current proﬁles of lS-HETE before, (A)
and after, (B) hydrogenation. The PFB, TMS
derivatives were subsequently prepared. Ions
monitored were (A): m/z= 391 and (B): m/z= 399.
The split peak in (B) is described in the text. ..........

Figure 3-1. Competition between radical hydrogen abstraction and

B-scission, also known as cis, trans to trans, trans

isomerism. ..............................
Figure 3-2. Autoxidation products of linoleic acid and arachidonic

acid. ..................................
Figure 3-3. The GC-FID chromatogram of the overnight

autoxidation of arachidonic acid after methylation. The
production of HETE isomers was not detected by mass
spectrometry. The actual retention time for the HETE
isomers is approximately the same as the

plasticizer. ..............................
Figure 3-4. The "ene" mechanism. IP is the ionization

potential. ...............................
Figure 3-5. Photooxidation Apparatus. .....................
Figure 3-6. The chromatograms from a GC-FID of the methylated

photooxidation products (HODDs-ME) of methyl
linoleate (18:2-ME) after (A) 6 1/2 hours of irradiation
and (B) 21 hours of irradiation. ..................

Figure 3-7. A graph of the percent yield of the photooxidation
products of 18:2-ME verses irradiation time. The
reaction was terminated after 27 hours. .............

Figure 3-8. The photooxidation isomers of linoleic acid. ..........

Figure 3-9. The GC-EI-MS mass spectra of 13-OH 18:2. (A) Only
the methyl ester was prepared, m/z= 310. (B) The
methyl ester, TMS ether was prepared m/z= 382. .......

Figure 3-10. The normal phase HPLC chromatogram of separated
photooxidized linoleic acid isomers. AUFS=
absorbance units full scale .....................

Figure 3-11. The lipoxygenase mechanism. ..................

Figure 3- 12. The speciﬁcity of the most common plant enzymes
with (A) 18:2 and (B) 20:4 as substrates. ............

Figure 3-13. The RP-HPLC chromatogram of a 50 ul injection of
(+,-) S-HETE-ME (Cayman Chemicals). The HPLC
ﬂow rate was 1.0 ml/min; 7mm: 235 nm; 65:35:0.03
MeOH/HzO/HOAC with 0.5 m_M_ oxalic acid adjusted to
pH 5.6 with NH40H. ........................

Figure 3-14. The RP-HPLC chromatogram of a 50 111 injection of
15-HETE ME made by the Graff rocedure. The
HPLC flow rate was 1.0 ml/min; ax: 235 nm;
65:35:003 MeOH/HzO/HOAC with 0.5 m_M_ oxalic acid
adjusted to pH 5.6 with NH40H. .................

Figure 3—15. The GC-FID chromatogram of 15-HETE-ME made by
the Graff procedure. The temperature program was
180—280°C at 5°C/minute, 15 m DB-l megabore
column. ................................

Figure 3-16. The Sigma 3B GC-ECD chromatogram of 15-HETE-
ME made by the Graff procedure after DDQ oxidation.
The temperature program was 180-280 5°C/min, 6 ft.
OV-l packed column. .......................

Figure 3—17. The GC-EI-MS reconstructed mass chromatograms of
synthesized 15-HETE as the methyl ester, TMS ether.
The lower trace is the total ion current (TIC) proﬁle,
showing all the eluting compounds. The upper trace at
m/z= 406, the molecular weight of 15-OTMS ETE-ME,
shows which peak is the desired compound. ..........

Figure 3-18. The EI mass spectrum of the peak in Figure 3-17 that
corresponded to the derivatized 15-HETE. The
fragment ion at m/z= 335 conﬁrms the position of the
hydroxy group to be on the co-6 carbon. .............

Figure 3-19. The GC—FID response comparison of 1 1.11 each of
20:4-ME and 15-HETE-ME. The area for 1.25 mg of
20:4-ME was compared to the area for an unknown
amount of 15-HETE-ME. This procedure was used to
determine the percent yield for the lipoxygenase
reaction. ................................

xi

Figure 4-1.

Figure 4-2.

Figure 4-3.

Figure 4-4.

Figure 4-5.

Figure 4-6.

Figure 4-7.

Figure 4-8.

Figure 4-9.

Figure 4- 10.

The structure of PGBl-ME, MW: 348.

The ECD chromatograms of a fraction from cell wash
that should contain 12-I-IETE. (A) The wash
derivatized as the PFB ester, TMS ether. (B) The same
fraction was selectively oxidized and then methylated
with diazomethane. The arrows point to the appropriate
elution time of the 12-HEA, PFB ester, TMS ether (A),

or the 12-oxo-ETE-ME (B).

Mn02 can oxidize only allylic alcohols under most

reaction conditions. ........................

The (A) El and (B) ECNI mass spectra of 13-oxo-
ODD-ME (MW: 308), the Mn02 oxidized 13-HODD.

The (A) FID and (B) ECD chromatograms of the
marker compounds 2-Br 16:0-ME and 18:0-PFB. The
column was a J&W 15 m megabore DB-l. The ﬂow
rate was 15 ml/min. The temperature program was

ISO-280°C at 5°C/min. ......................

The chromatographic proﬁles of various compounds as
detected by FID and ECNI (EC-NCI).

The selectivity of DDQ for allylic alcohols in steroids

(from ref. 13). ...........................

The ability of DDQ to oxidize carbon-carbon bonds and
the possible effects on an allylic alcohol

(from ref. 13). ...........................

The reaction between HETEs and DDQ.

The TIC proﬁle (A) and reconstructed mass
chromatograms (B & C) for the oxidation reaction of
DDQ with photooxidized HETEs. The molecular
weight of the oxo-ETE-ME is 332; 330 is the molecular
weight of the overoxidized unwanted secondary

oxidation product.

xii

81.

83.

85.

86.

90.

92.

95.

96.

97.

Figure 4-11. The possible products of the reaction of PGBI with
DDQ/dioxane which are consistent with the ECNI mass
spectrometric data and comments from ref. 16 which
pertain to the formation of secondary oxidation
products. ............................... 101.

Figure 4-12. The ECNI TIC and reconstructed mass chromatograms
of PGBI after DDQ oxidation in dioxane (A). The TIC
of PGBI after DDQ oxidation in CH2C12 (B). The
structures corresponding to the various masses in (A)
are found in Figure 4-11. ...................... 103.

Figure 4-13. The B1 (A) and ECNI (B) mass spectra of oxidized
PGBl, the 15-oxo-PGB1-ME (MW: 348). The E1
fragments at m/z== 317 and m/z= 249 are explained in
the text. ................................ 104.

Figure 4- 14. The ECD and FID chromatograms showing 15-oxo-
ETE-ME after oxidation of 15-HETE with (A) MnOz, 6
1/2 hours; (B) PDC, 4 hours; and (C) PDC,
7 1/2 hours. .............................. 108.

Figure 4-15. The ECNI mass spectra of (A) 15-oxo-ETE-ME, (B)
15-oxo—ETE-PFB and (C) 15-HEA-PFB, TMS. The
column used was a DB-l 15m megabore, w rate
12ml/min. The CH4 pressure was 7 X 10 ' torr. ....... 113.

Figure 4-16. The ECNI SIM proﬁles of (A) 15-oxo-ETE-ME, (B)
15-oxo-ETE-PFB and (C) 15-HEA-PFB, TMS. The
proﬁles were obtained on the JEOL HX-110 mass
spectrometer with a DB-l column. Note the poor gas
phase properties of the 15-oxo-ETE-PFB. The peak
shape is similar to that of the 15-HETE-PFB, TMS. ...... 114.

Figure 4-17. The ECNI SIM proﬁles of the HETE HPLC fraction
from islet cells: (A) the reduced, derivatized portion,(B)
the oxidized, methylated portion The arrows point to
the appropriate elution time of the 12-HEA, PFB ester,
TMS ether (A), or the 12-oxo-ETE-ME (B). The same
sample is also presented in Figure 4-2 as the ECD
chromatogram. ............................ 1 16.

xiii

Figure 5-1.

Figure 5—2.

Figure 5-3.

Figure 5-4.

Figure 5-5.

Figure 5—6.

Figure 5-7.

The reverse phase HPLC chromatogram of 30 111 of
epithelial cell wash after ozone exposure. The ﬂow rate
was 1.0 ml/min, range: 0.01 AUFS and A: 280 nm.
The arrows show where LTC4 and LTB4 standards
would elute. The area corresponding to LTC4
isequivalent to the injection of 20 n g of standard

ch4. ................................

The reverse-phase HPLC chromatogram of the islet cell
wash. The fractions collected are noted at the bottom.
Fractions I and H had the same retention times as
standard peptido-LTs and fraction IV contained 12—
I-IETE. The chromatogram was sent by Dr. Parviz
Hayat, from the lab of Dr. Pek, University of

Michigan. .............................

The modiﬁcation of LTC4 for analysis by GC-ECNI-
MS. The Birch reduction and the hydrogenation of

LTC4 are shown. .........................

The superimposed mass chromatograms of 18:0 PFB
and 20:0 PFB. The areas were compared for the crude
approximation of the percent recovery from the Birch
reduction and the hydrogenation reactions. The peak
heights are not relative to each other but are normalized

to themselves. ...........................

The HP 5985A GC-ECNI-MS SIM proﬁles of (A)
LTD4 standard, modiﬁed according to the text, TMS,
PFB, (B) islet cell wash sample and (C) a blank
derivatized similarly. The temperature program was
60-200°C at 20°C/min (30 m DB-5 megabore, on-
column injector), ZOO-300°C at 7°C/min, source
temperatrge= 200°C, EM: 3000 V and CH4 pressure:

2.2 X 10 torr. ..........................

The JEOL HX-llO GC—ECNI-MS SIM proﬁles of
another islet cell sample after the Birch reduction and
derivatization (PFB ester, TMS ether): (A) standard
LTD4, (B) sample and (C) blank. The temperature
program was 180-280’C at %°C/Mn (DB-1 15m
megabore), CH4 pressure 5 X 10'

torr, source temperature 200°C. ................

The Silanization of glass to derivatize the active sites on

the glass surface. .........................

xiv

121.

Figure 5-8.

Figure 5-9.

Figure 5-10.

Figure 5-11.

Figure 5-12.‘

Figure 5-13.

The reverse phase Ultrasil ODS C13 HPLC
chromatogram of 50 ng of LTC4, D4, and E4, and 25 ng

LTB4. The ﬂow rate was 1.0 ml/min, range: 0.02

AUFS, and 7»: 280 nm. ...................... 133.
The reverse phase HPLC chromatograms of 100 pl of

LTB4 spiked (A) and unspiked (B) cell wash. The flow

rate was 1.8 ml/min, range: 0.01 AUFS and A: 280

nm. The area corresponding to LTB4 in (A) is

equivalent to the injection of 60 ng of standard

LTB4.

The ECNI spectrum of derivatized, stande LTB4.

The peak at m/z= 479 represents the (M-181)" ion and

the peak at m/z= 389 represents the (M-l81-90)’ ion.
MS4parameters: source T= 200°C, CH4 pressure: 2.1 x

10 torr. GC temperature program: 60-200°C at

20°C/min, 200-300°C at 7°C/min, on-column

injection. ............................... 137.

The HP 5985A GC—ECNI—MS SIM proﬁles of TMS,
PFB derivatized: (A) standard LTB4, (B) LTB4 spiked
plasma and (C) unspiked plasma. The GC—MS
parametersare listed in Figure 5-5. AU= area units. 138.

The SIM proﬁles of 18O-labelled LTB4. The 24-hour
reaction incorporated 70% of the total LTB4 with two
0 ato (m/z= 483) and 29% of the total LTB4
with one 0 atom (an= 481). .................. 140.

The physical set-up for the apparatus used in the Birch
reduction. ............................... 142.

XV

Table 2- 1.

Table 2-2.

Table 2-3.

Table 3-1.

Table 3-2.

Table 3-3.

Table 3-4.

Table 4- 1.

Table 4-2.

Table 4-3.

LIST OF TABLES

Derivatives for the Analysis of Hydroxy Fatty

Acids by GC ...........................

Double Bond Derivatives of Fatty Acids for Electron

Impact ...............................

Double Bond Derivatives of Fatty Acids for Chemical

Ionization .............................

Overview of the Success of HETE and HODD

Reactions ...............................

Speciﬁcity of Plant Enzymes ..................

Product Analysis of Soybean Lipoxygenase .........

Speciﬁcity of Animal Lipoxygenases .............

The Optimization of the Manganese Dioxide

Reaction ..............................

The Comparison of the Success of the Various

Oxidation Reagents with HETEs ...............

Properties of Various Products of Chemically Modiﬁed

15-HETE .............................

xvi

38.

55.

60.

77.

87.

KEY TO ABBREVIATIONS

AA
20:4
HPETE

S-HETE

12-HETE

15-HETE

oxo-ETES
1 5-oxo-ETE-ME

15-HEA

TMS
PFB
DDQ
PDC
HODDs
BSTFA

xvii

arachidonic acid
eicosatetraenoic acid
hydroperoxyeicostetraenoic acid
hydroxyeicosatetraenoic acids
5-hydroxy-6, 8, 11, 14-
eicosatetraenoic acid
12-hydroxy-5, 8, 10, 14-
eicosatetraenoic acid
lS-hydroxy—S, 8, 11, 13—
eicosatetraenoic acid
oxo-eicosatetraenoic acids
15~oxo-5, 8, 11, 13-eicosatetraenoic
methyl ester
15-hydroxyeicosanoic acid
methyl ester

trimethylsilyl ether
pentaﬂuorobenzyl ester
dichlorodicyanobenzoquinone
pyridinium dichromate
hydroxyoctadecadienoic acids
N, O—bis-trimethylsilyl-

trifluoroacetamide

18:2

13-HODD
oxo-ODDs
l3-oxo-ODD-ME

PGB 1

15 -oxo-PGB 1
RIA

HPLC

GC-MS
AUFS
EI

CI
ECNI
ECD
FID

xviii

octadecadienoic acid

13-hydroxy-9, ll-octadecadienoic acid
oxo-octadecadienoic acids

13-oxo—9, ll-octadecadienoic

methyl ester

prostaglandin B1

15—oxo-prostaglandin B1
radioimmunoassay

high performance liquid
chromatography

gas chromatography-mass spectrometry
absorbance units full scale

electron impact

chemical ionization

electron capture negative ionization
electron capture detector

flame ionization detector

CHAPTER I: INTRODUCTION

The objective of this research project was to develop an assay for allylic hydroxy
unsaturated fatty acids; speciﬁcally those formed from the reaction of arachidonic acid
with lipoxygenase enzyme (Figure 1-1). This reaction produces hydroxyeicosatetraenoic
acids (HETEs). The assay incorporated the knowledge that allylic hydroxy moieties can
be selectively oxidized to CL, B—unsaturated ketones, and that 0:, B—unsaturated
ketones are electrophilic. It was expected that the oxidized species would be responsive
to electron capture negative ionization (ECNI) mass spectrometry (MS) which had shown
success previously for other on, B—unsaturated ketones (1). Also, it was anticipated that
the oxidation reaction would be a discriminating alternative to sample derivatization,

greatly reducing background interferences.

5- -HETE 11- -HETE
OOH
_ / COOH _ _ coon
12-HPETE — — \/ ’ — 15-HPETE
— - 000“ S-HPETE "0° 11-HPETE COOH
12'HETE ARACHIDONIC ACID 15- ‘HETE
_ _ COOH _ _ COOH
OH OH
Figure 1-1. The various lipoxygenases that metabolize arachidonic acid. (source:

The Leukotrienes; chemistry and biology 1984, L. W. Chakrin and D.
M. Bailey eds., Academic Press: Orlando, p. 198).

The reason for the development of this assay was to have a simple, easy method
for determining the concentration of HETEs in biological samples. It was imperative that

the assay be sensitive, due to the low concentration (high potency) of the analyte. Also,

2

the technique had to be applicable to a variety of biological ﬂuids. It was hoped that the
prepared oxidized compound would increase the stability of the analyte. Finally, the
technique would have to be reliable, allowing for reasonable, reproducible recoveries

(>80%) of the analyte.

As can be seen in Figure 1-2 the release of arachidonic acid, from the
phospholipids, leads to its transformation by two enzymes, cyclooxygenase and
lipoxygenase. Speciﬁcally, the reaction of 5-lipoxygenase with arachidonic acid
produces 5-hydroperoxyeicosatetraenoic acid (S-HPETE) that is reduced to S-HETE
(Figure 1-3). S-HPETE is primarily shunted to form the epoxide leukotriene A4 (LTA4)
which is converted to the peptido-leukotrienes (peptido-LTs) LTC4, D4 and E4 or the
dihydroxy eicosanoids- the LTB4s. The basis for this project was that if the presence of
S-HETE was known, the presence of leukotrienes could be assumed. The validity of this
stems from the fact that S—HPETE is a precursor to both compounds and although most of
the 5-I-IPETE would be shunted to LTA4 some would be non-enzymatically reduced to 5-
I-IETE. Thus monitoring for S-HETE would give some preliminary information about
whether the 5-1ipoxygenase pathway was activated. If results from a quick test were
positive; a more rigorous analysis of a sample could be done to determine the exact
leukotrienes present. Other samples might contain only HETEs (12-HETE, lS-HETE,
etc.), but the same oxidative methodology could be used for their analysis. The molecule
lS-I-IETE was used in these experiments to determine the feasibility of methodology

developed.

The ability to analyze biological ﬂuids for HETEs and LTs is important for the
determination of the etiologies of many inﬂammatory and hypersensitivity diseases (e. g.,
asthma (2, 3), psoriasis (4, 5), acute arthritis (6) and bowel inﬂammatory disease (7)).
One of the physiological effects of asthma (6, 8, 9), bronchoconstriction, has been

associated with the presence of the peptido-LTs: LTC4, D4 and E4. In another example,

_._-——

COOH
Moo, IZ—HPETE W“
‘OOH
\ / IS-HPETE

OOH
COOH _ COOH
W<= m
8449575 Arochidonic Acid

0” HHT

 

-Gt -C 5'61
3 u 1\ VOH

5,
9 g COOH

s
. COOH
\ \ g

H
c” um: \‘
\S,’ OH s
\ \ COOH
' _ “ 0 SH
P602
0
LTE4\‘Glu-Cy\$ MOON

S”, OH
l\ \ H: (a)
‘ V\ LTF, PGE2

I m

Figure 1-2. The arachidonic acid cascade. The large arrows are the enzymes
cyclooxygenase (right) and 5-lipoxygenase (left). (source: The
Biochemistry of Arachidonic Acid Metabolism 1985, W. E. M. Lands
ed., Martinus Nijhoff Publishing: Boston).

_ _.. COOH
Andiidonic acid
Lipoxl'genase
'OOH OH
_ Peroxidase __ _
S—H‘PETE S—HETE

Dchydrm

0 COOH
5,12—DiHETE} / / /
<- — —
S-D‘HETE
5' ' _ 95””
Loukotriono A4 \

Hydro/u: Gluflfhlone—S-
trons/erase

on ’o/H \

OH
00H OOH
W W
H
W _ C5 11 EH2

Loukotriono 8‘ (LTB4l H CO NH CH2 COOH
H
u—oxidation :0 CH2 CH2 H COOH
H2
OH OH Lookotriom C‘
COOH
/ / 'y-Giutomyl—
"mud“:
\ OH

ZO-Hydroxy-LTB‘

l ..
l 00H
u—oxidanon / / / ’4,
OH i 0H — CSHt 1 H2

00“ H co NH cu coon
/ / — 2

”2
COOH
\ LOUKOU” D4
20—Corboxy— LTB4
(YIN/nyl—
glycincw

 

Loukotricm E 4

Figure 1-3. The 5-lipoxygenase pathway showing the formation of the various
leukotrienes. source: The Leukotrienes; chemistry and biology 1984,

L. W. Chakrin and D. M. Bailey eds, Academic Press: Orlando, p.
199).

5
inﬁltration of neutrophils (granular leukocytes) in the ﬂuid of psoriatic lesions is

associated with a high concentration of LTB4, as is the ﬂuid from acutely inﬂamed
arthritic joints (6, 8, 9, 10). Psoriatic lesions are also known to contain high
concentrations of 12-HETE and 13-hydroxyoctadecadienoic acid (13-HODD) (5). Figure
1-4 shows the effects of selective lipoxygense eicosanoids on a polymorphonuclear
(PMN) leukocyte. [The leukocyte is responsible for the removal of foreign substances

(e.g., bacteria) by ingesting them (phagocytosis)].

”34 \O

 

 

 

 

PM” CHEMOTAXIS
LEUKOCYTE _...
chg, LTD4 : ADHERENCE
tray 5—HETE
Figure 1-4. The effects of assorted arachidonic acid lipoxygenase products on a

polymorphonuclear (PMN) leukocyte (a white blood cell).

By having analytical methods available to ascertain the concentration of
eicosanoids, not only can certain eicosanoids be associated with diseases, but the efﬁcacy
of drugs or inhibitors developed against these diseases can be improved. In order to
determine whether lipoxygenase products of arachidonic acid play a role in the mediation
of the effects of any disease, inhibitors of the enzymes involved in the production of the
lipoxygenase products had to be designed and evaluated (11, 12). By selectively
blocking the production of arachidonic acid metabolites (from both cyclooxygenase and
lipoxygenase), changes in the condition of the diseased tissue may be linked to certain

eicosanoids. The amount of the eicosanoid is measured before and after the inhibitors are

added to see how the inhibitor affected the enzyme it was supposed to block. The

6

determination of which cells promote the production of inﬂammatory mediators and
which enzymes are activated in a certain disease, would lead to a better understanding of
an inﬂammatory reaction. The increased knowledge could possibly lead to the design of
drugs that more effectively eliminate inﬂammation. Some inhibitors of the arachidonic

acid cascade are shown in Figure 1-5.

A problem with the identiﬁcation of the lipoxygenase eicosanoids is that they are
potent mediators of inﬂammation that are considered to be autacoids (compounds
produced endogenously with a short half-life and local biological stimulus) (13). Thus,
they are are labile and of low concentration in normal amounts of biological ﬂuids (50-
100 pmoles). It was necessary that analytical methods be developed for the analysis of
each of these molecules taking into consideration their potency. Originally, assays of
muscle tissues were used to detect eicosanoids, because eicosanoids have the ability to
induce muscle contraction. As will be presented in chapter H, muscular bioassays are not
speciﬁc for a certain analyte and do not give reliable information for any samples except
standards. Radioimmunoassay (RIA) has also proven ineffective in reliably determining
the presence of lipoxygenase metabolites of arachidonic acid in biological ﬂuids. Cross-
reactivity of the antiserum with structurally-similar molecules will cause responses not
representative of the true analyte concentration. Only when HPLC has been used as a
preliminary puriﬁcation step, has RIA data agreed with gas chromatography-mass
spectrometry (GC-MS) data. Because GC-MS is a reliable technique, many current
eicosanoid analytical methodologies have been developed to be used with this tool. The
differences in eicosanoid structure, as can be seen in Figure 1-3, make the development
of GC—MS methodology for one eicosanoid not adequate for another compound. These
structural differences require derivatizations and modiﬁcations unique to each analyte.
Thus, the concerted effort of this research was to develop a GC-MS technique primarily
for the mono-HETEs and possibly LTB4. The peptido-LTs were not considered because

they lack the structural moiety that this research was based on, the allylic alcohol.

BOUND 20:4 IN MEMBRANE PHOSPHOLIPIDS

 

 

1 T PHOSPHOUPASE A2
CYCLOOXYGENASE UPOXYGENASE
2, 3, 5 2
‘j‘ “—4
PROSTAGLANDINS 5-HPETE \
5 S—HETE
1= STEROIDS
2- ETYA, 3w755 LEUKOTRIENES
3= NSAIDS

4== NORDIHYDROGUARIAREIIC ACID. VITAMIN E
5= 5—AMINOSAUCYUC ACID

Figure 1-5. Inhibitors of the arachidonic acid cascade. The horizontal lines are

where the inhibitor works. (20:4: arachidonic acid, ETYA=
eicosatetraynoic acid, BW775= a Burroughs-Welcome competitive
inhibitor, NSAIDS= nonsteroidal antiinﬂammatory drugs).

8
In order to develop a realistic GC-MS method, the fact that the analyte

concentration would be low and the biological matrix complex had to be considered .
Also, the analyte would have to be converted to a more stable structure. So, the
metabolism of the analyte would have to be well understood to assure that the
determination of the chosen metabolite was representative of the in vivo concentration
(14). Possible degradation of the analyte ex vivo would cause it to be an undesirable
choice as a product to be measured. This is one of the reasons why we chose to monitor
the compound S-HETE as a measure of the activity of 5-1ipoxygenase. Although the LTs
are usually of more interest, because they are more potent, only LTB4 shows little
degradation ex vivo. The peptido-LTs degrade quickly by loosing amino acid residues
(15). Monitoring LTC4 for the peptido-LT concentration may not be valid if most of the
LTC4 has actually degraded to LTE4. Recently, methodology developed for the analysis
of peptido-LTs has taken this degradation into account by converting all of the peptido-

LTs to the same molecule (16). The actual procedure is discussed in chapter V.

Another consideration in the analysis of inﬂammatory ﬂuids is the way in which
they were obtained. Many methods of obtaining a sample may actually cause the
concentration of inﬂammatory mediators to increase. If cells are injured during sample
collection, lipoxygenase products of arachidonic acid may be formed. This obviously
leads to incorrect determinations of in vivo concentration of the analyte. Preferable
techniques would involve collection of ﬂuids local to the site of inﬂammation without

scraping or puncturing the skin or cell membranes.

Finally, the analysis must exhibit sufﬁcient sensitivity, speciﬁcity and selectivity.
Although bioassay and RIA are quite sensitive, speciﬁcity and selectivity are poor.
However, with GC—MS selectivity is obtained in the form of molecular weight, structural
fragmentation and retention time information for the analyte. Sensitivity can be

increased by changing the ionization technique. Thus, GC-MS is a ﬂexible instrumental

9
technique. The details of GC-MS are covered in chapter II.

The methodology that was developed, taking into account many of the above
points, was a selective oxidation of the allylic alcohols of the lipoxygenase pathway from
the substrates octadecadienoic acid, 18:2 (18 carbons, two double bonds) and
eicosatetraenoic acid, 20:4 (20 carbons, four double bonds). The compounds formed
were the hydroxyoctadecadienoic acids (I-IODDs) and HETEs, respectively. Since the
alcohols are allylic, selective, mild oxidizing reagents were used for the transformation of
the HETEs and HODDs to oxo—ETEs and oxo-ODDs, which are 0t, B—unsaturated
ketones. The beneﬁt of this method was that the ketones formed were electrophilic; thus,
electron capture detection for GC or electron capture ionization for MS could be used.
These techniques are known for their high sensitivity. The low concentration of analyte

would be balanced by the high sensitivity of the detection method.

This methodology has been applied to the standards of 5- and 15—HETE. The
improvement of volatility of the the oxidized compound over the original molecule made
it unnecessary to derivatize the ketone moiety. The induced electrophilicity through
chemical modiﬁcation made the procedure quite selective because no general
electrophilic derivative was required. Also, since the bulk of the biological matrix would
not contain allylic alcohols, it should not be transformed into electrophilic 0t, [3-
unsaturated ketone molecules. This made the methodology speciﬁc for the analyte by
decreasing the interference from the sample matrix. Unfortunately, by choosing this
structure modiﬁcation over general derivatization, we have decreased the potential
detection limits for the analyte. The oxidized HETEs and HODDs are not as
electrophilic as halogen-containing derivatives. The reduction of background
interference tailors the use of this technique for samples that contain potentially
interfering compounds with concentrations that are much larger than that of the analyte.

Although they might be modiﬁed, it would be unlikely that they would oxidized to CL,

10

B—unsaturated ketones and, thus, would not be responsive to electron capture.

It was crucial to discern the role that the oxo-E'I‘Es played in the metabolism of
arachidonic acid so that we could be sure that using them as a representation of the HETE
concentration was valid. Skoog et al (17) reported the presence of 5-oxo-ETE in samples
in an anaerobic environment. This was conﬁrmed through earlier reports of the
formation of 13-oxo-ODD (18) in an anaerobic atmosphere. Bryant et al have found 5-
oxo-ETE is produced in monoclonal (MC)-9 mast cell cytosol (19). Fruteau de Laclos et
al have recently reported the presence of 12-oxo-ETE in human platelets (20) They
believe the transformation to be the heme-catalyzed degradation product of 12-HPETE
and not necessarily related to presence of oxygen. Skoog and Fruteau de Laclos both
determined the degradation of the HETE to oxo—ETE to be approximately 10% of the
total HETE concentration. By oxidizing the remaining 90% of the HETE present, we
would be able to assess the total HETE concentration regardless of the oxygen content in

the atmOSphere or existence of heme.

The success of the developed methodology in a biological matrix has been
examined in samples of lZ-HETE from islet cells. The preliminary results of these
experiments determined this oxidation to be a reasonable alternative to the current
derivatization methods for the HETEs (16). Future experiments by other graduate
students will offer more insight into the feasibility of this methodology for general

biological samples with complex matrices, e.g., lipid extracts from mammary tumors.

10.
11.

12.

13.
14.

15.

16.
17.

18.

19.

1 1
CHAPTER I REFERENCES

Vessman, J. in Electron Capture, J. Chromatogr. Library, A. Zlatkis and C. F.
Poole eds. 1981, 20, 137-149.

Lewis, R. A. and Robin, J-L., J. Allergy Clin. Immunol. 1985, 76, 259-264.
Lewis, R. A., Chest 1985, 87 58-108.

Mallet, A. I., Barr, R. M. and Newton, J. A., J. C hromatogr. 1986, 378, 194-
200.

Camp, R. D. R., Mallet, A. I., Woollard, P. M., Brian, S. D., Black, A. K. and
Greaves, M. W., Prostaglandins 1983, 26, 431-447.

Goetzl, E. J., Wong, M. Y. S. and Matthay, M. A., in Advances in Inﬂammation
Research 1986, I. Ottemess et al eds., Raven Press: New York, I I , 47-55.

Donowitz, M., Gastroenterology 1985, 88, 580-587.
Piper, P. J ., Br. Med. Bull. 1983, 39, 255-259.

Goetzl, E. J., Proc. Natl. Fed USA 1983, 42, 3128-3131.
Bray, M. A., Br. Med. Bull. 1983, 39, 249-254.

Arai, Y., Shimoji, K., Konno, M., Konishi, Y., Okuyama, S., lguchi, S., Hayashi,
M., Miyamto, T. and Toda, M., J. Med. Chem. 1983, 26, 72-78.

Bach, M. K. in The leukotrienes: chemistry and biology 1984, L. W. Chakrin
and D. M. Bailey eds., Academic Press: Orlando, 163-194.

Marcus, A. J., Am. J. Med. 1985, 78, 805-810.

Massicot, J. G., Soberrnan, R. J., Ackerman, N. R., Heavey, D., Roberts 11, L. J.
and Ausren, K. F., Prostaglandins 1986, 32, 481-494. .

Orning, L., Kaijser, L. and Hammarstrom, 8., Biochem. Biophys. Res. Comm.
1985, 130, 214-220.

Balazy, M. and Murphy, R. C., Anal. Chem. 1986, 58, 1098—1 101.

Skoog, M. T., Nichols, J. S., Harrison, B. L. and Wiseman, J. S., Prostaglandins
1986, 31, 577-593.

deGroot, J. J. M. C., Veldink, G. A., Vleigenthart, J. F. G., Boldingh, J., Wever,
R. and vanGelden, B. F., Biochim. Biophys. Acta 1975, 377, 71-79.

Bryant, R. W., She, H. 8., Ng, K. J. and Siegel, M. I., Prostaglandins 1986, 32,
615-627.

12

20. Fruteau dc Laclos, B., Maclouf, J ., Poubelle, P. and Borgeat, P., Prostaglandins
1987, 33, 315-337.

CHAPTER II: CONVENTIONAL METHODS OF ANALYSIS

Eicosanoids have been determined by four analytical techniques. They are
bioassay, radioimmmunoassay (RIA), high performance liquid chromatography (HPLC)
and mass spectrometry (sometimes accompanied by gas chromatography). These
methods are quite different in their sensitivity and speciﬁcity. The most desirable
technique would have both a low detection limit and high speciﬁcity. As will be pointed
out each method has its advantages, but a combination of analytical techniques usually

affords the researcher with both of the desirable qualities.

This chapter will brieﬂy address the methods of bioassay and RIA since they were
not used in this research. The important development of speciﬁc HPLC methods will be
considered in the chapters on model compound preparation and collaborative projects, so
only general aspects of eicosanoid HPLC will be discussed. The thrust of the chapter
will be in the area of mass spectrometry, where the ﬁgures of merit of the available
ionization methods and their derivatization requirements will be compared with the

procedures of sensitivity and selectivity.

The early ﬁnding that spurred interest in lipoxygenase eicosanoid analysis was
the reaction of guinea pig ileum to the infusion of snake venom (l). The unusual
muscular response from the unknown compounds in snake venom, as compared to
histamine, gave them the label of slow reacting substances (SRS). After discovering the
existence of SRS compounds, great effort was made to determine their structure. In
1979, Murphy, Hammarstrom and Samuelsson determined the structure of leukotriene C4
(LTC4) (2, 3). Further discoveries that the precursor of leukotrienes, 5-HPETE and other
HETEs were present in many mammalian tissues, has given researchers new insights into

diseases that thus far have had unknown origin.

Many quantitation procedures for the various lipoxygenase eicosanoids have been
13

14
modelled after methods originally developed for prostaglandins. The structures of the

molecules have similar features, and sensitivity requirements for analysis are
approximately the same. Since the lipoxygenase eicosanoids are potent mediators of the
inflammatory response, their concentrations in biological ﬂuids are very small. This
requires analytical techniques to be sensitive at concentrations that are much less than

one microgram/milliter.
BIOAS SAY

The technique of bioassay is the most commonly used method for eicosanoid
determination. The assay uses a muscle tissue to determine the biological response
(activity) of a compound. Different types of muscle tissue are used for eicosanoid
bioassays. SRS compounds are determined with guinea pig ileum strips. Lung
parenchymal strips of guinea pigs are used for the measurement of LTB4 because its
contractile effects are much smaller than the peptido-LTS. The tissue is kept viable in a
buffered solution and the leukotriene is periodically pumped into the buffer. The
contraction of the muscle is determined through a strain gauge which is connected to a
transducer. The response of the transducer is recorded in mV per minute. Each
eicosanoid gives a unique response with respect to concentration and time. Histamine is

used as a standard for comparison.

Bioassay cannot be used to determine the concentrations of speciﬁc analytes in
bulk unknown samples, but it can be used to determine picomolar concentrations of pure,
synthetic LTs and HPLC puriﬁed LTs (5). The temporal response of the tissue will only
conﬁrm the presence of a lipoxygenase eicosanoid. Bioassay is non-speciﬁc; i.e., non-
eicosanoid compounds may give similar responses. The technique also requires some
knowledge of tissue preparation. The beneﬁts of bioassay lie in its ability to determine
the presence of labile compounds in real time. The immediate degradation of compounds

can be determined by a series of tissue baths. It can also be used to determine the chiral

15

centers of new compounds that have been synthetically prepared because only one

enantiomer will have the proper conformation to elicit biological activity.
RIA

The technique of RIA has been used quite successfully in the area of eicosanoid
measurement. Detection limits as low as 1 picogram are reported, thus RIA is a sensitive
method for analyte measurement. Unfortunately, it does have selectivity problems due to
the cross-reactivity between molecules of similar structures. This problem is common
for RIAs developed for peptido-leukotrienes (LTC4, LTD4) and the speciﬁcity of

antiserum will vary.

Antiserum is prepared by covalently bonding the eicosanoid to a large protein like
bovine serum albumin (BSA) (6). The eicosanoids are not, by themselves, antigenic and
are called haptens. In eicosanoid RIA, the most common linkage is between the carboxyl
group of the eicosanoid and amino groups on the protein. The physical location of the
covalent bond of the hapten to the BSA is critical. Different locations of bonds may
inactivate the hapten, and antiserum may not be raised against the antagonist. The type
of bonding may also affect the cross-reactivity of the antiserum if speciﬁc groups on the
molecule are bound. If a chemical group that makes related molecules differing from one

another is hidden, these compounds will take on structural similarity.

The problem of cross-reactivity is best understood by the example of the
development of an antiserum for LTC4. There are many compounds that have very
similar structures to LTC4 such as LTD4, E4, F4, as well as some degradation products,
the leukotriene sulfones. Coupling the peptido-LT to BSA with the fatty acid carboxyl
group results in antiserum which does not differentiate between the peptido-LTs.
However, coupling the BSA to the LT through the amino group from the peptide chain in

the LT increases speciﬁcity. Surprisingly, a most successful antiserum was made by

16

coupling polyamino-bovine serum albumin to acetylated LTC4 (7). This antiserum gave
cross-reactivities to other peptido-LTs of <0.5%. As a comparison many other antisera

have cross—reactivities of more than 20%.

After the antiserum is prepared, the radiolabelled antigen must be made. In most
LT and HETE procedures tritiated standards are used. This labelling is accomplished by
3H20 exchange and the standards are not completely stable. Other radiolabelling
techniques have been based on the incorporation of 125I or preparation of 3H-methyl
esters (8). RIA determines the inhibition of the binding of the radiolabelled species to the
antiserum. This is caused primarily by the presence of the unlabelled species. However,
the binding of the radiolabelled compounds may be hindered by pH differences and
protein concentrations. In the special case of eicosanoids, high concentrations of other
fatty acids may cause cross-reactivity lowering the binding ability of the radiolabelled
species and yielding an incorrect, high concentration response for the unlabelled analyte.
These possibilities make RIA less reliable for complex samples like plasma and serum
(6). RIA does have the advantages of low detection limits and high sample throughput.
It is most useful when samples have been puriﬁed by HPLC (9,10). This eliminates
ambiguous results. Quantitation by mass spectrometry can be used to conﬁrm RIA

results.
HPLC

HPLC has been the most widely used method of analysis for the lipoxygenase
eicosanoids. This is probably due to, in part, the fact that most laboratories are equipped
and familiar with a liquid chromatograph. The development of a selective and sensitive
method for detection of HETEs and LTs has been primarily accomplished with UV
absorbance detectors. Both normal phase (NP) and reverse phase (RP) techniques may
be used for the HETEs as well as LTB4 but the sulﬁdopeptidyl LTs can only be analyzed

by RP-HPLC because of their solubility requirements. A recent review by W.S. Powell

17
of general HPLC techniques for eicosanoids is found in Biochemistry of Arachidonic

Acid Metabolism (1 1).

HPLC has been used as a subsequent puriﬁcation step in the preparation of all of
the eicosanoids. The isomers of the different photooxidation products of linoleic and
arachidonic acids have been puriﬁed by NP-HPLC which is discussed in Chapter H1.
The sulﬁdopeptidyl LTs need continuous puriﬁcation prior to use in any experiments due
to their chemical instabilites. The decomposition of LTC4 is well-known (12) to occur
within hours of biological production if precautions are not taken. Special additives to
the mobile phase decrease the rate of decomposition. The development of a suitable
aqueous mobile phase and the decomposition of peptidyl LTs is covered extensively in

chapter V.

To reduce the complexity of the liquid chromatographic analysis of biological
ﬂuids, extensive sample preparation is required. The most widely used procedures are
those employing C18 Sep-Pak cartridges. Powell (11) has developed a sequential elution
system which utilizes a C18 Sep-Pak and separates all eicosanoids into their speciﬁc
groups by changing the polarities of the ﬂushing solvents. These cartridges also remove
salts and denatured proteins that might cause interferences in the HPLC analysis. It is
also important that all samples and standard solutions are passed through 0.45 um pore
ﬁlters. These steps will remove ﬁne particulate and proteins that may clog frits in the

column and decrease the its lifetime.

As already stated, the evaluation of the selectivity and sensitivity of HPLC as an
analytical method for eicosanoids is based on the use of UV detector. The mobile phase
may be optimized for the best separation but the lowest obtainable detection limits are
determined by the quality of the HPLC system and the molar absorptivities (e) of each
LT or HETE at their kmax. The HETEs have a hmax at 235 nm (e=30,000), LTB4 at
270 nm (e=50,000) and the peptido-LTs at 280 nm (e=40,000). Since each compound

 

18
has its own lmax, the selectively of HPLC is relatively high. Most mobile phase

programs are designed to elute LTs and I-IETEs from biological samples sequentially
according to their polarity so there is no overlap of the components and the wavelength
of the detector may be changed during the analysis from 280 nm to 235 nm. Detection
limits for each of the compounds varies and is greatly dependent on noise from the
reciprocating solvent pumps. For example, experimental data showed that the analysis of
LTB4 and the peptido-LTs at 280 nm produced detection limits of 2 ng and 4 ng,
respectively. Detection limits of less than one ng could have been achieved with a
different system that had quieter pumps (more pulse dampening). Details of these

ﬁndings are highlighted in chapter V.

A recent publication (13) gave a new RP-HPLC separation which employs
ﬂuorescence as a detection method. Fluorescence has been shown to improve the
detection limits of compounds over those achieved by conventional UV methods. In this
procedure ﬂuorescent derivatives of the peptido-LTs were prepared. Detection limits for
these ﬂuorescent compounds were reported to be 1 nanogram. Other ﬂuorescent
derivatives of eicosanoids have been reported to have detection limits of less than 1 ng
(11). Fluorescence is more sensitive, but non-ﬂuorescing molecules must be derivatized.
If the HPLC technique is being used as a puriﬁcation step, it is unlikely that a ﬂuorescent

derivative is desired. Thus, its usefulness is limited to detection only.

The use of HPLC as the only method of analysis has its drawbacks. The
conﬁrmation of an analyte is accomplished by comparing the retention times of the
standards to the samples. Slight differences in the composition of the solvent media of
the standards and the samples may change retention times by as much as 5 minutes. This
reduces the conﬁdence in the peak identiﬁcation. Also, samples may contain many other
compounds that may have absorption maxima close to the wavelength used for analysis.

If a contaminant is present in a large amount, extraneous peaks in the chromatogram are

19

likely to lead to an incorrect peak assignment for the analyte. However, a beneﬁt of
HPLC is that it is non-destructive. Samples can be collected, pooled and evaluated by
some other technique. The most useful instrumental method for unambiguous

determination is mass spectrometry.

Bioassay, RIA and HPLC cannot be used for the undeniable identiﬁcation of the
analyte. However, these methods are usually more sensitive than MS, and HPLC
puriﬁcation increases the reliability of the two assay techniques. HPLC is also useful for
mass spectrometry. Mass spectrometric analysis requires high picogram amounts of
analyte to be present. If one samples does not contain enough analyte, HPLC fractions
can be pooled from many samples. The combined fractions then may contain sufﬁcient

amounts of analyte for mass spectral analysis.
MASS SPECTROMETRY

Once a sample has been puriﬁed for mass spectrometric analyte determination,
the choice of ionization method has to be made. Of the available methods, each one
gives different information and has different derivatization requirements as well as
detection limits. Eicosanoids (fatty acids in general) have been examined by positive ion
electron impact (PEI), positive and negative chemical ionization (PCI and NCI) and both
positive and negative ion fast atom bombardment (FAB). FAB is the only ionization
technique of those mentioned, that does not require the analyte to have appreciable
volatility. The other methods require the analyte to be modiﬁed somehow to increase
volatility. Since eicosanoids have low vapor pressures, derivatization or structural
modiﬁcation is necessary. The chemical derivatives of analytes for MS analysis will be

described in detail after the types of ionization methods are explained.

Mass spectrometry of eicosanoids has been most recently reviewed by Murphy

(14) in Biochemistry of Arachidonic Acid Metabolism. The different modes of ionization

20

useful for eicosanoid analysis are discussed with emphasis on the type of information that
is obtained. The choice of ionization mode for eicosanoid analysis is dependent on the
type of information desired (e.g., structure, concentration). It is helpful to know the
approximate concentration of the analyte (ug, ng, pg, etc.). This can be determined
preliminarily by HPLC or RIA. Low nanogram range samples are below the detection
limit of GC-PEI-MS, but samples properly derivatized for GC-NCI-MS will provide

much information.

Electron impact was the ﬁrst widely used ionization technique. Electrons are
emitted from a tungsten ﬁlament with 70 eV of energy. In the gas phase, the electrons
bombard the analyte and ionize it by removing an electron. The 70 eV electrons impart
so much energy into the analyte that it will fragment in pieces indicative of its structure.
Thus, PEI is very useful for structural assignment. Fatty acids and hydroxy fatty acids
have been analyzed by GC-PEI-MS most commonly as the methyl esters (ME),
trimethylsilyl (TMS) ethers. A representative mass spectrum of 13-OTMS 18:2 MB is
shown in Figure 2-1. It is apparent from the mass spectrum that the abundance of the
molecular ion is small. The fragment ion, m/z= 311, conﬁrms the identity of the hydroxy
group to be in the 13-position. The fragment is the (It-cleavage product from the
trimethylsilyl ether. Another isomer would have a different set of (at-cleavage products,
which would conﬁrm the position of the hydroxy group. It is the molecular weight plus
the GC retention time that give a ﬁrm identity to peaks eluting from a GC column. The
problem with electron impact is that the total number of ions produced is spread over
many mass units, and the higher abundance ions are usually of low mass, and lack
important structural and molecular weight information. A way to produce more higher
mass ions is to use a softer ionization technique which does not impart as much energy

into a molecule. Examples are chemical ionization and fast atom bombardment .

21

5353on

+83% can +32% 0.8 mom Ea Em "NE .2329: $6 .28 meow comawgmcoo 8a :m 95 mmm s mucoawac
2:. .33 "BE m2 we: 225-2 .28 an 852?: $06: a to 888” awe ma >0 E 8:5 <

 

 

 

 

 

 

 

 

 

 

NE
00? 00M 000 OVM ONM 00m OQN DON OVN OWN OON 00.“ 00.“ 0?.“ ON.“ 00." 00 00 0V
_:E_:E_.:.—.:L. .—.._._::_.::_.E__. l

_ I _ __ .- A- =14.....r:_ 4
hmm Nmm
NON T
nmm ..
.s m
n X I
Mk.
mzho

:n / mNN

nzoooo

HZBWZWHBH

mﬂﬂdﬁH>lﬂ

.3 ear

 

22

Chemical ionization requires the analyte to be in the gas phase when it reacts with
the reagent gas. Typical reagent gases are methane, ammonia and isobutane. The
reagent gas is ionized by energetic primary electrons which form odd electron (radical),
ionized species and secondary electrons (reaction 1). In PCI, the ionized gas reacts with
other gas molecules and protonated gas ions are formed (reaction 2). These protonated
ions react with analyte molecules (M) and proton transfer occurs (reaction 3). The more
basic a reagent gas is the more selective the protonation. The basicity (proton afﬁnity) of
the analyte must always be higher than the reagent gas for proton transfer to occur

(reaction 3).

Reaction 1. CH4 + e' --------------- > CH4“ + 2e'
Reaction 2. CH4 + CH4+° ------------ > CH5+ + CH3’
Reaction 3. CH5+ + M --------------- > CH4 + (M+H)+

° = odd electron (radical) species

Positive chemical ionization has been widely used for structure determinations.
One important outcome has been the location of double bonds positions in fatty acids.
This is outlined in Tables 2-2 and 2-3 later in this chapter. Recently, it was reported that
branched and straight chain fatty acid methyl esters have been distinguished by methyl
ether CI (15). Because it does not provide enough ion abundance at high masses;
methane PCI has not been used extensively in eicosanoid quantitation. Instead, the

fragmentation from PCI (14) proves most useful for structure determination.

Ammonia and isobutane PCI have proven successful for the determination of pure
9- and 13- hydroperoxy isomers of linoleic acid, HPETEs, LTA4 and LTB4. This method
has yielded quantitative results for pure analyte. The underivatized

hydroperoxyoctadecadienoic acids are placed on a direct insertion probe (16). A large

 

23
(M-t-NH4)+ ion is formed with ammonia CI, but isobutane CI causes fragmentation that
yields primarily (Minn-Hp)+ and (M+H-H202)+. Large quantities of (M+H-H20)+ and
(M+H-H2037)+ have also been recorded for ammonia PCI of 5- and 15- HPETE as well
as (M+H-H20)+ ions for ammonia and isobutane PCI of LTB4 and LTA4 (17). It is these

fragmentation ions that were used for quantitation.

Negative Chemical Ionization has proven to be the most successful method for
determining the concentration of eicosanoids in biological matrices. It is a soft ionization
technique which creates mostly M'° ( molecular analyte anions). The name negative
chemical ionization (NCI) implies that negative reagent gas ions react with the analyte
ions. Hydroxide ion NCI is one example (18). The reagent gases produce stable
hydroxide ions that then participate in acid/base ion-molecule reactions with the analyte.

The reactions of OH' N C1 are demostrated below:

 

 

 

 

A. N20 + e' > N2 + O"
B. O" + i-C4H10 > t-C4H9° + CH
C. OH’ + RCOOR’ > RCOO’ + HOR’
in general:
D. OR' + MH > (M-1)' + ROH
R: H, CH3

OH' NCI is much different than the method that is used most frequently in
eicosanoid analysis, electron capture negative ionization (ECNI). ECNI employs a
reagent gas (CH4, i-C4H9 or NH3) that when bombarded with high energy electrons >100
eV produces secondary electrons. These secondary electrons have their energies reduced
to that of a thermal electron (approximately 0 eV) through a series of inelastic collisions

with the high pressure reagent gas in the mass spectrometer source. These thermal

24
electrons may be captured by analytes which have a reasonable electron afﬁnity. The
electron afﬁnity is based on the energy level of vacant low lying orbitals in the molecule
(19). Thus the requirement for ECNI response is that the analyte be electrophilic.
Although biological matrices contain many components, most are not be electrophilic. If
the analyte is preferentially ionized over the matrix, the sensitivity and speciﬁcity of the
ionization technique is enhanced. Some analytes are inherently electrophilic, but the

eicosanoids are not. Appropriate derivatization or chemical modiﬁcation is necessary.

The sensitivity of ECNI is attributed to the selective ionization of those species
that are electrophilic. The rate constant of the electron capture reaction is also important
and compound dependent. Thus, the ECNI response will vary for different molecules. It
has been concluded that the ratio of positive to negative ion current, under equal
chemical ionization conditions, is equal to the ratio of the rate constants for the ion—
molecule reaction in PCI to the electron capture reaction in ECNI (20). The rate

3

constants of PCI are 10'9 cm mol'3 s'1 but ECNI rate constants may be much higher due

to greater mobility of the electron compared to the positive reagent ions.

Ion formation by electron capture is possible by three mechanisms:

 

A. associative resonance capture (0 eV)
AB + e' -------------- > AB”
B. dissociative resonance capture (0-15 eV)
AB + e' > A” + B'
C. ion pair formation (>10 eV)
AB+e' ............. >A++B'+e'

Thermal electrons produce analyte ions from both mechanisms (A) and (B). The

25
structure of the analyte will dictate whether or not dissociative resonance capture will

occur. In the case of pentaﬂuorobenzyl ester derivatives of eicosanoids, the major ion in
the ECNI spectrum is the (M-181)’ ion. The loss of 181 is the cleavage of the
pentaﬂuorobenzyl group [(M-PFB)'] leaving behind a highly stable carboxylate anion
(see Figure 2-2. below). It has been postulated that the cleaved pentaﬂuorobenzyl group
does not compete for thermal electrons because an extra electron would make the anion
formed less stable. The pi system in the benzene ring would no longer be aromatic (21).
In comparison, chemically modiﬁed compounds, such as the ones prepared in this

research (0t, B-unsaturated ketones), primarily follow associative capture.

 

Figure 2-2. The fragmentation of a PFB ester under ECNI mass spectrometric
conditions. The stable carboxylate anion is formed.

The ﬁnal ionization method that will be discussed in this chapter is fast atom
bombarbment (FAB). This technique does not require the analyte to have appreciable
vapor pressure. Thus, derivatization is unnecessary. The technique has been used in
eicosanoid analysis primarily for the structure determination of pure peptido-LTs. The
peptido-LTs are good candidates for FAB because they form ionic solutions in the FAB

matrix prior to ionization. This increases analyte ion current.

Fast atoms are produced from Xe or Ar gas. The gas is ionized and accelerated

through an electric ﬁeld into a plasma of electrons where the positive ion (Arir or Xe+)

26
picks up an electron but does not lose signiﬁcant energy to slow down. Then the fast
atoms bombard the surface of a probe covered with glycerol (a FAB matrix) and the
analyte. Both glycerol and the analyte are desorbed from the probe as ions . Some ions
receive excess energy and fragment. FAB can be used with both the positive and
negative ion detection. Negative ions produced by FAB are mostly (M-H)' ions (loss of a
proton) for molecules that, in solution, are not charged like LTB4 (14). Positive ions
produced by FAB are protonated (M+I-I)+ ions. Also, Na+ and K+ adducts are commonly
formed. FAB is not a selective ionization technique such as ECNI. Signiﬁcant
interference from the glycerol background makes this technique less sensitive than other
mass spectrometric methods and not analytically useful for biological samples without

prior puriﬁcation.
SAMPLE INTRODUCTION

Methods of sample introduction to the source of a mass spectrometer for E1 and
CI include a direct insertion probe or a gas chromatograph. The direct probe allows for
rapid analysis of an analyte. The probe is heated and compounds escape into the vapor
phase. There is little separation of components of a complex sample. The separation of
different components of the sample will be observed when reconstructed proﬁles of
selected ions representative for each component are visually displayed. Usually

eicosanoid biological samples are too complex to be analyzed on a direct probe.

Normally, biological ﬂuids are analyzed for eicosanoids by gas chromatography-
mass spectrometry. Sample components are separated chromatographically that so only
one compound enters the mass spectrometer source at a time. If other compounds, that
may be in the sample, were present in the source simultaneously with the analyte, they
would compete with the analyte for ionization. This would result in less analyte being

ionized and higher background, lowering the signal-to-background rado.

27
Gas chromatography does require the analyzed compounds to be in the gas phase.

Eicosanoids may be derivatized or chemically modiﬁed to make them more volatile.
Lipoxygenase eicosanoids are derivatized by reagents that remove reactive hydrogens
from the hydroxy and carboxyl groups. The choice of derivative is dependent on the
mode of ionization. The most commonly used derivatization reagents are presented in

Table 2-1.

In many cases, especially for the peptido-LTs, derivatization greatly increases the total
mass of the analyte; it can make the compound impossible to determine by MS. Also
some derivatives do not always improve the volatility of the analyte enough and poor
chromatographic peak shape results. Extreme GC conditions (very high temperatures)
may be necessary to improve the chromatographic properties. The problem of poor
chromatographic peak shape is evident in the case of 15-HETE. Preparing 15-HETE for
analysis by ECNI requires the 15-OH group and the COOH group to be derivatized to a
TMS ether and pentaﬂuorobenzyl ester (PFB) ester, respectively. These two derivatives
increase the mass of the parent compound by 252 mass units. The GC-ECNI-MS proﬁle
of the derivatized lS-HETE is shown in Figure 2-3A. The excessive fronting of the peak
is caused by the compound condensing in the GC. This can be eliminated by increasing
GC temperatures to 300°C, but a more realistic approach is to reduce the double bonds of
15-HETE with hydrogen and 5% Rh/Al prior to TMS, PFB derivatization (22). The
resulting 15-hydroxyeicosanoic acid (15-HEA) is then derivatized and the
chromatographic peak shape is improved greatly (Figure 2-3B). (The split peak in (B) is
caused by the injection technique (on column injection) and does not depict the presence
of more than one compound). The reason for the improvement of the peak shape after
the reduction of the double bonds is not known. It may be the relationship between the
saturation of double bonds and the PFB derivative. The peptido-LTs have a related
problem. Derivatization of all the hydroxy, amino and carboxylic acid groups would

greatly increase the total mass of the compound. The extra mass may make the

28

mmmUOU-mZ-
Ede

0©Eaeoﬁoa

buggy—0m

$000-2- .EUOUﬂZ-
293 028808
-80:th

$68-2- .mmoouiz-

0288000

merged-
.2050 as:

mmeo-mmU-OOU-
Amman c080
Ruck—Bosqmuﬁm

250x308-
62.: use
EEEseE
cane meow-
82v £0-08-
Bum 3502

GEES—Or.—
EF<>EWG

Emma:
severance.
EcoEoEng—mﬂﬁm

EEC
QUEEE Boa.
0:00w0eosaE.

Ewing
Eon ou0o<

(«Emm—

02805
Rnaﬁeoscﬁﬁm

{Sham
+§Om

0cm£0EoN§Q

Elm—04$:—
ZOE<N~H<>EWG

55.35: 32:55 0333: "52
:euanES ﬁgs—0:0 2562. "Hum
55.35: 5.58.0 02:25 unma—

H02

HUA 4mm

HUZ 40m 4mm

HUZ

HUQ 4mm

H.502
ZO~F<N~ZO~

DU Wm mO~O< >E<h WNCmm—WE .mO m~m>A<Z< HEP an: mmakatﬁmm—G

3.5-8803
mmz- .92.

EC.

3000.

QDOMU m0 m..>._.

.ﬁé mam—«G.

29

 

I5- HETE (TMS, PFB)

m/z=39l A /

 

 

 

I5-HEA (TMS, PFB)

RELATIVE INTENSITY

 

 

 

 

 

I4 l5 I6 l7 l8 I9 2o 2|
TIME (mins)

Figure 2-3. Selected ion current proﬁles of 15-HETE before, (A) and after, (B)
hydrogenation. The PFB, TMS derivatives were subsequently
prepared. Ions monitored were A: m/z= 391 and B: m/z= 399. The
split peak in (B) is described in the text.

30
compound difﬁcult to elute from a GC column. Also, with a molecular weight over 1000
the mass range of the mass spectrometer may be at its limit, and unit resolution between
sequential masses may not be possible. It is advantageous to cleave off the peptide side
chain and reduce the double bonds prior to derivatization. A detailed procedure for the

analysis of sulﬁdopeptidyl-LTS is found in chapter V.

Another way of improving the gas phase properties of the analyte is structure
modiﬁcation. Oxidation of hydroxy groups to ketones will improve vapor phase
properties and chromatographic peak shape. Oxidations employed in our laboratory also
increase the electrophilic character of the analyte. The nascent electrophilic analyte can
then be analyzed by ECNI. The oxidations are also more selective than the general
derivatizations. This adds more speciﬁcity to the methodology and analysis. The

beneﬁts of oxidation over derivatization are covered in Chapter IV.

Derivatives can be used in structural assignments. One important example of this
is a derivative of fatty acids that helps with the determination of double bond positions.
Early work in fatty acid determinations required the knowledge of structural features,
most commonly the location and geometric conﬁguration of double bonds. The location
of double bonds has not been easy to assign by mass spectrometry. During ionization
double bonds tend to migrate. Many reactions that are geometrically speciﬁc (0504) can
be used to determine cis and trans double bonds. OsO4 oxidizes the double bonds to cis
diols and eliminates bond migration, and hence, ambiguity. Derivatives are then made
of the diols and the compounds are analyzed by various mass spectrometric techniques.
The fragmentation is driven by the heteroatoms now at the site of the original double
bond, which reveals the double bond location. Other derivatives are made of the
carboxylic group and a unique fragmentation pattern discloses the positions of the double
bonds. Recently tandem mass spectrometry has been used for the determination of

double bond positions in underivatized fatty acids.

 

 

31

All of these methods have drawbacks, which are explained in the comments in the
tables. Most frequently double bond positions of monoenoic and dienoic acids are easy
to assign from the mass spectra, but polyenoic acid mass spectra give no useful
information. Another problem has been the assignment of double bonds in hydroxy
polyenoic acids. The hydroxy group will change the fragmentation of the molecule.
Tables 2-2 and 2-3 summarize the types of derivatives, the ionization mode used, and the
information obtained from the determination which have been published. Work in our
laboratory is being conducted to explore new methods for the determination of the double

bond positions is fatty acids, especially those containing hydroxy groups.

It has been shown that a variety of analytical techniques is needed for the
determination of structure, bioactivity and concentration of lipoxygenase eicosanoids and
in general, all fatty acids. The analysis of biological samples generally requires more
than one analytical tool. Although bioassay is important, the most regularly used
techniques are HPLC with RIA, HPLC with mass spectrometry or both. The available
amounts of biological ﬂuids and an approximate analyte concentration will determine

whether mass spectrometry is feasible.

Mass spectrometry does conﬁrm the identity of the analyte more accurately than
any other technique. The information obtained from a mass spectral analysis is not
questionable unlike RIA data. Eventually it will become the most widely used method
for eicosanoid analysis. It is now important that easier derivatization techniques be
developed for the facile analysis of HETEs and LTs. Many existing methods require
highly trained persons to assess the presence and concentration of these elusive

compounds. New methodology is addressed in Chapter IV.

32

0:808 0:: 0.96%
>66»: :2?» 0.63 6: :3»
6:0: 03:0: 6 6602 >30

6:0: 03:0:
8:833. 86: .~30. 6 008::
880 :0:S:08w0.c 0300556

0333.0:
0523:0808 002 30¢

0:0: 836:8 3:880:80.
6 06$ 8:06:02 0:00 05:00
065.20: 088.600: 6800

6880895 0300856

.80: 836:8 b83080. 30%
6:00 03:00 6 t 05 u:

80:3 55808.0. N + :3 + «:0
8083:0885 £008

0.808800

06:030:
06:00:08

06:06

06:00:08

06:00:08

66: 06:030:
06:00:08

66: 06:030:
06:00:08

0:286.

8m 58
e030

OEEOEQ

8363.5: 20: 0:80:80

:8
050 68
$0200 803% 185:0

Ch 08
088060302:

GS
050230808

@NV
.650 MSG.

as
:050 3508
20:0 .86? £0 \eOmO

30:80.65
0>:a>t0Q E0ma0m

BOISE ZOKHQHAH mOh mQ~U< >.:.<.m RC mghksﬁmmn— QZOm HAM—DOG .N..N mun—med.

33

0030000 A0:00
003000000000 0: $060
06:0300 :00 0000: :0:
080830 0:30:80 =080

0:00 80830 80880.8

0:00.00
:0 0080 +E+$0 =080
80000 80800:: b02885
”060068 6 EwBB
0000008 00:00 06:00 m v

0:2 800608 b08808:
.6 00:00:30 00000008
085008 32 00w 8:020

006w000: 0: >000 0:0:
0:0t008m bag—0:80

0808800

0080 03:08 38> HMS;
~Z\mm0\m2>

OZ

0:00:08“

£2

0:000:63

000 80w00m

06:00:08

06:0300

0060 06:00:08
06:0300

0060 06:00:08
06:0300

03080~

$0
$0

060 00000360....

Ame
3:5 02,0

ANS

:9
:050 3508

.20 :20: 200000

30:0:0000

0>00>E0Q 80w00m

 

QUE 5 0>080&

ZO~E<NHZO~ J<UHEHEU ZOE mm=0< WERE,”— mO mmah<>~¢mn GZOm HAM—DOG .n.N mun—mag

4
3

:0_:_00080000 00:02:00 2:820:60 HQ<U

8.20 :o 008% 96
00:00 06:00 .6 :00000:

:0: 0:0: 0000::w06 9505
8-2: :0 008% 05

0808800

QZBEQONZV
.m0

000 ::0m00m

0060 06:0300
06:00:08

0060 06:0»60
06:00:08

0:30:40:

PPZOU .mé min—<8

A5
m<m
60:
:00: 00:0:
80:00 0:0 0200 3:0:
600063 0>00w0z
2.32
30:80:00

60000085 068%

10.

11.

12.

13.

14.

15.
16.
17.

18.

35
REFERENCES CHAPTER II

Feldberg, W. and Kellaway, C., J. Physiology 1938, 94, 187-226.

Murphy, R. C., Hammarstrom, S. and Samuelsson, B., Proc. Natl. Acad. Sci.
USA 1979, 76, 4275-4279.

Hammarstrom, S., Samuelsson, B., Murphy, R. C., Clark, D. A., Mioskowski, C.
and Corey, E. J ., Biochem. Biophys. Res. Commun. 1979, 91, 1266-1272.

Piper, P. J. and Vane, S. R., Nature 1969, 223, 29-35.

Piper, P. J. in The Leukotrienes: their biological signiﬁcance, 1986, P. J. Piper
ed., Raven Press: New York, 59-66.

Salmon, J. A., Br. Med. Bull. 1983, 39, 227-231.

Lindgren, J. A., Hammarstrom, S. and Goetzel, E. J., F EBS Lett. 1983, 152, 83-
88.

Moonen, P., Klok, G. and Keirse, M. J. N. C., Prostaglandins 1985, 29, 443-
448.

Beaubien, B., Tippins, J. R. and Morris, H. R., Biochem. Biophys. Res.
Commun. 1984, 125, 97-104.

Taylor G. W., Chappell, G. G., Clarke, S. R., Heavey, D. J ., Richmond, R.,
Turner, N. C., Watson, D. and Dollery, C. T., in The Leukotrienes: their
biological signiﬁcance, 1986, P. J. Piper ed., Raven Press: New York, 67-90.

Powell, W. S. in Biochemistry of Arachidonic Acid Metabolism, 1985, W. E. M.
Lands ed. Martinus Nijhoff Publishing: Boston, 375-403.

Westcott, J. Y., Clay, K. L. and Murphy, R. C., J. Allergy Clin.1mmunol. 1984,
74 363-368.

Dowbrowsky, R. T., O’ Sullivan, G., Ballas, L. M., Fleisher, L. N. and Olsen, N.
C., J. Chromatogr. 1987, 10, 137-160.

Murphy, R. C. in Biochemistry of Arachidonic Acid Metabolism, 1985, W. E.
M. Lands ed. Martinus Nijhoff Publishing: Boston, 418—435.

Killmer, L., abstract from 35th ASMS Conference Denver CO, May 1987.
Plattner, R. D. and Gardener, H. W., Lipids 1985, 20, 126-131.

Jamieson, G. C., Gut, J. and Trudell, J. R., abstract from 35Lh ASMS Conference
Denver CO, May 1987.

Roy T. A., Field, F. H., Lin, Y. Y. and Smith, L. L., Anal. Chem. 1979, 51, 272-
278.

19.

20.
21.
22.
23.
24.
25.

26.
27.
28.

29.
30.
31.

32.
33.
34.

35.

36.
37.

 

36

Chapman J. R., Practical Organic Mass Spectrometry 1985, John Wiley and
Sons: New York.

Doughtery, R. C., Anal. Chem. 1981, 53 625A-636A.

Strife, R. J. and Murphy, R. C., J. Chromatogr. 1984, 305, 3-12.

Balazy, M. and Murphy, R. C., Anal. Chem. 1986, 58, 1098—1 101.

Neihaus, W. G. and Ryhage R., Anal. Chem. 1968, 40, 1840-1847.

Dommes, V., Wirtz-Peitz, F. and Kunau, W., J. Chrom. Sci. 1976, 14, 360-366.

McCloskey, J. A. and McClelland M. J ., J. Am. Chem. Soc. 1965, 87, 5090-
5093.

Brooks, C. J. W. and Maclean, I., J. Chrom. Sci. 1971, 9, 18-24.
Oliw, E. H., J. Chromatogr. 1983, 275, 245-249.

Vincenti, M.,Gulielmetti, G., Cassani, G., and Tonini, C., Anal. Chem. 1987, 59,
694-699.

Andersson, B. A. and Holman, R. T., Lipids 1974, 9, 185-190.
Eagles, J ., Fenwick, G. R. and Self, R., Biomed. Mass Spec. 1979, 6, 462-464.

Suzuki, M., Ariga, T., Sekine, M., Araki, E. and Miyatake, T., Anal. Chem.
1981, 53, 985-988.

Lankelma, J., Ayanoglu, E. and Djerassi, C., Lipids 1983, 18, 853-857.
Ariga, T., Araki, E. and Murata. T., Anal. Biochem. 1977, 83, 474-483.

Bambagiotti, M., Coran, S. A., Giannellini, V., Vincieri, F. F., Daolio, S. and
Traldi, P., Org. Mass Spec. 1983, 18, 133-134.

Bambagiotti, M., Coran, S. A., Giannellini, V., Vincieri, F. F., Daolio, S. and
Traldi, P., Org. Mass Spec. 1984, 19, 577-580.

Tomer, K., Crow, F. and Gross, M., J. Am. Chem. Soc. 1983, 103, 5487-5488.
Jensen, N., Tomer, K. and Gross, M., Anal. Chem. 1985, 57, 2018-2021.

CHAPTER III: AUTOXIDATION, PHOTOOXIDATION AND
LIPOXYGENATION 0F FATTY ACIDS

The preparation of model compounds for this project was a challenging part of
the research. The allylic hydroxy fatty acids are unstable, and they were not
commercially available until recently. The formation of oxidized fatty acids in vegetable
oil and other products has been a problem in the fats and lipids industry. The oxidation
of fatty acids produces species that add unwanted ﬂavors and odors (i.e., rancidity) to oils
(1). The odors and ﬂavors are linked to the thermal decomposition of the hydroperoxy
species. The earliest information on the oxidation of fatty acids is found in journals
which pertain to the chemistry of fats and oils. Lipid chemists have done exhaustive
research to determine the causes of fatty acid oxidation. The fatty acid that has been used
most ﬁ'equently as a model compound in studies is linoleic acid (18:2). It is a major
component in plant lipids. Researchers became interested in making allylic alcohols to
determine the mechanism of production. Thus, the methodology for preparing allylic

hydroxy fatty acids from parent acids is well documented.

The oxidation of fatty acids is known to occur by three different mechanisms.
Each mechanism results in different products that have been oxidized at a speciﬁc carbon
(regiospeciﬁcity). This chapter presents the mechanisms of the three oxidation methods.
It includes the considerations necessary in model compound preparation in order to
obtain maximum yield and purity of the isolated products from the starting compounds.
The preparations are described in detail for the four methods that were evaulated; each
procedure and its potential pitfalls are regarded with insight. It was observed that many
problems can arise from incomplete descriptions in published papers. The published
procedures were altered during this research many times. In order to minimize future
problems with contamination the ﬁnal step-by-step syntheses are presented here. Some

of the more speciﬁc reactions and the glassware Silanization procedure are found in other
37

38

chapters of this thesis.

The ﬁrst step in oxidation is the formation of the hydroperoxy acid. The most
selective way of preparing the hydroperoxy unsaturated fatty acids is by enzyme
oxidation. This method insures regiospeciﬁcity that is not possible with more general
oxidations such as photooxidation and autoxidation. Because the enzyme reaction was
not initially fulﬁlling the needs for milligram quantity yields, the processes of
photooxidation and autoxidation were utilized for gross productions of isomeric allylic
alcohols which were then used for the optimization of the selective oxidation reactions
found in chapter four. The reactions that were tried, their subtrates and the results
obtained are found in Table 3-1.

TABLE 3-1. OVERVIEW OF THE SUCCESS OF HETE AND HODD
REACTIONS

HETE: hydroxyeicosatetraenoic acid, HODD: hydroxyoctadecadienoic acid

REACTION TYPE SUBSTRATE % YIELD (gram amounts)/
# OF ISOMERS FORMED

lipoxygenase method:

Sigma Chemical 18:2 0.06% (.05 mg)/ two isomers
" 20:4 0%/ one isomer
Graff 20:4 0%/ one isomer
Lands 20:4 6% (3 mg)/ one isomer
photooxidation 18:2 70% (70 mg)/ four isomers
" 20:4 20% (10 mg)/ eight isomers
autoxidation 20:4 ' 0%/ six isomers

In each oxidation method the hydroperoxy acid, not the hydroxy acid, is formed.
The hydroperoxy species is very unstable and may isomerize. Therefore, to prevent a
mixture of products, it is very important to reduce the hydroperoxy group to the hydroxy

group as soon as possible.

39

There are many reducting agents (enzymatic and non-enzymatic) that can be used
for this purpose. The three utilized in this research were NaBH4 (Aldrich),
triphenylphosphine (TPP) (Aldrich), and glutathione/glutathione peroxidase (Sigma
Chemical). Each of these reducing agents has its own procedure and, in some cases, such
as with TPP, severe problems. NaBH4 is the least expensive reagent and the easiest to
use. The reaction procedure is explained in the autoxidation section. The only "trick" to
using NaBH4 successfully is to add water after the reduction step to neutralize the excess

reagent. This makes the ﬁnal extraction much easier.

TPP is a reagent speciﬁc for hydroperoxy groups. It is very easy to use, but can
produce severe contamination problems. It also complicated the HPLC puriﬁcation of
lS-HETE. The complete procedure and drawbacks are explained later in the chapter in
the Lands lipoxygenase method for the preparation of 15-HETE. The reduction may be
accomplished enzymatically by glutathione (GSH) and glutathione peroxidase (GSH-PX).

The reaction for this reduction method is shown below:

GSH-PX

 

2 GSH + PHDETE > GSSG + HETE + H2

This is the way that the hydroperoxy compounds are reduced in cells. This enzymatic
method does not produce compounds that could potentially coelute with the HETE
because the GSH, GSSG (oxidized GSH) and the GSH-PX are water soluble and are not
extracted into the organic phase with the model compounds. The only disadvantage of
this method is that the GSH-Px is very expensive and not a realistic choice for reducing

large quantities of oxidized fatty acid.

 

40
AUTOXIDATION

Autoxidation is a free radical process with initiation, propagation and termination
steps. Fatty acids such as arachidonic, linoleic and linolenic with methylene-interrupted
double bonds are the best substrates. The initiation step is least understood but it is
thought to be due to decomposition of hydroperoxides formed by photooxidation in the
presence of trace iron and copper. This is explained later in detail in this chapter. One
way to control~ autoxidation is to add a chelating agent to remove metal ions or a
compound with which free radicals preferentially react. For example, BHT (butylated
hydroxytoluene) is widely used for this purpose in the food industry.

Autoxidation was brieﬂy examined as a method of preparation for
hydroxyoctadecadienoic acids (HODDs) and hydroxygicosatetragnoic acids (HETEs).
The rate of autoxidation is about a thousand-fold slower than photooxidation (2). Most
photooxidation reactions that have been conducted in our laboratory were at least forty
hours long. The procedure was not expected to give great yields unless reaction times
were weeks long. Autoxidation also produces many by-products such as
dihydroperoxides and cyclic products. Other unwanted by-products are trans, trans-
conjugated dienes instead of the desired cis, trans. It is only the cis, trans isomers that are
of biological importance in the arachidonic acid cascade. Porter has found that the
amount of trans, trans versus cis, trans isomers formed is directly connected to the
thermodynamic/ kinetic control of the reaction. The concentration of the trans, trans
species may be minimized by the use of low temperature during the oxidation (3)
(thermodynamic control). Also high concentrations of fatty acids favor the cis, trans
isomers (kinetic control). The addition of Vitamin E (a-tocopherol) greatly increases the
amount of the cis, trans isomers. Vitamin E increases the chances of hydrogen
abstraction (kinetic control) instead of B-scission (thermodynamic control). The latter is

the cleavage of oxygen from the carbon [3 to the site where radical hydrogen abstraction

 

41

occurred (Figure 3-1). A drawback of using Vitamin E is that it causes a decrease in the
formation of hydroperoxy products. Great care should be exercised when a-tocopherol
is added due to the possibility of completely consuming all the peroxide radicals and
forming no product (hydroperoxyeicosatetraenoic acid, HPETE). Only trace amounts of

a-toc0pherol should be added in an autoxidation reaction.

02 00‘ OOH
.0- r—\-- 0———\/—V—< __..H- M—
—\\_/ _02 — / /

1-... ar-

ﬂ—_ ﬂ_ _ -_ /_ .—
\\ /

02H»

'00 H00

>—\/———/ >W=/

Figure 3-1. Competition between radical hydrogen abstraction and B-scission,
also known as cis, trans to trans, trans isomerism.

 

Porter (4) has described the autoxidation mechanism in detail for linoleic and,
arachidonic acids. The possible products for the reactions are shown below in Figure 3-
2. The autoxidation of arachidonic acid to HETEs was investigated by Brash, Porter and
Maas (5). The mechanism of the propagation step was of interest to see if the reaction
was a purely random process. This was found to be the case. The abstraction of a
hydrogen radical and addition of the oxygen radical was not stereoselective unlike those

in enzyme oxidations.

 

___ COOH

   

, t_ ‘K\\——-15-OOH
11—OOH 12—OOH

Figure 3-2. Autoxidation products of linoleic acid and arachidonic acid.

DIRECTIONS FOR THE AUTOXIDATION OF ARACHIDONIC ACID

Autoxidation of arachidonic acid was suggested by an employee of Cayman
Chemical because previous trials at HETE production had not been successful. The
autoxidation of AA produces only six isomers. The production of HETEs by

autoxidation was accomplished in the following manner:

A known amount of AA (Nu-Chek Prep, Elysian, MN) (ca. 50 mg) is placed in a silylated
ﬂask. To the ﬂask is added a small amount of oc-tocopherol. The ﬂask is ﬁlled with ca.
100 ml of ethanol or toluene and oxygen is bubbled into the solvent. The reaction is
terminated by treating the solution with NaBH4 for one hour (20 minutes at 0°C and 40
minutes at room temperature); water is added until a white cloudy solution remains. This
solution is acidiﬁed to pH 3 with 6 N HCl and then extracted with ether, hexane or
hexane/ethyl acetate (1:1). The solvent is removed under a stream of nitrogen and the

reduced residue is again dissolved in methanol. It is this ﬁnal solution that may be

 

 

43
puriﬁed by normal phase (NP) HPLC.

After an autoxidation reaction was conducted with arachidonic acid, the reaction
mixture was methylated with diazomethane for GC-FID analysis. The resulting
chromatogram is displayed in Figure 3-3. The largest peak is phthalate ester possibly
originating as a contaminant from the ethanol, glass-distilled petroleum ether or the
HPLC grade ethyl acetate all of which were used in ca. 100 m1 volumes and concentrated
to dryness. The problem of phthalate ester contamination will be addressed at length in

the lipoxygenase section.
PHOTOOXIDATION

Photooxidation was also examined as a way of preparing large amounts of_
HODDs from linoleic acid and HETEs from arachidonic acid. The mechanism of
photooxidation is much different from that of autoxidation. The reaction is initiated by
singlet oxygen instead of an alkyl radical. Molecular oxygen is promoted to the singlet
state by a sensitizer and light. Common sensitizers are chlorophyll and methylene blue.
The reaction mechanism is shown below. The ground state sensitizer is promoted to the .
singlet state by absorption of a photon. Through intersystem crossing the sensitizer
decays to the triplet state. When ground state triplet oxygen reacts with the triplet state
sensitizer, it is promoted to the reactive singlet state. All reactions occur with the singlet

state oxygen.

3 02
1 9

sens + hv1 > sens > sens

 

 

 

>02

 

00060006 0:: 00 0800 0:: b80808800

0: 80800: Ema 0:: :0: 08:: :00:0:0: 00800 0::. 0308000000 0008 .3 00:00:00 :0: 003 80800: Ema

6 :0::00008 0::. 800000.008 0:00 0:00 06006000 6 :00000880 23806 0:: 6 80:»008020 95-00 0::. .m-m 0::wE

 

 

A 62:2: 0.2:

 

 

 

 

Ina/41511.1 . . ital

 

 

M

 

 

 

 

(‘31an AHVHIISHV) SSNOdSBH 80103130

 

 

 

 

45
The oxidation proceeds through an "ene" reaction which is presented in Figure 3-4.

singlet oxygen attack occurs preferentially at the
more highly substituted double bond (lowest IP)

1 2 3

l l I

_ 1 \ _
—c—C—c + 02 c—c—c\
H OOH

above reaction resembles the 'ene' reaction below

 

 

0 ,0
/
+ = |
\ \
H o 0
Figure 3-4. The "ene" mechanism. IP is the ionization potential.

Photooxidation is not affected by the introduction of antioxidants but may be
terminated by singlet oxygen quenchers. It is thought that autoxidation in fats and lipids
is initiated from the hydroperoxides produced by photooxidation. (6,7) The sensitizers
necessary for photooxidation are present in the fats and lipids at very low concentrations
even after bleaching. Sunlight excites the sensitizers and the unwanted hydroperoxy

products are formed.

It was necessary to assemble a photooxidation system in our laboratory. The
method which seemed to be easily implemented was that of Thomas and Pryor (8)
involving the reaction of methyl linoleate (18:2 ME) with methylene blue. The light
source used in the photooxidations must produce light at the Kmax of the sensitizer used.
The Max of methylene blue is 653 nm; thus high intensity visible light was needed for
excitation. The most easily accessible source of high intensity visible light was a slide
projector. All photooxidations were accomplished at temperatures less than 10°C to

minimize cis, trans to trans, trans conversions of the double bonds in the fatty acid.

 

46
DIRECTIONS FOR PHOTOOXIDATION

Approximately 50 mg of an unsaturated fatty acid is added to a vial. The methyl
ester of the acid is prepared with diazomethane [prepared by the non-alcoholic
instructions on the Diazald bottle (Aldrich)]. The methylated product is transferred to a
125 ml ﬁlter ﬂask (an Erlenmeyer ﬂask with a sidearm). Approximately 25 m1 of
methanol is added to the ﬂask. There should be enough methanol in the ﬂask to submerge
the pipet which will be used to bubble in the oxygen. A small amount of methylene blue
(Aldrich) is added to the ﬂask containing a stir bar. The reaction is conducted in the cold
room (5°C). The ﬂask is placed on a magnetic stir box and plugged with a rubber stopper
that has a pipet in it through which the oxygen ﬂows. Excess oxygen escapes through the
sidearm. When the oxygen starts to bubble, the magnetic stir box is turned on. Finally,
the projector is set at the highest intensity and focused on the solution. The schematic

for the irradiation system is shown in Figure 3-5.

oxygen (from tank) /\

slide projector

?©?<0 ..

18:2 ME + methylene blue
in methanol

:excess 0 2

 

 

stir box

 

 

 

Figure 3—5. Photooxidation Apparatus.

The apparatus is periodically checked to make sure that the oxygen is bubbling,
the stir bar is rotating, and that there is plenty of methanol in the ﬂask. One milliliter

aliquots are removed from the reaction vessel every 6 hours. After photooxidation, the

47
compounds are in the very reactive hydroperoxy form. They are initially reduced with
sodium borohydride (NaBH4) to the hydroxy form. The aliquot is transferred to a vial
and about 2 milligrams of NaBH4 is added. The vial is placed in the freezer for 20

minutes and then left at room temperature for an additional 40 minutes.

At the end of the hour, water is added to the vial with the reduced photooxidation
product to double the volume. A pH electrode is used to monitor the pH of the solution
as it is adjusted to 3.0 with 6 ﬂ HCl. A cloudy white solution will result. The sample is
transferred to a separatory funnel for extraction. Three 10 ml portions of 50% ether in
hexane are used as the organic phase. The ether layer from each extraction is collected.
They are pooled and then washed with water two times. The combined, washed ether

layers are dried with anhydrous NaZSO4.

The photooxidation reaction was conducted for 28 hours and the chromatograms
of aliquots removed at 6 1/2 and 21 hours are presented in Figure 3—6. The concentration
of photooxidized products with time is shown in Figure 3-7. The light source was of low
intensity (30 watts), as compared to the type of source used by Thomas and Pryor (8)

(650 watts), making it necessary to irradiate the reaction mixture longer.

There should be four cis, trans isomers made (# of possible isomers = 2n where n
is the number of double bonds); the yields of these isomers (2,8) and the structures are
presented below in Figure 3-8. There were more than four compounds made in the
synthesis reported here; some of the cis bonds must have been converted to trans. Thus
instead of having cis, trans isomers exclusively, the trans, trans isomers were produced;

the trans, trans isomers cannot convert back to cis,trans.

To determine the identity of the peaks in the chromatograms, the products were
aIlalyzed by GC-EI-MS. Initially, the sample was analyzed after treatment with only

diazomethane; this would leave a free hydroxyl group on any alcohol. If 13-OH 18:2

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

48
—— r
6 1/2 hours
AA
(1)
t
% k
E 5T" L‘l A
< l
(I i
t .
a)
3% 18:2-ME HODDs ME
m
a)
Z
O
a.
(1)
Lu
(I
a:
O
F.
8 \
m x 21 hours
0
1 J L/A/K
{1... r T I —'r
18:2-ME HODDS ME
TIME (MINS) >
Figure 3-6. The chromatograms from a GC-FID of the methylated photooxidation

products (HODDs-ME) of methyl linoleate (18:2-ME) after (A) 6 1/2
hours of irradiation and (B) 21 hours of irradiation.

49

PHOTOOXIDATION OF 18:2

PERCENT YIELD VERSUS REACTION TIME

 

 

100 -—-
.L ‘
80 ——
o
d 60 ~—
;
i— --
2
DJ
0
E 40 --
a.
20 ~-
1 1 n 1 i l
0 l I 7 1 l l
o 10 20 .30
PHOTOOXIDATION TIME (HOURS)
Figure 3-7. A graph of the percent yield of the photooxidation products of 18:2-

ME verses irradiation time. The reaction was terminated after 27
hours.

 

 

50
were present the underivatized hydroxy acid methyl ester (13—OH 18:2 ME) would be

10—OOH —\ /— 9-OOH

COOH

‘lZ—OOH —/ \—13-OOH

Q—OOH (357.)
10-OOH (177:)

iZ-OOH (177.)
13—00H (327:)

All isomers are initially cis, trans

Figure 3-8. The photooxidation isomers of linoleic acid.

produced. This compound, if present in the sample, would have a molecular weight of
310 daltons. When the sample was analyzed by GC-EI-MS, the mass spectrum did not
indicate a molecular ion but instead a peak at m/z= 292 was present. This peak represents
an ion corresponding to (M — 18)+ (loss of water). The loss of water is very common
when free hydroxyl groups are exposed to E1 conditions. Another aliquot of the
compound was derivatized with BSTFA (Pierce) as well as diazomethane. The resulting
compound would now have a molecular weight of 382 (13-OTMS 18:2 ME). The
corresponding spectra for the two differently derivatized compounds are shown in Figure
3—9. It can be seen that when only the carboxylic acid group is derivatized, there is little
structural information. Only low mass ions are present. However, when. the
trimethylsilyl ether is made, higher mass ions with structural information are of large
abundance and can be used for determining the location of the hydroxy group in the

molecule.

As can be seen in Figure 3—6., the mixture of photoxidized products is complex.

It was necessary that the sample be puriﬁed. HPLC had proven successful for separation

 

51

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

CH
MW-310
7o
- 93
e7

- as x a

' 292
R _
E
L -
A 108
T - ‘2‘ 149 I
I l t 135
V .

_ ti , l ' '1 5‘: l ml I 1.11.; x“ -L l h

UITITI TTVIIIIII UUII I'll II r11 TIIIIUIII

I so no 70 ea :10 130 sea :70 190 210 230 250 270 290 310 330 330
N —— 0°":
'1‘
E 225 \ 311
N
S OTMS
I
T
Y 73

.. x a

149 I”
.. 311
279
93 225
130 157 292
.. I 332
so:
— - A l 4
[TI TTTTIIIITITIII TITU ITII I I IIWIIIIIIITTTIIITII

 

 

BO 00 100 120 140 160 $80 200 220 240 260 280 300 320 340 350 380 400
m/z

Figure 3-9. The GC-EI-MS mass spectra of l3-OH 18:2. (A) Only the methyl
ester was prepared, m/z= 310. (B) The methyl ester, TMS ether was
prepared m/z= 382.

52
of the hydroxy linoleate isomers.(8) Normal phase HPLC was used because the products

were methyl esters with methyl linoleate as the starting compound. The molecules
formed are conjugated dienes with a maximum absorption at 235 nm, so the HPLC
detector conventional wavelength of 254 nm was not appropriate. It was discovered that
the conventional wavelength of 254 nm was normally obtained with a low pressure
mercury lamp. The Hg lamp did not have a strong emission around 235 nm. However,
the Biochemistry departmental Beckrnan HPLC detector had a cadmium lamp. The Cd
lamp did have an emission line near the desired wavelength and with a ﬁlter, produced an

intense line at 229 nm.

The separation of the the photooxidized linoleic acid isomers by HPLC was
completed by an undergraduate student. The separation was optimized by varying the
percentage of isopropanol in the mobile phase (hexane) so that the isomers would be
separated properly. The collection of fractions was also optimized by determining the
time between each injection and the number of injections possible without peak overlap

from previous injecu'ons. The best liquid chromatogram is shown below in Figure 3-10.
A

    
 

9-OH 1812 ME

 

 
 
 
 

   
   
 

 

LIJ
‘2 x. 235 nm
0
3) range- 1.00 AUFS
g ﬂow rate- 1.2 ml/min
Ct 0.75% lsopropanol In hexane
O
5
til—J TIME (MINS) >
UJ void volume void volume
0
Figure 3-10. The normal phase HPLC chromatogram of separated photooxidized

linoleic acid isomers. AUFS= absorbance units full scale

 

53
The ﬁrst samples that were collected were examined by GC-FID and found to be
contaminated. This contamination was due to plasticizers. The source of contamination
was believed to be the use of plastic pump priming syringes. The syringes were changed
to glass and this eliminated the problem. Other samples were collected and the puriﬁed

allylic alcohol isomers were stored for later use.

Photooxidation proved to be successful in producing monohydroxy isomers of
linoleic acid (18:2). It was not as successful for arachidonic acid. The larger number of
double bonds increases the number of possible isomers from four to eight; this does not
include the possibility of cis to trans isomerization. Also the HPLC separation of the
monohydroxy isomers of arachidonic acid is much more difﬁcult than that of the
monohydroxy isomers of linoleic acid. Even Porter et a1 (9) do not get complete

separation of their arachidonic acid photooxidized products.
LIPOXYGENASE

The lipoxygenase enzyme catalyzes oxygen uptake by polyunsaturated fatty
acids. The enzyme requires the substrate to contain a 1,4-cis-pentadiene unit. The
enzymatic oxidation is initiated by removal of a hydrogen radical from the methylene
group in the pentadiene unit. The fatty acid radical reacts with oxygen to form a triplet
state peroxy radical and is ﬁnally transformed into a hydroperoxy 1,3-cis,trans-diene.

The lipoxygenase mechanism is found in Figure 3-11.

Preparation of allylic hydroxy model compounds for feasibility studies of their
oxidation to or, B—unsaturated ketones required the use of the lipoxygenase enzyme.
This preparation gives only one or two isomers (the enzymes may not be pure) as
compared to the other mechanisms of oxidation. The lipoxygenase enzymes are a group
of oxygenases that vary in their position of oxygenation. The most commonly used

lipoxygenases are plant enzymes from tomato, potato or soybean. The soybean

Figure 3-11.

lipoxygenase is used most frequently because it is commercially available from Sigma or
ICN Biochemicals.
The important reason for choosing a certain plant enzyme is its positional speciﬁcity.
Different plant enzymes will mimic the speciﬁcity of animal enzymes and are easier to
obtain in large quantities.
oxidize linoleic acid at the (0-10 position((o-16 for arachidonic acid), which yields 9-
hydroperoxyoctadecadienoic acid (9-HPODD) or 5-hydroperoxyeicosatetraenoic acid (5-

HPETE). The soybean lipoxygenase-1 produces an oxidized product at the 03-6 position
which yields primarily (l3-HODD) 13-hydroxyoctadecadienoic acid or

hydroxyeicosatetraenoic acid (IS-HETE). A list of plant enzymes is presented in Table

54

d H
H H
R1 R2

Fe (III)
SLOW ‘

F n

e ( ) 02
FAST 'C

o- o /—

 

Fe Viv m:‘>;HK_:J

The lipoxygenase mechanism.

3-2 (10) on the following page.

(I prefer the ICN product due to its high activity and white color.)

The tomato and potato lipoxygenase enzymes primarily

TABLE 3-2.

ENZYME SOURCE

combined soybean
isoenzymes

soybean-l
soybean-2

H

potato tuber

II

tomato

55

SPECIFICITY OF PLANT ENZYMES

IE”:

6.6
6.6
9.0
5.5
6.3
6.5

REGIOSPECIFICITY*
03-6: (0-10 for 18:2

100:0 (0°C)
44:56
100:0 for 20:4
77:23
25:75
35:65
5:95
90% (S-I-IETE) 20:4
4:96

* T= 25°C unless otherwise
speciﬁed.

56
It is important to note that the regiospeciﬁcity of the plant enzymes is pH

dependent and substrate dependent. The substrate dependence is evident with soybean
lipoxygenase. The speciﬁcity for linoleic acid differs with pH and temperature, but it
does not with arachidonic acid. The speciﬁcity is shown in Figure 3-12. The pH

dependence will be examined later in this chapter.

/-—tomato. potato (VI-10)

___ COOH
s/
1
\— soybean (pH 9)
(VI-6)
/—tomato. potato (VI-16)

COOH

  

\- soybean (VII-6)

Figure 3-12. The speciﬁcity of the most common plant enzymes with (A) 18:2 and
(B) 20:4 as substrates.

The kinetics of the soybean lipoxygenase enzyme have been studied extensively.
A recent review can be found in Biochemistry of Arachidonic Acid Metabolism (10).
Interest in the lipoxygenase kinetics helped determine how to obtain the best yields of
HODDs and HETEs. Since the enzyme had never been used in our laboratory, it was
important that the reaction and its mechanism be well understood. Initially, it was
difﬁcult to prepare large amounts of the desired hydroxy compounds by enzyme
oxidation and this may have been due to a lack of knowledge about the characteristics of

the enzyme. A possible problem could have been the age of the enzyme and its loss of

57

activity. This can be determined spectrophotometrically. The conversion by
lipoxygenase of the diene unit into a conjugated diene allOws for the spectrophotometric
monitoring of product formation and hence the enzyme’s activity. The conjugated diene
formed has an absorption maximum at 235 nm. This is the wavelength that is also used

in ultraviolet detection by HPLC.

The studies of the kinetics reveal that the hydroperoxide product concentration is
important in the activation of the enzyme. The oxygen concentration is important due to
the possibility of product decomposition. There is also some evidence (11) [which has
been disputed by other laboratories (10)] of enzyme inactivation from adhesion th the
walls of the reaction vessel. Other considerations include pH, addition of Ca2+, removal
of iron from the water used, the stability of the formed hydroperoxide, temperature of the
reaction, the concentration of substrate verses enzyme and ﬁnally the extraction,
reduction and puriﬁcation of the hydroperoxide product. All of these parameters will be

introduced and described in detail.

The concentration of hydroperoxide has been shown to be responsible for the lag
time in the production of oxygenated species (11). Also the reduced hydroperoxy
species, the hydroxy unsaturated acid, greatly inhibits the formation of the hydroperoxy
products giving an extremely long lag time. This same time lag was seen by Smith and
Lands (10) by supplying the reaction vessel with ample amounts of glutathione and
glutathione peroxidase which immediately reduced the hydroperoxide formed. This
shows that hydroperoxide is needed for activation throughout oxidation. This same
phenomenon is true for animal-based enzymes such as those found in leukocytes (10).
Addition of antioxidants or radical trapping agents shut down the supply of lipid
hydroperoxide and limit the degree to which the lipid is oxidized. This can also be

expressed initially as a time lag.

The lipoxygenase enzyme (formation of hydroperoxides) may be inhibited by free

58
radicals. This agrees with the reaction mechanism that was developed by deGroot (12).
The participation of iron in this mechanism makes it important that the water used for the
buffer solution is free of any iron (H) ion that might deactivate the enzyme. This can be
accomplished by passing distilled water through an ion-exchange resin (Chelex-lOO Bio-
Rad).

Addition of Ca2+ ion was not tried in our laboratory. For future work the addition
of 0.5 mM Ca2+ would be advantageous. According to Galpin and Allen (13), the
addition of Ca2+ increases the amount of non-micellar fatty acid in solution. The
formation of micelles is a great concern when preparing the substrate (linoleic or

arachidonic acid) for the lipoxygenase reaction.

Since the enzyme reaction takes place in aqueous solution, direct addition of the
fatty acid is impossible. Linoleic and arachidonic acid are not water soluble. The two
choices for substrate addition are the preparation of ammonium or sodium salts or the
"low-ethanol method". Both methods were tried and the low ethanol method was more
successful. The preparation of the fatty acid salts is achieved by adding a dilute solution
of ammonium or sodium hydroxide to the neat acid. The salt is then added to a buffered
enzyme solution. The pH for the buffering solution is chosen to be nine because it
promotes mainly (0-6 peroxidation. The low-ethanol method is accomplished by
dissolving the fatty acid in pure ethanol and then adding it to the buffered enzyme
solution. The reason it is called the low-ethanol method is because the percentage of
ethanol in the buffered enzyme solution should not exceed 3%. It has been found that
concentrations above 3% hinder the production of hydroxperoxide products. (11, 14).
The high concentration of ethanol may yield homogeneous solutions of enzyme-substrate
but the enzyme can be inhibited. This inhibition could be caused by either competition
for the hydrophobic site of the enzyme or a change in the enzyme conformation. Both of

these methods allows the addition of a large amount of substrate without inhibition of the

59
lipoxygenase enzyme. This so-called substrate inhibition is thought to be due to micellar

formation of the fatty acid.

Another critical parameter is temperature. This reaction may be performed
successfully at both 0°C and 25°C. Determining the appropriate temperature to run the
experiments is really substrate dependent. The hydroperoxides formed from linoleic acid
are much more unstable than those of arachidonic acid. This trend can be seen in Table

3-2 on the next page (15).

The production of hydroperoxides of linoleic acid was (Expt. 1) 100% 03—6 after a
ﬁve-minute reaction. The major component was the hydroperoxide. After changing the
pH to 7 (Expt. 2), the major product was still hydroperoxides, but now the
regiospeciﬁcity had changed. Changing the temperature (Expt. 4) to 25°C now produced
major components that were by-products. However, (Expt. 7) showed no by-product

formation at the higher temperature.

Two ﬁnal considerations from the previous tables are the length of time of the
experiment and the difference in the method of substrate addition. Depending on
whether the low-ethanol or salt method was used, the temperature dependence as well as
the substrate concentration varied. The change in substrate concentration may be
connected to each method’s ability to increase the critical micellar concentration (CMC)
of the substrate. The other consideration has to do with reaction time, Long reaction
times tend to increase the by-products as well as isomerization. As mentioned earlier, the
hydroperoxides (especially those of 18:2) are thermally unstable and isomerization is a
problem. Isomerization may not be simply positional, but geometric rearrangement of
double bonds may also occur and this is most unfavorable since separation is difﬁcult.
(16). Low temperatures have been suggested for the reaction, extraction and other stages
of the puriﬁcation process. (17) The isomerization can be minimized by reducing the

hydroperoxide to its hydroxy species immediately after extraction. This is the procedure

8:00:55 500:: n + 80080 00.88 H ++

 

02 o o ++ m ON 0 :xom .505 h
3... on ++ + om mm a Nuwﬁ .200 o
85 o 0 ++ Om c a mum: £3 m
we mm ++ ++ m mm N. ~an .005 0
cm ow ++ + m0 o 0 ~an .305 m
0w 2 o ++ w o N. my: 505 m
02 o o ++ m o a mum: .505 5
0.8 2-8 30:00:95 000080050035 00:05:: 0.
”0008005 00 0000008800 0‘5. 005. 05300809 ma 0:00.53 :95

Wm<ZmU>XOQHA Z<mm>0m m0 maﬁa—<20. ROSA—Oma— .m-m mam—air

61
that has been used in our laboratory. A high concentration of hydroperoxides in solution
is a source of collisionally-induced isomerization. Light may produce by-products as well
as unwanted isomers through photooxidation. Therefore, it is most important that
reduction be accomplished swiftly after extraction or that the hydroperoxides be stored in

a dilute solution in amber vials.

The thermal degradation of the hydroperoxides is also a problem. This is most
likely to occur at longer reaction times and high temperatures (>0°C). It is these
decomposition products that yield the odors mainly associated with rancidity. The
products and mechanisms of formation are described in detail by Chan et a1 (1, 18). The
immediate reduction of the hydroperoxides directly to hydroxy acids by glutathione and
glutathione peroxidase is a possible solution to this problem but it has already been
discussed that this would actually inhibit the enzyme. This isomerization is more easily

eliminated by using a low temperature and short reaction times.

Most of the above discussion was concerned with linoleic acid as a substrate. The
important model compounds formed from arachidonic acid are much more stable and
thermal degradation is not as important. A method that stresses the use of cold
temperatures for HETE production is the Graff Procedure (19). The extraction of the
HPETEs is accomplished with solvents kept on dry ice. The Graff procedure was tried
but yields were small and contamination from phthalates a problem. The source of
contamination was discovered to be in the diethyl ether which was used during the
extraction step. Plasticizers in the plastic cap on the can were extracted by the ether
fumes. Since a large amount of the ether was used and then evaporated to dryness the
phthalate contaminant was concentrated. This problem may be eliminated by covering

the can opening with aluminum foil before capping.

The phthalate contamination of the prepared HETE is just one of a series of minor

problems that led to poor yields and incorrect concentration determination of the 15-

62
HETE. Phthalates are quite a problem in the analyses that were completed. Their Max:
which is close to the km” of the HETEs, may make the collection of the wrong
component a problem, but they complicate analyses much more. The phthalates make it
impossible to determine the concentration of a solution of HETEs by GC. They elute
from the GC at approximately the same time as the derivatized (TMS, ME) HETEs and
the oxidized HETEs, too. The phthalates are electrophilic and show up in ECD
chromatograms. Also they are responsive to electron-capture negative ionization
(ECNI). Thus, they cannot be tolerated in a sample and, if found, the source must be
immediately deﬁned. It is important that these problems be explained for those who

intend to use the various procedures available in the literature.
PRODUCTION OF 15-HETE BY GRAFF METHOD

The preparation of lS-HPETE was tried using the preparation scheme found in
Methods in Enzymology, volume 86 (19). The only difference between this procedure
and many others is the extraction of the lS-HPETE from the aqueous enzymatic solution.
lS-HPETE is not stable and may isomerize. This was prevented by keeping the

extraction medium extremely cold (-70°C).

Arachidonic acid in ethanol is added to Tris-HCl buffer (Sigma Chemical)
adjusted to pH 9. This solution is sonicated before adding it to more Tris buffer which
has been equilibrated at 30°C. Oxygen is bubbled in at a rate of at least 30 ml/min. The
soybean lipoxidase is added and the reaction is terminated after 5 minutes by pouring the
reaction mixture into a separatory funnel with 100 ml of cooled (-25°C) ethyl
acetate/petroleum ether (1:1) and 20 ml of 0.2 _l\_l HCl (adjusts the pH to 3). The organic
phase is collected twice, anhydrous NazSO4 is added and the solution is placed on dry

ice.

The ﬁnal extraction of the organic phase occurs after the aqueous phase is frozen

63
solid. The organic layer (ethyl acetate/petroleum ether) is poured into and passed
through a fritted glass funnel and then evaporated to dryness on a rotary evaporator. The
residue (15-HPETE plus other reaction by-products) is dissolved in methanol. The
methanolic solution is cooled and NaBH4 is added to reduce the lS-HPETE to 15-HETE.
The procedure is followed as previously described. It is this ﬁnal solution that is puriﬁed
by reverse phase (RP) HPLC.

The above synthesis was completed many times with one minor adjustment.
Instead of using ethyl acetate/petroleum ether (1:1), ethyl ether/hexane (1:1) was
incorporated. This mixture is a little less polar. Unfortunately, as will be pointed out

later, this change caused many problems.

The RP-HPLC of 15-HETE was accomplished using methanol/water/acetic acid
65:35:0.03 with 0.5 ml_Vl_ oxalic acid adjusted to pH 5.6 with NH40H (20). This is a very
common mobile phase for the analysis of lipoxygenase products of arachidonic acid by
HPLC. Other RP-HPLC mobile phases involve acetonitrile and/or triﬂuoroacetic acid
(TFA) instead of the methanol and acetic acid. TFA is a volatile acid that can be
removed easily under vacuum. It is a better choice over acetic acid if fractions are to be
collected for further analysis. The wavelength used was 235 nm, as this is the Km” for
conjugated diene moieties. Most recent HPLC work has been accomplished using non-
polar mobile phases with a preparatory silica column (normal phase). Normal phase has
proven to be more successful and will be covered in the ﬁnal lipoxygenase section

explaining the most useful synthesis for 15-HETE.

The determination of the retention time of 15—HETE was the next important step
in the RP-HPLC analysis. Since commercially prepared (+,-) S-HETE ME (Cayman
Chemical) was readily available, an aliquot of this standard was sufﬁciently diluted with
methanol for detection by HPLC (1-10 ng/ml). Analysis by HPLC gave the

chromatogram presented in Figure 3-13; note the large symmetrical peak at 15 minutes.

)

DETECTOR RESPONSE (ARBITRARY UNITS)

 

 

 

 

 

15 mins
TIME >
Figure 3-13. The RP-HPLC chromatogram of a 50 [.11 injection of (+,-) S-HETE-

ME (Cayman Chemicals). The HPLC ﬂow rate was 1.0 ml/min;
Km”: 235 nm; 65:35:0.03 MeOI-I/HzO/HOAC with 0.5 mM oxalic
acid adjusted to pH 5.6 with NH4OH

65
After successive trials, it was concluded that this peak was probably the (+,-) 5-HETE
MB. This was assumed because, as seen in Figure 3-13, there were no other large peaks
in the chromatogram. A methylated sample prepared by the Graff synthesis was then
injected into the chromatograph and a representative chromatogram is displayed in
Figure 3-14. It can be seen by the labelled peak that there is a compound with the
approximate retention time of the standard (+,-) 5-HETE MB. This peak was collected
and determined by GC to contain one compound but it was not immediately veriﬁed by

El-GC-MS to be the actual 15-HETE ME.

The Graff synthesis was tried again with a larger amount (100 mg) of arachidonic
acid. The extraction procedure remained the same; analysis of the reaction mixture
produced a RP-HPLC chromatogram similar to the one in Figure 3-14. Once again a
fraction corresponding to the largest peak in the chromatographic run was collected. It
was assumed that it must be 15-HETE because it absorbed at 235 nm. This fraction was
analyzed by GC after DDQ oxidation (see oxidation chapter) and again only one peak
eluted Figure 3-15.

Due to various problems with the Hewlett Packard GC—MS systems, the identity
of this component was not determined immediately. Before using the GC—MS, the
sample was also chromatographed by GC-ECD. The chromatogram is shown in Figure
3-16. The peak at eleven minutes is the same compound which eluted at seven minutes
in the GC-FID trace. Thus, the component present in the collected HPLC fraction that
had absorbed at 235 nm, showed good chromatographic properties when oxidized with
DDQ and was proven to be electrophilic by GC-ECD. This happened to be only a

coincidence.

When the sample was analyzed by GC-EI-MS the mass spectrum obtained proved

this compound to be a phthalate ester. The most intense peak was m/z= 149.

66

 

 

 

 

 

ﬂ
AA
(I)
t:
Z
3
>-
(I
<
D:
t
m
[I
s
UJ
(I)
Z
O
a.
(D
UJ
0:
CE
0
l—
O
LU
F.
LLI
D
t I I + I I
15 mins
TIME >
Figure 3-14. The RP-HPLC chromatogram of a 50 ttl injection of 15-HETE ME

made by the Graff procedure. The HPLC ﬂow rate was 1.0 ml/min;
Km”: 235 nm; 65:35:0.03 MeOH/HZO/HOAC with 0.5 ml_\_/l_ oxalic
acid adjusted to pH 5.6 with NH4OH

 

 

DETECTOR RESPONSE (ARBITRARY UNITS)

Figure 3-15.

 

67

plasticizer

 

 

 

 

7 mins

TIME >

 

The GC-FID chromatogram of lS-HETE-ME made by the Graff
procedure. The temperature program was 180-280°C at 5°C/minute,
15 m DB-l megabore column.

68

 

 

 

 

>

 

 

 

 

 

 

 

 

DETECTOR RESPONSE (ARBITRARY UNITS)

 

 

 

 

; plasticizer
I
I
I
_I
' 11 mins
TIME >
Figure 3-16. The Sigma 38 GC-ECD chromatogram of lS-HETE-ME made by the

Graff procedure after DDQ oxidation. The temperature program was
180-280 5°C/min, 6 ft. OV-l packed column.

69
There were no ions that would conﬁrm the presence of the oxidized lS-HETE ME
anywhere in the GC-MS data recorded for the sample. Thus, the fraction that had been
collected by HPLC was, in fact, a phthalate ester, and not 15-HETE. Since the phthalate
species absorbed at 235 nm, the Kmax for phthalate esters was researched. According to
the Handbook of Chemistry and Physics the maximum absorbance for n-butyl phthalate
is 225 nm. Since the concentration of the phthalate in this sample was so large the

absorbance at 235 nm was still large.
LANDS PREPARATION OF lS-HETE

The Graff procedure never produced worthwhile yields of 15-HETE. A method
by Lands et al (21) has proven to be the most successful of all those tried previously.
The actual method has been modiﬁed from this paper for the best yields. The differences
in the Lands method over all the others are that the reaction is conducted at 30°C and
phenol is added as a radical scavenger. Precaution was paid to silanize all the glassware
used. The reaction worked well because the paper was carefully written. The
experience gained from the previous failures was also helpful in determining the success

of the procedure.

In a one dram vial, a solution of 50 mg of arachidonic acid is prepared in 3.0 ml
of ethanol. In a silanized 250 ml Erlenmeyer ﬂask, 100 ml of 0.1 M Tris-HCl buffer
(pH=8.5-9.0), and 6.3 mg of phenol (antioxidant), (Fischer) is added with 3.6 mg
(260,000 units/mg) soybean lipoxidase (ICN Biochemicals). The enzyme solution is
stirred vigorously at room temperature while oxygen is bubbled in at a rate of at least 30
ml/min. The arachidonic acid solution is added to the enzyme solution over a 5 minute
period. The substrate/enzyme solution is stirred with oxygen for an additional 5 minutes
and then 5.0 ml of 1M citric acid is added along with 50 ml of diethyl ether. (The diethyl

ether is capped with aluminum foil liner.)

70

This mixture is transferred to a silanized 500 m1 separatory funnel and the ether
phase collected. The water phase is washed an additional two times with 50 ml of ether.
The ether fractions are pooled and washed three times with water. The ether fraction is
then placed in a silanized 250 ml Erlenmeyer ﬂask with anhydrous sodium sulfate. After
drying, the ether solution is transferred to a silanized 500 ml round bottom ﬂask and
evaporated to dryness on a rotary evaporator. The residue is dissolved in a solution of
N aBH4 in methanol and reduced as previously described. The lS-HETE is extracted and
prepared for HPLC puriﬁcation by ﬁnally dissolving it in the HPLC mobile phase (found
in the next paragraph) or pure HPLC hexane.

An aliquot of the prepared 15-HETE is methylated with diazomethane and the
TMS ether is made with BSTFA/pyridine. The GC-EI-MS spectrum and the
reconstructed mass chromatograms are obtained to determine (1) existence and (2) purity
of the nascent lS-HETE. If only one compound is present in the GC proﬁle and that
compound is 15-HETE, further puriﬁcation is unnecessary. However, if there is more

than one compound in the GC trace, HPLC is highly recommended.

It can be seen by the reconstructed total ion current (TIC) in Figure 3-17, that
there is more than one compound in the newly made 15-HETE. The upper trace of the
molecular ion m/z= 406 identiﬁes which peak is actually an HETE. The mass spectrum
in Figure 3-18 show the fragmentation that is expected for the 15-I-IETE isomer. Thus

the presence of other compounds makes it necessary for puriﬁcation.

The HPLC puriﬁcation procedure is from Porter et a1 (9). The puriﬁcation is
accomplished on a normal phase (NP) preparative column (EM Science LiChrosorb Si-
60, 10 mm, 250mm X 10mm). The mobile phase used, hexane/isopropanol/acetic acid
(989:10:1), elutes 15-HETE in less than 20 minutes. The ﬂow rate is set at 5 ml/min.
The Beckman HPLC system used was a with the UV detector conﬁgured for 229 nm.
(The cadmium lamp and the 229 ﬁlter must be placed in the detector.) The lS-HETE is

 

71

.0c:00800 00500 0:: m: 00:00 :0:? 950% 52-5-5 35.0.2 00 £053 0300—08 0::
.090 "mi: :0 000:: 00000 0.:- 00500800 000020 0:: =0 wax/0:0 .0580 65.5 80:50 :0: _0:0: 0:: m: 000:: 032
05- .H0::0 925. .:0:00 03:08 0:: 00 5.5.5.2 00050550,: 00 080Hw0:080£0 0008 0080000000: 22-5-00 2F .2 -m 2&5

caum

Odm OON OWN OON OBN OON OON OVN ONN ONN OdN OON ON« ON.“ Oh.“ 00.0 0nd

 

 

 

 

 

 

.._.rrrr..r-F.r..~.-.._I..._.-.._.-.pp...._..t.__...._...._.p.._...I_I..._.ILIE..tI

I w

I H.

H

I m

r z

I 0

HH... momma.“ lacs 1 a

2

H

._...._...._...p_..r._...._...p_...._...._...._...-_..pt_..p._...._...r_.r+._.pt+

m

H- >

I H

r H.

4

I .H

.- m

I 0

mos vmm taco.“ ﬁ-

 

 

 

 

72

00000 0-8 00: 00 00 0: 0::on 080.3 00: 00 000000 00: 0500000 wmm "NS: :0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

00: :00Ew0b 000. 50500-2 000003000 0:: 0: 000000000000 :00: S-m 00:05 0: 0.000 00: 00 80.00000 0008 5 000- .me 0000.5
N>c
00v 00v can can chm own onn own ann own can can omw com
_IH..._-I.LII.._....Ipprvtpii _...._ .._.I-.EI .. _..__.I_r_......_-.
—_ ._ .— _ .._J_ .H— _ . .2. .4 J:- . : .:_.1_
«ma mom I a
wow I
mHn I H
I m
z
0
v x I a
mmm I 2
H
cam chm omm omm ovm 0mm 0mm 00m 00m and one and 000 000 m
_JHI—F1n1b Fin—u —"—<q_qbtuqb qu—Wuttw b-"RP-u—wut11umuq - bd-———!uq—~ unm-V— “um—“u —n ~— :w :rm p P _Mm F m “ ”Lulu?” “ “leIh— >
_ __ _ _ 0:15:30 _ L: .. H
DVN H.
new mad mm: I 0
I H
I m
and I 0
mmm 1

 

 

 

 

73
injected into a 100 III loop and the detector range is set at 2.0 AUFS. The lS-I-IETE
should be in a reasonable amount of mobile phase- ca. 1-2 ml; this decreases the amount
of injections needed and the preparative column can handle a large amount of material

without showing overload.

The 15-HETE is collected in a silanized-glass, screw-top bottle (125 ml) with a
teﬂon-lined cap. The fractions are pooled and placed in the 500 ml silanized round
bottom ﬂask and the mobile phase is removed in vacuo. The residue is ﬁnally

reconstituted in hexane/ethyl acetate 1:1 (2 ml) and stored in the freezer in the dark.

An aliquot of the normal phase (NP) HPLC puriﬁed lS-HETE is methylated with
diazomethane. It is co-injected into a GC—FID with a known amount of 20:4-ME (Figure
3-19). The area response of the compounds is directly related to their weights and thus
the mg/ml for 15-HETE-ME (15-HETE) can be calculated by ratio. 15-HETE can be
quantiﬁed by comparing the ratios of areas generated by GC-FID of 15-HETE-ME to
standard solutions 20:4-ME. This is accurate because the FID response is based
primarily on the number of carbons present in a compound. The response increases as
the number of carbons increase. The only difference between 15-HETE-ME and 20:4-

ME is one oxygen so the ratios of the areas can be related to the concentration.

The above method has produced milligram quantities of the lS-HETE for our
laboratory. When starting with 50 mg of arachidonic acid approximately 6% yield is
expected. The HPLC puriﬁcation gives reasonably pure product with no arachidonic
acid contamination without using the silicic acid column. The NaBH4 is not a problem
unlike the triphenylphosphine. Triphenylphosphine and the silicic acid column which are
used in the Lands’ article produced problems that could not be tolerated. These problems

are explained in the following paragraphs.

 

 

DETECTOR RESPONSE

74

 

 

 

 

 

TIME (MINS) >

 

20:4 ME 15—HETE ME

1.25 119 of 20:4-ME y ttg of 15-HETE-ME

1.25 rig/2.3 x 10 *6 area units: y ug/ 1.0 x 10 *6 area units

Figure 3-19.

y= 0.35 pg

The GC-FID response comparison of 1 ul each of 20:4-ME and 15-
HETE-ME. The area for 1.25 pg of 20:4-ME was compared to the
area for an unknown amount of 15-HETE-ME. This procedure was
used to determine the percent yield for the lipoxygenase reaction.

75

Initially, the residue from the enzyme reaction was applied to a 2 g silicic acid
column that had been equilibrated with 90:10 pet ether/diethyl ether. The column was
connected to a Milton-Roy LC pump to increase the ﬂow rate. Forty ml of 90:10 pet
ether/diethyl ether eluted through the column. This removed unreacted arachidonic acid.
The 15-HPETE then eluted with 50 ml of 80:20 pet ether/diethyl ether. The eluent was
transferred to the 500 ml round bottom and the mobile phase evaporated. This procedure
was eliminated because the lS-HPETE was coeluting with the arachidonic acid and much
of it was being discarded. Then the residue from the round bottom ﬂask was dissolved in
diethyl ether and triphenylphosphine (TPP) added to reduce the lS-HPETE to 15-HETE.
The molar ratio of HPETE:TPP was 1:1.5. Since there was no way of assessing the
amount of 15-HPETE present, 1.5 times the initial molar amount of arachidonic acid
should provide adequate excess of TPP. This reaction continued for 30 minutes at 0°C.
The TPP was removed by placing the reaction mixture onto the silicic acid column again
and repeating the elution procedure. The TPP was to be removed in the 90:10 mobile
phase and the lS-HETE eluted in the 80:20 solvent. The solvent is again removed in
vacuo and the 15-HETE is stored in hexane in the dark at 0°C. Unfortunately, the use of
the silicic acid column greatly reduced the yield of the lS-HETE. For this reason the use

of this column was completely removed from the procedure.

The TPP used proved to be a source of contamination during HPLC puriﬁcation.

The reaction for the reduction by TPP is shown below:

TPP + HPETE > TPP—oxide + HETE

 

The oxidized TPP or TPP—oxide absorbs at Km“: 228 nm. It is very soluble in alcohol.
The TPP-oxide was supposed to be removed by the silicic acid column, but it was not.
The TPP-oxide was so concentrated that some of it was collected from the HPLC instead
of lS-HETE. In a gas chromatographic analysis of the collected TPP-oxide, it was also

found that the TPP-oxide elutes from a DB-l column at approximately the same retention

76
time as 15-HETE ME. This makes it thoroughly impossible to use. This is why the
method was changed to employ NaBH4 as the reducing agent.

OTHER MPORTANT ISOMERIC HETES

The enzymatic preparation of S-HETE cannot be accomplished with soybean
lipoxygenase; only the (Io-6 position of arachidonic acid is oxygenated with soybean
lipoxygenase. The plant enzyme which yields 5-HETE is potato or tomato lipoxygenase.
These enzymes are not commercially available. However, there are documented
procedures for the production of S-PHSTE. They involve the use of plant enzymes which
have to be puriﬁed, animal enzyme or chemical synthesis for the generation of the
physiologically important HETE. Corey has found the potato lipoxygenase to be cleaner
than tomato lipoxygenase (22). These enzymes must be puriﬁed from the vegetables
directly. This was not attempted here. S-HETE may also be generated through synthetic
means. Corey has developed preparation schemes for both racemic and stereochemically

pure 5-HETE (23). They are larger scale preparations that are quite difﬁcult.

Animal lipoxygenases are much more speciﬁc as compared to the plant enzymes,
probably because the available plant enzymes are not 100% pure. The ﬁrst discovered
animal lipoxygenase was 12-1ipoxygenase in blood platelets. The product that had been
isolated from the tissue was 12-HETE. The enzyme produces 12-HPETE but it is highly
probable that glutathione and glutathione peroxidase or some specialized enzyme (24) are
present so the HPETE is easily reduced. See Table 3-4 for a summary of cell speciﬁc

lipoxygenases and their products.

Since there are so many different sources of the HETEs the methodology
developed here will be useful for many applications. It will be especially useful for
preliminary GC-ECD screening where other general derivatization techniques have

proven be be useless due to the high background from the matrix.

77

TABLE 3-4. SPECIFICIT Y OF ANIMAL LIPOXYGENASES

CELL

rabbit polymorphonuclear leukocytes

rat mast cells
VX2 carcinoma
guinea pig lung
human T lymphocytes
rabbit aveolar macrophages
platelets
eosinophils
mouse peritoneal macrophages
rat pituitary and vascular tissue
human epidermis

peritoneal, blood and lung
monocytes

neutrophils

endothelial cells

(species generated)
species isolated

(5-HPETE) S-HETE
5-, 11- and lS—HETE
11- and lS-HETE
11-, 12- and 15-HETE
5-, 11-, 12-, and lS-HETE
12- and lS—HETE
8-, 10- and 12-HETE
15-HETE
12-HETE

H

5-, 12- and lS-HETE
S-HETE, lS-HETE
15—HETE, ll-HETE

S"

95”.“?

10.

11.

12.

13.
14.
15.
l6.

17.

18.

19.

78
REFERENCES CHAPTER III

Peers, E., Coxon, D. T. and Chan, H. W. S., Lipids 1984, 19, 307-313.
Gunstone, F. D., J. Am. Oil Chem. Soc. 1984, 61, 441-447.

Porter, N. A., Weber, B. A., Weenen, H. and Khan, J. A., J. Am. Chem. Soc.
1980, 102, 5597—6001.

Porter, N. A, Acc. Chem. Res. 1986, 19, 262-268.

Brash, A. R., Porter, A. T. and Maas, R. L., J. Biol. Chem. 1985, 260, 4210-
4216.

Frankel, E. N ., Neff, W. E., Selke, E. and Weisleder, D., Lipids 1982, 17, 11-18.
Chan, H. W. S., J. Am. Oil Chem. Soc. 1977, 54, 100-104.
Thomas, M. J. and Pryor, W. A., Lipids 1980, 15, 544-548.

Porter, N. A., Logan, J. and Kontoyiannidou, V., J. Org. Chem. 1979, 44, 3177-
3181.

Papatheofanis, F. J. and Lands, W. E. M. in Biochemistry of Arachidonic Acid
Metabolism, 1985, W. E. M. Lands ed., Martinus Nijhoff Publishing: Boston, 9-
39.

Gibian, M. J. and Galaway, R. A., Biochemistry 1976, 15, 4209-4214.

deGroot, J. J. M. C., Veldink, G. A., Vleigenthart, J. F. G., Boldingh, J. and
vanGelden, B. F., Biochim. Biophys. Acta 1975, 377, 71-79.

Galpin, J. R. and Allen, J. C., Biochim. Biophys. Acta 1977, 488, 392-401.
Gibian, M. J. and Colanduoni, J., Lipids 1984, 19, 164-170.
Funk, M. 0., Isaac, R. and Porter, N. A., Lipids 1976, 11, 113-117.

Chan, H. W. S., Levett, G. and Matthew, J. A., J. Chem. Soc. Chem. Comm.
1978, 756-757.

Chan, H. W. S., Costaras, C. T., Prescott, F. A. A. and Swoboda, P. A. T.,
Biochim. Biophys. Acta 1975, 398, 347-350.

Chan, H. W. S., Prescott, F. A. A. and Swoboda, P. A. T., J. Am. Oil Chem. Soc.
1976, 53, 572-576.

Graff, G. in Methods of Enzymology 1982, 86, W. Lands and W. Smith eds.,

20.
21.

22.

23.
24.

79

Academic Press: New York, 386-392.
Muller, M. and Sorrell, T. C., J. Chromatogr. 1985, 343, 231-218.

Crawford, C. G., vanAlphen, G. W. H. M., Cook, H. W. and Lands, W. E. M.,
Life Sciences 1978, 23, 1255-1262.

Corey, E. J., Albright, J. 0., Barton, A. E. and Hashimoto, S., J. Am. Chem. Soc.
1980, 102, 1435-1436.

Corey, E. J. and Hashimoto, S., Tetrahedron Lett. 1981, 22, 299-302.

Skoog, M. T., Nichols, J. S., Harrison, B. L. and Wiseman, J. S., Prostaglandins
1986, 31, 577-593.

CHAPTER IV: DEVELOPMENT OF METHODOLOGY

This chapter highlights the main goal of this research project which involved
developing a new strategy for the analysis of allylic alcohols by GC—MS. As is presented
in chapter 11, the most amenable mass spectrometric ionization technique, ECNI, requires
the analyte to be electrophilic. This chapter discusses the progress of an alternative
procedure which leads to the formation of an analyte with inherent electrophilicity. The
stepwise manner in which reaction conditions were optimized for the various reagents
used is illustrated. The importance of reﬁning the methodology for samples containing
biological concentrations of analyte is addressed. The optimized scheme is compared to
the existing approach for the analysis of the allylic alcohols. Finally, suggestions for the

continued improvement of the new method are stated.

When the allylic hydroxy unsaturated acids were prepared from arachidonic acid
and linoleic acid, they were oxidized to at, B—unsaturated ketones. The labile
hydroxyoctadecadienoic acids (HODDs) and hydroxyeicosatetraenoic acids (HETEs),
which were not electrophilic, were converted to the unsaturated ketones. In general,
many compounds containing at, B—unsaturated ketones had already shown
electrophilic character (1, 2) especially those in steroids (3, 4). A goal of this research
was to determine if oxidizing these eicosanoids would make them electrophilic. A
speciﬁc example of an eicosanoid that has been analyzed by detectors that base their
response on the electr0philicity of the analyte are the B prostaglandins (5). The structure
of prostaglandin B1, PGB 1, is shown in Figure 4—1. Later in this chapter, PGBI is

discussed in detail because it was used in methodology studies.

The interest in selectively oxidizing HETEs and HODDs had originated with the
successful experiments conducted in the area of oxidation of exogenous steroids. The

improvement in volatility and electrophilicity of the steroids after oxidation, made them

80

81
amenable to analysis by ECNI-MS. For example, the corticosteroid dexamethasone

becomes electrophilic when it is oxidized eliminating the need for a general electron-
capturing derivative. The use of a selective oxidation reagent over a general derivative

sometimes even eliminates tedious sample preparation.

0

I \/\/\/ COOCH3

/

OH
PGB1 ME

 

Figure 4-1. The structure of PGB 1-ME, MW: 348.

The success of the steroid project inspired the interest in oxidation as an
alternative to derivatization. The physiological interest in eicosanoids as well as their
structure made them an excellent choice as an analyte that had the potential for selective
oxidation. The derivatization procedure for hydroxy unsaturated acids had been well
established by others, but the derivatives did not always lead to a compound that had
improved vapor phase properties over the original analyte. An example of this is found

in the chapter H discussion of the derivatization of lS-HETE.

Another problem with a general derivative is the likelihood that other endogenous
compounds may be derivatized along with the analyte. If these species are of a higher
concentration than the analyte, severe interferences may result. ElectrOphilic derivatives
of eicosanoids that were originally designed to be used in gas chromatography with an
electron capture detector (ECD) are a good example of this problem. In a biological

sample, the pentafluorobenzyl ester ﬂ’FB) derivative greatly enhances the response of the

82
matrix as well as the analyte. If the background is high, the sample cannot even be
qualitatively evaluated without the aid of GC-MS, since GC-MS can selectively monitor
for the analyte. An ECD chromatogram of a cell extract after PFB, TMS derivatization is
shown in Figure 4-2A. It can be seen that little information can be obtained. An ECD
chromatogram of the same sample selectively oxidized (Figure 4—2B) illustrates that, in

this case, there is the ability to determine the presence of the analyte immediately.

For this project, the lipoxygenase eicosanoids of interest were those compounds
that contained an allylic alcohol (HETEs and leukotriene B4, LTB4). The allylic alcohol
has the advantage of being selectively converted to an at, B—unsaturated ketone by
mild oxidizing reagents. The speciﬁc reagents that were tried in this research were
manganese dioxide (MnOz), pyridinium dichromate (PDC) and dichlorodicyano-
benzoquinone (DDQ), all purchased from Aldrich. Initially Mn02 was used successfully
to oxidize the 13-HODD to 13-oxo-ODD. All three reagents were researched as
possibilities for the oxidation of the HETEs. Each converted 15-HETE to 15-oxo-ETE
but the advantages of one reagent over another had to be evaluated. It is felt that PDC or
Mn02 are the most worthwhile reagents to use. Dichlorodicyanobenzoquinone has
caused many problems and it should not be used. This chapter presents the data and the
methodology developed for each reagent. Also, the sensitivity and chromatographic
properties of lS-oxo-ETE compared to the presently established methodology for HETE
analysis (6).

THE SELECTIVE OXIDATION OF THE HODDS

The mildest reagent which has been used extensively in this research is MnOz.
Manganese dioxide is a ﬁne black powder which oxidizes the analyte in a heterogenous
manner. The reaction mechanism is thought to be triphasic (7); the substrate adsorbs to
the MnOz, is oxidized, and then desorbs from the MnOZ. There is an equilibrium on the

Mn02 surface between the substrate and its oxidized product, and reactions never reach

83

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A
full scale: 10'11 amps
AA
(I)
I:
Z
I)
>—
[I
<
[I
b
$ | I I
S 5 10 15
ur
(I)
(2) full scale— 10'12 am 5
a. — p B
(I)
LU
CE
(I:
O 1
l.—
0
ur
|.._
ur
D
I I | I I
I 5 10 15
TIME (MINS) >
Figure 4-2. The ECD chromatograms of a fraction from cell wash that should

contain 12—HETE. (A) The wash derivatized as the PFB ester, TMS
ether. (B) The same fraction was selectively oxidized and then
methylated with diazomethane. The arrows point to the appropriate
elution time of the 12-HEA, PFB ester, TMS ether (A) or the 12-oxo-
ETE-ME (B).

84
100% completion (7). The more readily oxidized compounds are those that form a stable
carbonium ion, which is why the allylic alcohol is selectively oxidized. The reaction

mechanism for the Mn02 oxidation of benzyl alcohol has been postulated (8).

The oxidation is generally completed in CH2C12 or petroleum ether (pet ether). It
has been demonstrated that the Mn02 oxidation proceeds at a much higher rate in pet
ether (9) than in CH2C12. Thus, it is the preferred solvent. The difﬁculty in using MnOZ
as the oxidizing reagent is realized when an attempt is made to extract the analyte from
the reagent. The ﬁne black powder is not easily removed. In preliminary studies,
exhaustive ﬁltering with a 0.45 um pore ﬁlter proved to be the most facile way of
removing the MnOz. Most recently, Mn02 has been removed from the pet ether by
centrifugation. The solid MnOz is then washed 2-3 times with excess pet ether to
completely remove the analyte from the Mn02 precipitate in the centrifuge tube. The
development and optimization of the Mn02 oxidation are presented in the following

paragraphs.

The initial preparation of an at, B—unsaturated ketone from an allylic alcohol
was accomplished with enzymatically prepared 13-HODD. The Mn02 reaction and its
selectivity are shown in Figure 4-3. The oxidation of the 13-HODD was accomplished
with MnOz in petroleum ether. The reactions was conducted for 3 hours. The reaction
was Stirred constantly to expose fresh Mn02 to the analyte. The MnOz was removed by
passing the solution through a ﬁlter. The EI mass spectrum of the l3-oxo—ODD-ME
from the reaction is presented in Figure 4-4A. The fragmentation pattern confirmed that
the at, B—unsaturated ketone had been prepared. The methodology had to be reﬁned to
yield reproducible and complete oxidations. This also involved perfecting the extraction
protocol so that the analyte would be separated from the oxidizing reagent. The reﬁned

methodology would then be applied to the oxidation of the HETEs.

The Mn02 oxidation was optimized by an exhaustive study where the reaction

85

 

 

OH 0
/
O—CHJ O—CH:3
II II
0
M002
13-0H 18:2—ME > 13—KETO—18z2—ME
Mn02
12-OH 18:0—ME \/ > 12-KETO-1820—ME
Figure 4-3. Mn02 can oxidize only allylic alcohols under most reaction

conditions.

86

 

- as 91

67
.4
J 95

  

 

 

 

 

 

 

 

177
J I 194
Ill: A ll4|‘.' .

R 'l“"l

50 BO 70 BO 90 1.00 110 120 130 140 150 160 :70 180 :90 200
‘E
L 4
A. - x 5
T 308
I -I
v d

_ 277

237

 

 

 

1
200 210 220 230 240 250 280 270 280 290 300 310 320 330 340 350

 

 

 

 

 

I
N
T B
E
N
i . ,0 303
T /
Y _
___ O--CH:5
- H
0
11111111 11111111ITT11‘I1111I11111111 111111711 141A1L11 1111]
110 130 150 170 190 210 230 250 270 290 310 330 350
m/z
Figure 4-4. The (A) El and (B) ECNI mass spectra of 13-oxo-ODD-ME (MW:

308), the MnOz oxidized 13-HODD.

87

time, Mn02 stoichiometry and extraction method were varied. The conversion of the
analyte to the desired product was measured. An optimum reaction would yield at least
an 80 percent conversion of the analyte to the desired product in a reproducible manner.
The analyte used for the study was a photooxidized product of linoleic acid. It was found
that the optimal conditions included a 6—hour reaction time with a 25-fold excess of
Mn02 by weight which was extracted with a series of 0.45 mm pore ﬁlters. The percent

conversion of 13-HODD to 13-oxo-ODD is presented in Table 4-1.

TABLE 4-1. THE OPTIMIZATION OF THE MANGANESE DIOXIDE
REACTION

HOURS FOR REACTION PERCENT CONVERSION
3 85
6 97

After their preparation, it was important to determine the electrophilicity of the
CL, B—unsaturated ketones by gas chromatography with an electron capture detector
(ECD). The conjugated carbonyl compounds capture a thermal electron by resonance
capture and a negative ion is formed. In terms of signal response FID chromatograms of
the a, B-unsaturated ketones could be compared to the electron impact (BI) total ion
current (TIC) proﬁles, but the electron-capture negative ionization (ECNI) TIC proﬁles
would resemble the ECD chromatograms. A problem that arose when initially trying to
analyze the a, B—unsaturated ketones was that there was no way to correlate when they
would elute from a gas chromatograph with electron capture detection as compared to a
gas chromatograph with ﬂame ionization detection because the columns installed in the
two detectors were different. In order to relate retention times of the ECD
chromatograph (Perkin-Elmer Sigma 3B) to that of the FID chromatograph (Varian

2100), marker compounds were needed. The column in the Sigma chromatograph was

88
much shorter and the chromatograms looked entirely different than those examined by
FID. Hydrocarbon markers normally used with FIDs could not be used; they were not
electrophilic. Markers were needed that were FID/ECD compatible, i.e., electrophilic

and contained more than 12 carbons.

The importance of marker compounds must be stressed. Typically, when a FID is
used, hydrocarbon standards are coinjected to allow for the calculation of retention
indices (RIs). The "picket fence" produced by the hydrocarbons is used for this purpose.
Changes in ﬂow rate and column packing result in retention times that are variable; with
absolute numbers that may vary by a half a minute or more. Changes in temperature
programming, column length or instrument will also greatly affect the retention time. It
is more advantageous to calculate the retention indices because the number is a ratio and

does not vary as much.

A series of compounds that would mimic the picket fence effect of the
hydrocarbon standards was desired. It was eventually concluded that pentaﬂuorobenzyl
(PFB) esters of fatty acids would work. Since they had to be prepared, it was
advantageous that the procedure was relatively simple. Initially monobrominated
hydrocarbon standards with 14,16, and 18 carbons were tried but they eluted too quickly.
Pentafluoropropionyl (PFP) derivatives of fatty alcohols were not retained long enough
either. Finally 2-bromohexadecanoic acid was examined. It was commercially available
from Aldrich. When methylated it could be used as the early eluting goalpost. (Because
only two markers were actually needed, the use of the word goalpost reﬂects the
appearance of the markers in the chromatogram.) The later eluting goalpost was more
difﬁcult to ﬁnd. Monobrominated 18- and 20-carbon acids unfortunately were not
commercially available. An undergraduate student was assigned the task of determining
a useful compound. The compounds tried were 20:4, 12-OH 18:0, 16:0 and 18:0. The

PFB esters were made of each of these acids as was the TMS ether of the free hydroxyl

 

89
group in 12-OH 18:0. Both 20:4 and 12-OH 18:0 were found to be ineffective markers.
20:4 PFB had too long of an elution time and the hydroxy acid took too long to make.
However, the PFB esters of 16:0 and 18:0 eluted well after the l3-oxo-ODD ME. The
PFB ester of 18:0 has been employed as the ending marker in this research. Other
students in our laboratory have used the series of PFB esters as markers for ECD and
ECNI-MS to determine when compounds elute from a gas chromatograph as compared to
a gas chromatograph-mass spectrometer. Chromatograms of the markers used in this

research obtained with both an ECD and a FID are presented in Figure 4-5.

At ﬁrst the electrophilic nature of the analyte could be tested by GC with the ECD
only. The technique of ECNI-MS had not been learned. The electrOphilic character of
the 13-oxo-ODD-ME was determined with the aid of marker compounds. First, the
oxidized sample was coinjected into a chromatograph with the markers and detected by a
FID. The retention index of the analyte was determined and the identity of the 13-oxo-
ODD-ME was conﬁrmed by GC-EI-MS. The same markers were then coinjected with
the analyte into the ECD gas chromatograph. The retention index was calculated, and it
was found that the l3-oxo-ODD did indeed have electrophilic characteristics. The 13-
oxo-ODD—ME was then analyzed by GC—MS with methane ECNI conditions. The mass
spectrum showed only one ion, the molecular ion. The mass spectrum is shown in Figure

4-4B

The response of the a, B—unsaturated ketone was compared to 12-OH 18:0 as
the TMS ether, PFB ester derivative. The relative response of these two electrophilic
molecules was 1:9 ketone: PFB ester. Figure 4-6 shows the comparison of various
standard fatty acids analyzed by two methods, FID and EC-NCI (ECNI). The tOp trace
shows the response to 60 ng of each compound by a FID. They are essentially the same
because the FID is a universal, mass sensitive detector where the response is dependent

on the number of carbons in the molecule. However, the selective detection by EC—NCI-

DETECTOR RESPONSE (ARBITRARY UNITS)

)

 

FID

 

 

 

 

 

TLhL

2-Br16zo-ME 18:0-PFB

 

ECD

 

 

 

 

 

 

 

IWIK

2-Br 16:0-ME 18:0-PFB

TIME (MINS) >

Figure 4-5.

 

The (A) FID and (B) ECD chromatograms of the marker compounds
2-Br 16:0-ME and 18:0-PFB. The column was a J&W 15 tn megabore
DB-l. The ﬂow rate was 15 mI/rnin. The temperature program was
l80-280°C at 5°C/min.

 

91
MS shows a response only for l3-keto (oxo)-18:2-ME and lZ-OTMS 18:0-PFB. For
each compound approximately 100 ng of sample was injected on-column as compared to
the 60 ng injected into the gas chromatograph. Another important feature of the oxidized
13-HODD over the 12-OTMS 18:0 PFB is the retention time. The PFB ester greatly
increases the retention time of a compound as compared to a methyl ester. Also it has
been determined experimentally in our research and by others (6) that the carbon-carbon
double bonds of the allylic hydroxy acids have to be reduced before the PFB ester of the
acid is made. This oxidation of the carbon-carbon double bonds improved the vapor

phase properties of the original unsaturated compound as was explained in chapter 11.

The success of the Mn02 oxidation of l3-HODD to the l3-oxo-ODD made it
reasonable to try its oxidation capabilities on lS-HETE. The preparation of copious
amounts of lS-HETE was not easy; thus, many of the initial Mn02 oxidations were
conducted with small amounts of the substrate. Because Mn02 was a heterogenous
catalyst, it was possible that the oxidized analyte was not extracted from the solid
reagent. Many preliminary reactions were tried, but the lS-oxo-ETE was never
successfully detected by E1 or ECNI GC—MS. At the time, it was felt that the lack of
oxidized product was due to the mildness of the reagent. It was discovered later that
Mn02 does successfully oxidize lS-HETE to 15-oxo-ETE. The application of Mn02 as
an oxidizing reagent is recommended because of its mild properties, but extremely long

reaction times (>10 hours) are necessary. This limits its usefulness.
THE SELECTIVE OXIDATION OF THE HETES

Although the original preparation of a, B—unsaturated ketones from
lipoxygenase products of linoleic acid was accomplished using the reagent MnOz, as
discussed above, this had proved unsuccessful at ﬁrst for the oxidation of the
lipoxygenase products of arachidonic acid, the HETEs. Thus, it was necessary to explore

other selective oxidizing reagents. The reagent had to have mild oxidative properties like

 

92

 

60 ng each ON—COLUMN
18:0-ME 12-OH 13—KETO 13-OTHS 12-OTUS
18:0-ME 18:2-HE 18:2-ME 18:0-PFB
7.5 11le 11.5 HINS 13.8 MINS 13.2 11le 19.7 1.4le
EC—NCI A- 10,000 A- 91.000
Tll A
M M _ W
(105 ng) (100 no)
Figure 4-6. The chromatographic proﬁles of various compounds as detected by

FID and ECNI (EC-NCI).

 

93

those of MnOz. Most importantly, it had to oxidize allylic alcohols faster than isolated
alcohols. It was necessary to know if the selectivity would be compromised if reaction
conditions were modiﬁed from those presented in the literature. For example, the
preferential oxidation by Mn02 for allylic alcohols is decreased when higher
temperatures are used during oxidation (7). This would lead to the production of
unwanted secondary oxidation products. If the selectively was not compromised, would
the oxidation form only 0t, B—unsaturated ketones? Some reagents may be strong

enough to form carbon-carbon double bonds as well as carbonyl groups. The oxidation
had to stop at the formation of the GI, B—unsaturated ketone. Formation of various

secondary products would yield compounds with different molecular weights and
chromatographic retention properties. It was important that the analyte be converted to

only one product.

Other beneﬁcial qualities of a reagent had to be deﬁned; the two most important
were reaction time and sample clean-up. Short reaction times were beneﬁcial because
reactions that required 24 hours for completion would not lead to high sample
throughput, making sample processing long and tedious. This would certainly outweigh
the advantages of a new technique. If the reagent could not be removed easily or the
percent extraction of the analyte was poor, the reagent was not desirable. The
contamination of the analyte with unused reagent could produce high background during
analysis or might further oxidize the analyte with time. Also, the inability to extract the
analyte from the reaction mixture could make further analysis impossible. It was
possible to choose the best reagent for the HETEs by assessing these attributes for a
given reagent. Each allylic alcohol might require a slightly different reagent or a

different reaction condition for the greatest success.

 

 

94
DDQ AND PDC

The oxidation reagent dichlorodicyanobenzoquinone (DDQ) was evaluated
because publications (10, 11, 12) showed its selective oxidation of the allylic alcohol in
the 3-position of gangliosides and sphingolipids. Other literature on DDQ disclosed its
vast use in the oxidation of allylic alcohols of steroids (13). The preferential oxidation of
allylic alcohols over saturated alcohols is shown in Figure 4-7. In both cases preferential
oxidation of the allylic hydroxy moiety is demonstrated. Thus, DDQ had satisﬁed the
major criterion of selectively oxidizing the allylic alcohols to 0t, B—unsaturated
ketones. Unfortunately, another reaction which might occur along with the
dehydrogenation of the alcohol is dehydrogenation to form a carbon-carbon double bond
(13). Figure 4-8 shows how this occurs with steroids and how it might affect an isolated

allylic alcohol.

The DDQ oxidation was found to be a kinetically favorable reaction with
successful formation of the ketone. The reaction is shown in Figure 4—9. The I-IETEs
were oxidized by reacting them with a 5:1 molar excess of DDQ for 2 hours. These
parameters were determined to be the most successful for the primary production of the
oxo-ETEs. The oxo-ETEs were puriﬁed by removing the reaction solvent under N2,
adding 0.1 _M NaOH, and then extracting with diethyl ether. Unfortunately, there were
problems with using DDQ as a reagent. First, it seemed to produce a multitude of
secondary oxidation compounds. The appearance of dehydrogenation (loss of H2) and
dehydration (loss of H20) compounds, conﬁrmed by El and ECNI-MS, showed the
reagent to be a much stronger oxidant than MnOz. These compounds were present when
the effectiveness of DDQ was initially tested on photooxidized arachidonic acid samples.
The photooxidation had yielded the eight isomeric HETEs and possibly dehydration
products. Since the initial mixture was complex, many of the compounds that were

thought to be secondary oxidation products actually could have been products of the

 

 

 

95

 

D o
:1:
U
U
0
1

HO . 0 .

OH

HO ' HO O
OH O

 

U

O

O
I

Figure 4-7. The selectivity of DDQ for allylic alcohols in steroids (from ref. 13).

 

96

 

OH OH
CH3 DDQ CH3
. ___...
HOCH OHC
O O

 

DDQ
W ’- A/W
OH 0
Figure 4—8. The ability of DDQ to oxidize carbon-carbon bonds and the possible

effects on an allylic alcohol (from ref. 13).

97

 

 

 

Figure 4-9. The reaction between HETEs and DDQ.

98
photooxidation. When purer standards became available it was discovered that the same
types of products were formed when pure lS-HETE or S-HETE was oxidized with DDQ.
The presence of (M - H2)+ and (M — H20)+ were concluded to be from the DDQ

oxidation and not from the initial photooxidation.

Because of its complexity, the photooxidized HETE solution hindered
determinations of the FID chromatographic properties of the newly formed 0t, [3—
unsaturated ketone. There was such an overlap of structurally similar compounds in one
area of the gas chromatogram that the ketone products were not adequately separated
from the initial reactants. This can be seen in the TIC proﬁle in Figure 4-10A. Thus the
gas phase properties of the oxo-ETE could not be determined at ﬁrst. This problem was
magniﬁed by the fact that the electron impact mass spectrometry (El-MS) reconstructed
mass chromatograms of the oxidized HETEs showed the formation of compound which
had a molecular weight two mass units less than expected which was consistent with the
formation of a ketone and a carbon-carbon double bond . This phenomenon was
reproducible and characterized by a series of fragment ion peaks appearing two mass
units below those corresponding to the expected ketone. These compounds coeluted and
made it impossible to accurately assess the vapor phase properties of the ketone. The
GC-EI-MS molecular ion mass chromatographic proﬁles for the molecules are shown in
Figure 4-10B and C. One of the reasons for oxidizing the allylic hydroxy group to the
ketone was to improve the chromatographic properties of the allylic alcohol. The
convoluted mass chromatograms had not produced conclusive results about the
chromatographic characteristics of the HETEs. Also, the complexity of the sample had

made it impossible to determine the optiminal oxidation conditions.

 

 

Pl‘d *4 *3 3‘ t‘lﬂ #1

I4 ta +4 0322 [U ra zzta

Figure 4-10.

 

99

 

 

 

 

 

1111 1111l1111[1111 1111l1111‘[1j11"1111l111111111

 

 

 

 

 

 

 

 

200 225 250 275 300 325 350 375 400 425 450
" 100x- 332
‘ l
V‘TIIIY HTTMI ‘I IIIII‘IEVIYIjij
100x- 330
'VV’V Y zsgﬁﬁl'v V ' 'VY‘ I V V‘ V V V V I 'YMM
200 225 275 300 325 :0 375 400 425 400
Scan

The TIC proﬁle (A) and reconstructed mass chromatograms (B & C)
for the oxidation reaction of DDQ with photooxidized HETEs. The
molecular weight of the oxo-ETE—ME is 332; 330 is the molecular
weight of the overoxidized unwanted secondary oxidation product.

 

 

 

100

Dichlorodicyanobenzoquinone and Prostaglandin B 1

Because of the confusing results obtained from using complex mixtures in the
preliminary studies, it became necessary to chose a purer starting material. The reaction
could be evaluated with lS-HETE, 5-HETE, or prostaglandins (PGs). The 15-HETE had
not yet been prepared in sufﬁcient quantities due to problems that were discussed in
chapter 111. The S-HETE compounds had to be purchased and was quite expensive,
therefore, experimental reactions with this compound were not economically feasible. It
was then decided that prostaglandins be tried. These were used because they contain an
allylic alcohol which is in the same position as that found in 15-HETE. They are pure,
commercially available and reasonably priced. Reactions with dichlorodicyano-
benzoquinone could be executed using a simple prostaglandin and then, when puriﬁed
lS—HETE was available, approximate reaction conditions already would have been
established. Prostaglandin Bl (PGBI), Sigma Chemical, was chosen due its structural
simplicity. It had no isolated hydroxy groups in the 9 or 11 position like many of the
other PGs. The structure of PGB] is found in Figure 4-1.

The solvent used in the DDQ oxidation greatly inﬂuenced the products generated.
For example, the PGBl was initially reacted with DDQ in dioxane for 10 minutes, 1, 2
and 3 hours. The molar ratios of DDQ2PGBl were 18:1. In each case the reaction
mixture contained four different compounds in different ratios. At ﬁrst the most
predominant compound was the unoxidized PGBl. However, after 3 hours the most
predominant compound was the dehydration product of PGBl. They were identiﬁed by
GC-ECNI-MS and reasonable structures that were deduced from the data are presented in

Figure 4-11.

Originally it was not known if the various products were due to the action of
DDQ, the solvent employed or the age of the PGBl which was about two months old and

may have degraded. The choice of an appropriate reaction solvent had been based on the

 

 

101

   

MW= 346

MW= 332
.o"\/\/\/COOCH3

MW: 348

  
  

o DESIRED COMPOUND

Figure 4-11. The possible products of the reaction of PGBI with DDQ/dioxane
which are consistent with the ECNI mass spectrometric data and
comments from ref. 16 which pertain to the formation of secondary
oxidation products.

102
solubility of DDQ (13). Many of the steroid reactions had been conducted in dioxane.

Dichlorodicyanobenzoquinone is most soluble in dioxane. It was necessary to compare
dioxane to another solvent to determine if the PGB] reactions discussed above were a
result of the choice of reagent, solvent or analyte. New PGBI was purchased and the
oxidation experiments were conducted again. The two PGBl reactions were conducted
with different solvents: dioxane and dichloromethane. The solubility of DDQ in CH2C12
is slightly less than in dioxane. The reaction times were 2 hours for each solvent and the
DDQ molar ratio to PGBl was 5:1. Chromatograms examined immediately after
oxidation and 18 hours later showed differences. The 18 hour sample disclosed the
formation of only one new compound. Thus the presence of four compounds was due to
both the dioxane and the old PGBl. It was then decided that CH2C12 would be used as
the solvent. To show the difference between the original and ﬁnal reactions, the ECNI
TIC proﬁles for a one hour DDQ/dioxane from the original reactions and for a 15 minute
DDQ/CH2C12 reaction are presented in Figure 4-12. The experiment with DDQ in
dioxane showed four compounds but CH2C12 gave only the desired 15-oxo-PGB1. The
EI and ECNI mass spectra of 15-0xo-PGB1-MB are shown in Figure 4—13. The EI mass
spectrum was obtained to conﬁrm that the desired molecule had been made. The
fragments at (a) m/z= 317 (M-31)+ and (b) m/z= 249 (M-99)+ show that the molecule is
(a) a methyl ester and (b) a ketone. Loss of -OCH3 is common for methyl esters and loss

of 99 is from an (it-cleavage at the lS-ketone.

The ﬁrst oxidation procedure for prostaglandin B1 began with methylation of 100
ug of PGB] with diazomethane. The excess diazomethane and ether were removed
under N2. The analyte was dissolved in 100 pl of dioxane or CH2C12. Approximately 1
mg of DDQ was added and the reaction mixture stirred for 0-3 hours. The various
reaction times were used to establish the optimum reaction time. It was found that the
DDQ reaction was 90% complete within 30 minutes (see Table 4-2). The solvent was

removed under N2 at the end of the desired time. One hundred microliters of 0.1 l_Vl_

 

 

 

 

.. 100x- 4360 . W Txc

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

wwwwwwwwwwwww ﬁvwwvvwﬁw .wvvvvv v 4v ..7 7v
.3 roox- 3124. 350
R '1
1
E T 1 1 1 1 1 1 1 1 I 1 I 1
L
A I. roox- 2459. 34.
T '2
I “ V-
T 1 r r r 1 1 1 T ‘1 1 “'7 r T
V
E I: 100x- :6. 346
3 ..
I a 1 v 1 1 1 1 1 A .l o v 1 T 1 1 1 1 1
N .: 100x- 2958. 332
T '2
_I
E
N V I I I 7 1* I I I r' 1"“1‘_' 1‘ v w
S
I -
zoox- 5152. no
'1‘
Y _ B
r
1111I1111111111111Tﬁ11111IIIAI11111111111III1111'114171111'1111 117
2:0 250 270 280 290 300 310 320 330 340 350 350 370 300 390
3C.”
Figure 4-12. The ECNI TIC and reconstructed mass chromatograms of PGB] after

DDQ oxidation in dioxane (A). The TIC of PGB] after DDQ
oxidation in CH2C12 (B). The structures corresponding to the various
masses in (A) are found in Figure 4-11.

104

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

55 249
" 217 A
... 91
169 348
133
105 121
.. 161
199
—I 147 176 317
J J
1111 1111 1111 1 11 1111 1111‘1111]1111‘[1111l1111 1111

R .
E 50 7O 90 110 130 150 170 190 210 230 2150 270 290 310 330 360 370
L m/z
A
T
I
v
E
I
N
T 348 B
E
N

.1
s
I d
T
Y ..

4 11 L
HTTTIW1HIII111I1111]111r]j1111111j11r111111'1111]
200 220 240 260 280 300 320 340 360 380 400
m/z
Figure 4—13. The EI (A) and ECNI (B) mass spectra of oxidized PGB], the lS—oxo-

PGBl-ME (MW: 348). The EI fragments at m/z= 317 and m/z= 249
are explained in the text

 

105
NaOH was then added. This minimized the extraction of the DDQ but it did not prove to
be consistent from sample to sample. The oxidized reaction mixture was extracted with
diethyl ether/hexane in a 1:1 ratio. The solvent was again removed with N2 and the
residue dissolved in a measured amount of suitable solvent such as hexane/ethyl acetate
1:1. The extracted solution always contained residual DDQ that made the solution
appear yellow. Dichlorodicyanobenzoquinone had electrophilic character and would lead
to high background during analysis by ECD or ECNI-MS. It was also possible that

undesirable side products would be produced over extended periods of time.

When the parameters for PGBI were optimized, the addition of the appropriate
stoichiometric amount of solid DDQ was not possible for small amounts of the analyte.
Thus, it became necessary to make solutions of DDQ in CH2C12. These solutions
allowed for the controlled addition of a 5-fold molar excess of DDQ regardless of the
analyte concentration. It was discovered that the DDQ solutions should be freshly
prepared. Such solutions lasted approximately two months if stored in the dark at 0°C.
The degradation of DDQ was obvious when the solution changed color (bright yellow to
dark yellow or orange). It has been found that the best way to evaluate the color change
is to prepare a solution and expose it to room temperature and room light. A color

change will occur within a few hours.
Pyridinium dichromate and Prostaglandin B 1

The many original problems associated with the use of DDQ led to further
research about other possible oxidation reagents. Articles that had described the
attractive feature of short reaction times of the oxidation of prostaglandins with
pyridinium dichromate (PDC) (14, 15, 16) were encouraging. The oxidation of the
allylic lS-OH group of a PG to a ketone had produced the CI, B—unsaturated ketone.
This reaction had proved helpful for the authors because the ketoprostaglandins absorbed

at a higher wavelength than the initial hydroxy prostaglandins (190 nm to 230 nm). The

 

106
ketoprostaglandins could be determined by HPLC with an UV detector without

interference of solvent absorption. Normally a general derivative had been prepared to

shift the prostaglandin absorption maximum to longer wavelengths.

The PDC procedure was evaluated in our laboratory with PGBl. The reaction
was a success after forty minutes and the separation of the PGBI was quite facile since
PDC is not soluble in nonpolar organic solvents. Initial PDC oxidations of PGBl started
with the methylation of PGBI. The excess diazomethane was removed under a N2
stream and 10 equivalents of PDC/equivalent of analyte were added in dry acetonitrile
(stored in 4A molecular sieves) . The mixture was left to react for forty minutes or less,
and the reaction was terminated by the addition of enough water to make the acetonitrile
concentration 90% by volume. Then the mixture was diluted with sufﬁcient 10 mM
phosphoric acid to make the acetonitrile 10% by volume. This enhanced the extraction
of the analyte (16) from the aqueous phase, which was accomplished by repetitive
washings with ethyl acetate or hexane/ethyl acetate 1:1. Pyridinium dichromate has no
appreciable solubility in these solvents (14). The ethyl acetate fractions were pooled and
dried. The residue was then dissolved in a measured amount of ethyl acetate or hexane.
The PDC reaction produced only one compound, the oxidized PGBl. The strength of the
separate oxidizing compounds (PDC and DDQ) suggested that PDC was a milder reagent
than DDQ. Also, the rate of the PDC reaction could be increased by adding acetic acid
and/or molecular sieves (17). It was found that molecular sieves complicated the
extraction of the analyte. However, in studies with HETEs, acetic acid increased the

reaction rate considerably.
Optimization of pyridinium dichromate and 15-HETE

After preparing ample amounts of lS-HETE, the PDC and Mn02 reagents were
reevaluated. The Mn02 oxidation was accomplished by placing a known amount of 15-

HETE from the stock solution in a silanized vial. A small amount of Mn02 was added

 

107
(the tip of a spatula) and pet ether added. The reaction was conducted at room
temperature for 6 1/2 hours stirring constantly. The Mn02 was removed by
centrifugation. The residue was methylated with diazomethane. The FID and ECD
chromatograms from this reaction are presented in Figure 4-14A. It can be seen in the
FID chromatogram that there are at least two compounds in the sample. The ﬁrst
compound is lS-HETE-ME and the second, hidden component is 15-oxo-ETE-ME. This
demonstrates that the Mn02 oxidation did not come close to reaching completion (12%)

and a much longer reaction time was required.

The PDC reaction of ﬁve separate aliquots of 300 pg of 15-HETE was monitored
for 1, 2, 3, 4 and 7 1/2 hours. It was found that the reaction reached completion after
seven hours (see Table 4-2). The chromatogram in Figure 4—14B shows the presence of
two compounds in the FID trace at 4 hours. Only one compound was found in the 7 1/2
hour sample (Figure 4-14C). It is interesting to note that the ECD chromatograms do not
show the presence of the allylic alcohol because it is not electrophilic. Thus, anyone
completing these oxidations for the ﬁrst time should be careful to analyze the reactions

by GC with a FID before concluding that the oxidation was 100% complete.

It has been established that the most promising method is the PDC oxidation. It
fulﬁlls all the criteria for an appropriate oxidation reagent. It is a selective oxidation
reagent that is easily removed and reaction times are reasonable. The procedure has been
modiﬁed from the above PGBl oxidation. A 1 mg/ml solution of PDC in acetonitrile is
freshly prepared. These solutions have the same degradations problems that DDQ
solutions have, they last approximately 2 weeks in the dark at 0°C. It is important that
the color be bright yellow and not golden yellow. It is not known if the color change
affects the PDC potency. These shade differences are best tested by the researcher by

leaving a fresh solution exposed to room temperature and light.

 

108

ECD FID

61/2HR

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A HIM L
If
3 A
E I 15-HETE-ME
<2
0: F
t
a \I
S 4 HR
UJ
(I)
Z
O
0.
(f)
LU
0: MN
(I z. I I
9 I I
O . B .
L” "I
l.—
LU
D
7 1/2 HR
C 15-0x0-ETE-ME - 15-0x0-ETE-ME
TIME (MINS) )
Figure 4-14. The ECD and FID chromatograms showing lS-oxo-ETE—ME after

oxidation of lS-HETE with (A) MnOz, 6 1/2 hours; (B) PDC, 4 hours;
and (C) PDC, 7 1/2 hours.

 

109

TABLE 4-2. THE COMPARISON OF THE SUCCESS OF THE VARIOUS
OXIDATION REAGENTS WITH HETES

 

REAGENT REACTION TIME PERCENT CONVERSION
IN HOURS OF HETE TO OXO-ETE
MnOz/pet ether
6.5 12
DDQ/dioxane
(before PGB] optimization)
0.66 23
2 50
DDQ/CH2C12 and PGB]
0.25 86
0.5 92
2 100
DDQ/CH2C12
(after PGBl optimization)
1 4O

PDC/acetonitrile

ADJNH
U1
p—A

7.5 100

l 10

The recommended sample processing procedure is as follows: the analyte is
placed in a silanized vial and the solvent removed with N2. Approximately 200 [.11 of the
PDC solution is added to the vial. A stir bar is added and the vial capped. The reaction
is conducted at room temperature constantly stirring. After 7 hours, the reaction is
stopped by adding enough lOmM phosphoric acid to make the acetonitrile 10% by
volume. The oxidized product is removed with hexane/ethyl acetate 1:1. The aqueous
layer is washed three times and the organic layers are pooled. The solvent is removed
under N2. At this time, the oxidized analyte may be derivatized to the methyl ester or

pentaﬂuorobenzyl ester.

It was discovered accidently that having some hexane/ethyl acetate present during
the oxidation increased the extraction efﬁciency, of the analyte by a factor of four.
Unfortunately, it also made the PDC more difﬁcult to remove. This was apparent by the
color of the organic phase. The organic phase is normally colorless, but in the
sampleswhere the hexane/ethyl acetate was present during the oxidation, the organic
layers were orange. Organic solvents may decompose PDC. This should be determined
before continuing the use of the nonpolar organic solvents during the oxidation. Also, it

is not known how the residual PDC affects the analyte’s stability with time.

Another possibility for the oxidation of allylic alcohols to on, B—unsaturated
ketones is enzymatic oxidation with alcohol dehydrogenase (EC 1.1.1.2). The reaction
incorporates NADP as an hydride acceptor. The speciﬁcity of alcohol dehydrogenase
varies from source to source; so it is important that the enzyme source is known because
many enzymes oxidize only primary alcohols.

alcohol dehydrogenase

 

1° or 2° alcohol > aldehyde or ketone
A

NADP+ NADPH

The reaction is not as speciﬁc as the chemical reagents but it may work. This

 

11 1
reaction was tried on 15-HETE that had been prepared with soybean lipoxygenase and
subsequently reduced with glutathione and glutathione peroxidase. The reaction mixture
initially at pH 9 was acidiﬁed to pH 7.8, and alcohol dehydrogenase and NADP (both
from Sigma Chemical) were added. Five minutes were allowed to elapse and the
solution acidiﬁed to pH 3 and extracted into an organic solvent. After analysis by GC
with a FID, it was apparent that initial experiments were not successful. It should be
tried again because, there may be beneﬁts to enzymatic oxidation such as a short reaction

time and an easy sample extraction.

COMPARISON OF OXIDATION TO CONVENTIONAL DERIVATIZATION

Most mass spectrometric analyses of HETES in biological ﬂuids have been
accomplished by general derivatization of the analyte and subsequent quantitation by
ECNI-MS. This procedure involves the production of a electrophilic pentaﬂuorobenzyl
ester and derivatization of the hydroxy group to a trimethylsilyl ether. Unfortunately,
this derivative has poor chromatographic properties so, prior to derivatization all the
double bonds in the molecule are reduced. Since this procedure is the accepted method,
it was necessary to compare the electrophilic character of the oxidized HETE and the
reduced, derivatized HETE. The substrate was lS-HETE. One aliquot of the S-HETE
was reduced with Rh and H2, then derivatized with pentaﬂuorobenzyl bromide (Aldrich)
and N, O-bistrimethylsilyltriﬂuoroacetamide (Pierce). Another aliquot was oxidized with
PDC and half of this aliquot was esteriﬁed with pentaﬂuorobenzyl bromide and the other
half with diazomethane [prepared according to non-alcoholic directions on Diazald bottle
(Aldrich)]. Each compound was injected into a JEOL HX-l 10 mass spectrometer and the
molecular ion current areas for each molecule were obtained. The response of the ketone
as the methyl ester was compared to the 15-HEA-PFB, TMS and the 15-oxo-ETE-PFB.
The areas were acquired under ECNI conditions by both selected ion monitoring (SIM)

of the most intense ion and regular scanning (SO-500 amu). Figure 4-15 shows the ECNI

112

spectra for the three compounds. Figure 4-16 shows the comparison of the peak shape of
the three compounds. Table 4-3 gives the comparison of the responses of the three
compounds to ECD and ECNI and their retention times. It is obvious that the PFB ester is
more sensitive, however, in a real sample the gain in sensitivity for the PFB ester can be
reduced by matrix interference. This was demonstrated in Figure 4-2. One of the
beneﬁts of using the oxidation instead of the derivatization is that the compound created
has a short analysis time. The methylated ketone elutes quickly in a reasonable
temperature program. Another beneﬁt is that it is not necessary to reduce the double
bonds of the GI, B—unsaturated ketone methyl ester. The only shortcoming of this

methodology is that the detectability is much less than that for a PFB ester, but for some

samples it should be adequate.

A ﬁnal comparison of the oxidation procedure to the published derivatization
methodology was completed with islet cell-generated lZ-I-IETE. A sample donated by
Dr. Pek at the University of Michigan was divided into two aliquots. Each half was
prepared for ECNI experiments, one derivatized as outlined above, the other oxidized and
methylated. The SIM proﬁles for the two samples are presented in Figure 4-17. The
reduced, derivatized aliquot (Figure 4-17A) shows the existence of two peaks at m/z=
399. The earlier peak is 12-HEA-PFB, TMS and the later eluting peak is probably 5-
HEA-PFB, TMS. The arrow in the oxidized sample mass chromatogram (Figure 4-17B)
points to the approximate location of the 12-oxo-ETE-ME. The signal to background
ratio (S/B= 2.4) is not acceptable for positive conﬁrmation of the presence of 12-HETE
even though it looks likely. It is obvious that the signal/background is much more
acceptable (SIB: 14) in Figure 4-17A. This comparison of S/B response demonstrates
the difference in electrophilicity of the two compounds, since each aliquot contained the

same amount of 12- and 5- HETE (approximately 10 ng of 12-HETE).

 

 

113

 

V I V V Yj V f

003.0.3CU'D GC.-ﬂﬂo-‘Ox

 

 

 

 

 

 

I V I I V V V

003.QSCUD nc*ﬂ.-m
4A A ‘3‘ A A$A A ‘g‘ A A$A4 A

 

 

 

 

OODDQDCUD "(Mﬂﬁwﬂm

 

v v 1 1 v v 1 v 7“ '—v W V’T Y 1

 

 

”I

60-

da-I 3a9
I
1

28‘
I

 

 

L1 41 I '1. .J.L ._ jL
1 V T Y Y 1 V V ‘I V’ *1 1 1 v j T Y Y ‘I ~‘I V V ‘V

250 300 350 480 43a
II/z

Figure 4-15. The ECNI mass spectra of (A) 15-oxo-ETE-ME, (B) lS-oxo-ETE-PFB

and (C) lS-HEA-PFB, TMS. The column used was a DB-6l 15m
megabore, ﬂow rate l2ml/min. The CH4 pressure was 7 X 10 ' torr.

 

I”

.A
“‘8
Igg

A A A A A A g

$

A

cnwuannaH ocnatwox
5?

A

 

 

 

A? q Q 1 12.1.
A A A A A A A A A AA A A A A A A A A A A A A A A A A A A

93

 

' V

”4.17... 1 R'T'

 

Cﬂoﬁ-CDORSH OC-twbﬂv-M‘m

 

 

an
.A
8 8'

ccv-u=3o.o:pq o4cuuop-n

 

~31?

 

 

 

Figure 4-16.

V v V Y Y Y

 

 

e399
7800 75130 ‘ ' aa'aa' ' V 05106 1 T 3600' 1 ' '95'00' ' ' £9006 ' Y Scan
The ECNI SIM proﬁles of (A) 15-oxo-ETE-ME, (B) 15-oxo—ETE-PFB
and (C) lS-HEA-PFB, TMS. The proﬁles were obtained on the JEOL
HX-llO mass spectrometer With a DB-l column. Note the poor

chromatographic properties of the 15~oxo~ETE~PFB. The peak shape
is similar to that of the 15-I-IETE-PFB, TMS.

 

\"E _ _.__._ _

 

 

115

TABLE 4-3. PROPERTIES OF VARIOUS PRODUCTS OF CHEMICALLY
MODIFIED lS-HETE

 

 

 

COMPOUND RETENTION MOLAR RESPONSE:
TIMEa ECD ECNI
(mins) (area) (height)
15-oxo-ETE-ME 8.4 1 1
lS-HEA-PFB, TMS 15.4 110 400
lS-oxo-ET’E-PFB 15.0 70b 3013

a: The retention time recorded is from the Varian gas chromatograph.

b= Theoretically, the response for the lS-oxo—ETE-PFB should be 85% (gravimetric
conversion) of the lS-HEA-PFB, TMS. The poor peak shape of the lS-oxo-ETE-PFB
makes the area and height calculations uncertain.

116

 

 

 

 

 

 

 

 

 

 

 

"75,216 A} ......... 19. . .- . . . . .15 ......... its ...... W-
R . 2.6
e
1 .
: I
1 99‘,
V
e 1
I 60
n 1
t
e 1
n 1
s 48-
I .
t
y
20
A
a v y 1 1 f v 1 1 I I v 1 I r v j v r v 1 v V v P399
3000 4000 4200 4400 4608 4893 55999:”
R 1 '
e 1
1
l.
t
l
V
C
1
O
t
e
n
8
I
t
9
20
B
’o v v f v v 1 v I 1 1 v v v 1 v I v 1 332
3000 ' 3§00 4000 4500 5000 5500 0000Scam
Figure 4-17. The ECNI SIM proﬁles of the HETE HPLC fraction from islet cells:

(A) the reduced, derivatized portion, and (B) the oxidized, methylated
portion. The arrows point to the appropriate elution time of the 12-
I-IEA, PFB ester, TMS ether (A) or the lZ-oxo-ETE-ME (B). The
same sample is also presented in Figure 4-2 as the ECD
chromatogram.

 

117

In conclusion, the oxidation procedure is as easy to use as the reduction,
derivatization technique. Application of this methodology to samples where the primary
interest is the presence of allylic hydroxy fatty acids in a matrix of many other higher
concentration fatty acids will prove its usefulness as an analytical method.
Improvements in the oxidation methodology lie in the areas of decreasing the reactitm
times, increasing the number of potential substrates and successfully oxidizing and
extracting small quantities of analyte. As the costs of leukotrienes decrease, it will
become feasible to use LTB4 in experiments. As with the ﬁrst experiments with 15-
HETE and MnOz, there has been no success with the oxidation of LTB4. Since the
sample size of LTB4 has always been restricted to high nanogram- low microgram
amounts, the extraction from the reaction mixture has probably been insufﬁcient. No
analyte was discovered when the samples were analyzed by selectively monitoring the
molecular ion of the oxidized LTB4 under ECNI conditions. If the reaction had not
worked, the unoxidized LTB4 would not have been responsive to ECNI. Since the
original concentration of LTB4 was small analysis by conventional E1 was not possible.
It was conﬁrmed by ECNI-MS that the standard LTB4 was present in the starting
material after derivatization to the PFB ester, di-TMS ethers. We hope that recent
improvements in the PDC and MnOz oxidations will make the the oxidation of LTB4 a
reality. In order to determine compounds such as LTB4 in biological ﬂuids it will also be
important to consider the extraction of very small amounts of the analyte after oxidation.
This can be determined by slowly decreasing the amount of the analyte (IS-HETE) in a
series of experiments. Although many biological samples may contain a much higher
concentration of hydroxyeicosatetraenoic acids compared to leukotrienes, the ability to
oxidize and efﬁciently extract small quantities of hydroxyeicosatetraenoic acids will help

with the extraction of the potent, expensive inﬂammatory mediators.

 

 

59909999?”

10.
ll.

12.

13.
14.
15.
16.
17.

1 18
REFERENCES CHAPTER FOUR

Bowie, J. H. and Janposri, S. J ., J. Soc. Perkin. Trans 11 1976, 1656-1659.

Vessman, J. in Electron Capture, J. Chromatogr. Library, A. Zlatkis and C. F.
Poole eds. 1981, 20, 137—149.

Her, G. H. and Watson, J. T., Anal. Biochem. 1985, 151, 292-298.

Her, G. H. and Watson, J. T., Biomed. Mars Spectrom. 1986, 13, 57-63.
Jouvenaz, G. H. et al, Biochim. Biophys. Acta 1970, 202, 231-234.
Balazy, M. and Murphy, R. C., Anal. Chem. 1986, 58, 1098-1101.
Evans, R. M., Quart. Rev. 1959, 13, 61-70.

 

Gritter, R. J., Dupre, G. D. and Wallace, T. J., Nature 1964, 202, 179-181.
Gritter, R.J. and Wallace, T.J., J. Org. Chem. 1959, 24, 1051-1056.
Kishimoto, Y. and Mitry, M., Arch. Biochem. Biophys. 1974, 161, 426-434.

Iwamori, M., Moser, H. W., McCluer, R. H. and Kisimoto, Y., Biochim.
Biophys. Acta 1975, 380, 308-319.

Ghidoni, R., Sonnino, S., Masserini, M., Orlando, P. and. Tettamanti, G., J. Lip.
Res. 1981,22, 1286-1295.

Walker, D. and Hiebert, J., Chem. Rev. 1967, 67, 153-195.
Corey, El. and Schmidt, G., Tetrahedron Lett. 1979, 5, 399-402.
Doehl, J. and Greibrokk, T., J. Chromatogr. 1983, 282, 435-442.
Doehl, J ., and Greibrokk, T., J. Chromatogr. 1985, 349, 431-438.
Czemecki, S. et al, T etrahedron Lett. 1985, 26, 1699-1702.

CHAPTER V: COLLABORATIVE PROJECTS

This chapter explores the three collaborative projects that expanded our
knowledge of eicosanoid analysis to include leukotrienes (LTs). Each project involved
the analysis of a biological sample. These samples included epithelial cell and islet cell
washes as well as equine plasma and bronchial lavage. An explanation of of each project
is given. A brief summary of the objectives of each project is presented, and a detailed
report of how the samples from each study were prepared for analysis appears in the
experimental protocol section at the end of the chapter. The current, accepted
methodology for LTs was employed for each sample. However, islet cell wash was also

analyzed by the oxidation methodology discussed in chapter IV.

In two of the projects, our interest in leukotrienes was derived from the fact that
LTB4 is present in ﬂuids from cells that have been challenged by an allergen. The goals
of these projects involved determining the presence of LTB4 in the biological ﬂuid.
LTB4 is known for its chemotactic properties. Chemotaxis causes the movement of
immune cells toward the site of inﬂammation. LTB4 is a mediator of neutrophil (white

blood cell) inﬁltration to inﬂamed areas.

The ﬁnal study, in which islet cells were employed , involved determining the
potential of eicosanoids as regulators of pancreatic islet hormone (insulin and glucagon)
secretion. Results of this study could possibly provide information explaining a cause of
diabetes. It is known that shifts in arachidonic acid metabolism result in regulation of
insulin or glucagon. By determining what types of arachidonic acid metabolites are
present in healthy islet cells and how they affect insulin production, diabetic cells could
be compared. The goal of this project was to determine if peptido-LTs are formed in

stimulated islet cells.

It should be stated here that all glassware used, that the analyte might come in
119

120
contact with in these experiments, was silanized prior to use. In each study, the analyte
contained a least one hydroxy group and these molecules tend to adhere to the active sites
on the glass surface. Since the concentration of the analyte is so small, it is possible that
all of the analyte could become permanently adhered to the glassware. To prevent this
problem, Silanization of the glassware is required. The process of Silanization can be

found in the experimental section.
PROJECT I

The ﬁrst project completed was the analysis of rabbit tracheal epithelial cell wash.
This wash was removed after the cells had been exposed to ozone. The presence of
chemotactic factors (LTB4) due to cell injury was to be examined. It had been shown
previously that canine tracheal epithelial cells had produced LTB4 after ozone exposure
(1, 2, 3). Ozone is a potent inﬂammatory agent (4, 5, 6) which causes lung

hyperresponsiveness.

The methodology for cell injury developed by Stephen Alpert, M. D., was to be a
model system that closely resembled in vivo ozone exposure. Dr. Alpert felt that
previous studies may have been too harsh and cells producing chemotactic agents were
not injured, but were dying (not viable). Dying cells might have their membranes
ruptured and membrane-bound phospholipids would be released. This could cause an
abnormally high pool of arachidonic acid to be released, and lipoxygenase enzymes
might be activated. This could cause the formation of arachidonic acid metabolites, but it

would not be due to inﬂammation.

If it was found that epithelial cells produce chemotactic factors that lead to
neutrophil inﬁltration, the contribution of these cells to the physiological effects of
diseases such as asthma and cystic ﬁbrosis would be known. Currently only endothelial

cells are thought to be involved in these diseases; thus the drugs used for treatment affect

 

121
only those cells.
Preliminary experiments with cell wash samples showed no formation of LTB4
(see Figure 5-1.). The samples did show the presence of peptido-LTs. This was
conﬁrmed by radioimmunoassays conducted by Michael Bach’s laboratory at Upjohn,

Kalamazoo, MI. Studies were concluded at this point.

A

 

 

 

 

 

U)

z

0

Cl.

(1)

DJ

E .
0: . MM

0

*— .

o .

E II LTC4 f

LU

0

TIME (MINS) > LTB4

Figure 5-1. The reverse phase HPLC chromatogram of 30 III of epithelial cell

wash after ozone exposure. The ﬂow rate was 1.0 ml/min, range: 0.01
AUFS and l: 280 nm. The arrows show where LTC4 and LTB4
standards would elute. The area corresponding to LTC4 rs equivalent
to the injection of 20 ng of standard LTC4.

PROJECT 11

The second collaborative effort was a preliminary study of LTB4 in equine
plasma and bronchial lavage. This study was not completed and will be continued by

other students in our laboratory.

The study of plasma and lavage for arachidonate metabolites is being conducted

on ponies which are challenged with an allergen which produces airway

 

122
hyperresponsiveness commonly called heaves. Heaves are the equine equivalent to
asthma. Thus, it is an excellent animal model system for human asthma. Although it has
already been shown that heaves are accompanied by neutrophil inﬁltration (suggesting

the presence of LTB4), arachidonate metabolites have not yet been detected.

Lavage and plasma from two ponies (Frisky and Augie) were spiked with LTB4.
These samples were used to test the extraction protocol which was developed by the
collaborating scientists. Also it was important that spiked and unspiked biological ﬂuids
be compared to determine possible basal amounts of LTB4. The actual determination of
the percent recovery of the spiked samples was also of interest. However, this could only
be accomplished by spiking the collected samples with radiolabelled or 18O-labelled
LTB4. Radiolabelled LTB4 recoveries could be determined by liquid scintillation
counting and 18O-labelled LTB4 recoveries by mass spectrometry. Although
radiolabelled LTB4 is available, 18O-labelled LTB4 must be prepared. The preparation
of 18O-labelled LTB4 is described at the end of the chapter.

Preliminary results showed that LTB4 had been extracted from all four of the
spiked plasma samples examined. In the two lavage samples analyzed, one that had been
spiked with LTB4, LTB4 was not found. Also a plasma sample that was not spiked did
give positive LTB4 response that was approximately one-fourth the response of a spiked
sample. This could be the basal level of LTB4. The percent extraction of LTB4 was not
determined with the ﬁrst group of samples analyzed because l8O-labelled LTB4 was not
prepared. Recently, 18O-labelled LTB4 was successfully made and future samples will
have their [LTB4] determined.

PROJECT III

The ﬁnal collaborative project involved the determination of peptido-LTs

produced from islet cells. Islet cells are known to contain 12-lipoxygenase (12-HETE)

123
but it is not known whether 5-lipoxygenase is present. In order for LTs to be present this
would have be the case. The cells had been stimulated with glucose and calcium
ionphore to produce arachidonate metabolites. HPLC analysis of the cell wash from the
stimulated cells showed peaks which corresponded in time to standard LTs. Fractions
were collected and sent to our laboratory from the University of Michigan to be analyzed.
A representative HPLC chromatogram is presented in Figure 5-2. Fraction H was
thought to contain the peptido-LTs and Fraction IV contained 12-HETE. These fractions

were to be analyzed by mass spectrometry.

Intact sulﬁdopeptide leukotrienes are very difﬁcult to analyze by mass
spectrometry. They are soluble only in water; thus, only fast atom bombardment (FAB)
ionization is appropriate. FAB, however, is not very sensitive, which makes it an
unrealistic ionization mode for the analysis of biological samples containing only
nanogram amounts of leukotrienes in the presence of microgram quantities of other
compounds in the matrix. The technique which affords the researcher better sensitivity is

GC-MS.

Gas chromatography requires the analyte to have appreciable volatility. The
labile sulﬁdopeptide leukotrienes have poor vapor phase properties. Possibilities for gas
phase analysis require the derivatization of carboxylic, amino and hydroxyl groups.
Although the derivatization itself is easy the derivatives have higher mass. For example
an isopropyl ester, pentaﬂuoropropionyl amido and an ether derivative of LTC4
(MW=623), is 418 mass units heavier than the underivatized LTC4. This takes the
molecular weight well over 1000 amu. Many mass spectrometers that are used routinely

do not have unit resolution at such a large mass.

An alternative proposed by Murphy and Balazy (7) is the reduction of
sulfidopeptide leukotrienes by cleavage of the cysteinyl sulfur bond (see Figure 5-3).

This is accomplished by using the Birch reduction; this reaction originally was used to

1—

 

 

124

.5322: mo 585323 .xom an mo 8: 2: 80¢ .3sz Etna .uQ E E8 33 EﬁwoﬁEoEo
2E. .mpmm-§ 3:838 >H cough new mkqoucmon Emwcﬁm mm 826 cogent 08mm 2: em: = new H £5:an

.8038 05 “a 88: can 880:8 mcowomt 05. .595 :8 “2mm 2: 90 8833825 0.5: $298.62.: 2E. .N-m. Ramm—

 

 

_.

E ;. 5.. _ “fufyzwg
, . . . :1. G
. . of. .5. i B
wkm1-N_. _. a .. $13 3
u .. ._ . .. . Li 0
. . . . _.; l
_ _ f O
”.2-.. H
..W t H
.1414- 3
.3: S
5 d
__ m: O
.”.__l_-_e N
h_._"“ S
r“ Z 3

___.

 

; -;.;':.:.:--.:.' ‘ ‘ "’
mm.-- ,..

L.,“...‘v
. I.-. .._

 

,’_.‘l

 

 

 

 

 

 

 

 

 

 

 

 

125

OOH

   

COOH

 

HZ/S‘Z. Rh 10 mins 0°C

OH

WOOH

Figure 5-3. The modiﬁcation of LTC4 for analysis by GC-ECNI-MS. The Birch
reduction and the hydrogenation of LTC4 are shown.

 

126

reduce conjugated unsaturated systems such as a, B—unsaturated ketones to alcohols

(8, 9). It has also been used in the preparation of peptides due to its capacity to cleave
sulﬁde bonds (10, 11, 12). The speciﬁc reagent used for the cleavage of most sulfur-
carbon bonds is lithium metal in liquid methylamine. In the case of peptido—LTs, the
peptide is cleaved from the leukotriene skeleton, and the residual double bonds are
hydrogenated. The resulting compound is then derivatized to the pentafluorobenzyl
(PFB) ester, trimethylsilyl (TMS) ether and analyzed by GC-MS. The reduction of
double bonds of the molecule is accomplished simply with hydrogen gas and a rhodium
catalyst. Fission of the sulfur-carbon bond is a much more sophisticated procedure, but it

can be accomplished quite simply if the correct equipment is obtained.

Two different batches of islet cell samples and leukotriene standards were
prepared for GC-MS analysis. In both batches the LT standards were used to conﬁrm
that the Birch reaction and hydrogenolysis had worked successfully. An internal
standard, linolenic acid (18:3) was added to each sample to approximately determine the
losses of the analyte through the sample preparation. Although 18:3 was not a substrate
for the Birch reduction, it was reduced during hydrogenolysis. By comparing the areas of
the (M-181)' ions for 18:0 (m/z=283, hydrogenated 18:3) and coinjected 20:0 (m/z=311),
approximate recoveries for the Birch reduction] H2 reduction were 50%. Figure 5-4

shows the proﬁle of the two compounds and their areas.

The sample fractions (labelled H in Figure 5-2) which were thought to contain
LTC4, D4 and E4 were compared to a blank sample. The blank was from an HPLC
"analysis" in which no sample was injected. The effluent was collected during the
mobile phase gradient at the same time that the peptido-LTs would elute. Since peptido-
LTs have a tendency to adhere to the column, if standards were initially injected to
determine retention times, some of the standard might be retained. If it eluted in a

subsequent analysis, a response that would not be completely due to the sample would be

 

 

127

338805 9 Bandage: 8a :5 .550 :98 9 03:28 8: 8m
8:30: Vang 2F acetone couacowoéz; 2: 98 8532 :85 05 Bob @382 2:083 05 mo couaemxohaw

 

25.8 05 new 339:8 083 32a 2E. .95 onom Ea mmm on? .8 maﬁwouanEo 9.88 Bmantoasm 2:. 41m ouswi
cnum
DNN OON aha on« ”Na cod an on ”N O
___np—ner—FhL—_nrmb—_~—_—L__——[C_—_—P -—-_
:m "NE. mmm NEE
Nov? u< mmmmm u< ..
En 98. 6 9. ow En. 3: no 9.9. .

 

 

 

 

 

 

 

 

 

 

128
recorded. This was found to be true. In Figure 5-5 comparison between a standard, a
blank and a sample is shown. Comparison of the areas showed the samples to have only

twice as much signal as the blank at m/z=399.

The second set of experiments showed an area unit increase of three for the
sample over the blank (Figure 5—6). It is clear from Figures 5-5 and 5-6 that experiments
must be continued and the concentration of metabolites in the cell wash increased before
a deﬁnite conclusions can be made about the presence of S-Iipoxygenase in islet cells.
This will require larger colonies of cells or pooled experiments. Hopefully this is

feasible.

The fraction containing 12-HETE (indicated by IV in Figure 5-2) was not treated
with the Birch reagent; it was not required. Instead it was split and half was oxidized
with pyridinium dichromate. The other half was reduced with H2 and 5% Rh/Al then
derivatized to the PFB ester, TMS ether. The nascent 12-HEA PFB, TMS compound was
also monitored by ECNI-MS at m/z= 399. The 12-HETE was oxidized to 12-oxo-ETE
which was then methylated and selectively monitored at m/z= 332. The comparison of

the two derivatives is explained in detail in Chapter IV.

 

 

 

txr<tHaturmw

Kt-JHMZLIJHZH

Figure 5-5.

 

 

 

 

 

 

 

 

 

 

129
m/z=399
‘ A=1100
(A) LTD4 standard
1111111111111j111 1111 1111 1111111111ﬁ11 111111111
200 210 220 230 240 250 260 270 200 290 300 310
Scan
m/z=399
- A=200
(B) cell extract sample
IWTIT11‘ITT11'11111111111111IW1ITITT1TIIII1II11T‘Ij

 

 

 

 

(C) blank

 

 

200

111 I

[1711 1111 1111 111111111 1111 1111 1117 1111IU1II

2:0 220 230 240 250 260 270 280 290 300 310

The HP 5985A GC-ECNI-MS SIM proﬁles of (A) LTD4 standard,
modified according to the text, TMS, PFB, (B) islet cell wash sample
and (C) a blank derivatized similarly. The temperature program was
60-200°C at 20°C/min (30 m DB-S megabore, on-column injector),
ZOO-300°C at 7°C/min, source temperature= 200°C, EM: 3000 V and
CH4 pressure= 2.2 X 10-4 torr.

 

 

 

130

Hux.11.1 14 R.T.

A A A A A I A AAPA A A

 

an
g
p

j (A) LTD4 standard

lSﬁu-HDGO‘DH OCWﬂDWO
U5
o
l

 

 

 

 

 

 

 

 

 

 

 

 

 

 

28«
d
4999 V V ' 4288 ' ' 14488 7 V ' 4588 V V 4888 588856”,
ﬁnal 8
R188 A . 1:4 4 .1? 1p RT
9 ‘ 131
i .
g 88~ (B) cell extract sample
V 1
e 1
. m/z: 399
I 594
n 1
t
e 1
n 1
c 484
1 1
t 1
u 1
ZB~
J
4388 ' ‘ ' T ‘ ' '
4 1 v v 1 f v .
288 4488 4688 4888 5888Scan
None 3
4 R.T
R188 AAAAA All A A A A4; +41§AA A A A AA A A1pA4 gL4_A_
. 8.31
C
1 1
3 1
‘ 88-
‘ J
V
I nk
° 3 (C) be m/z= 399
I 53..
n 1
1.
e 1
n l
s 48‘
‘ i
1» .
U 1
28-
a ”v— v V v r —V v r V V 1 1 ‘ ' '
4888 4288 4488 4688 4888 58885can
Figure 5-6. The JEOL HX-llO GC-ECNI-MS SIM proﬁles of another islet cell

sample after the Birch reduction and derivatization (PFB ester, TMS
ether): (A) standard LTD4, (B) sample and (C) blank. The
temperature program was 180-280°C at 5°C/min (DB-1 15m
megabore), CH4 pressure 5 X 10-5 torr, source temperature 200°C.

13 1
EXPERIMENTAL PROTOCOL

S ilanization of Glassware

The process of glassware Silanization inactivates the hydroxy groups on the glass
surface by chemically converting them to methoxydimethylsilyl ethers. The glassware is
prepared by soaking it in a solution of 5% dimethyldichlorosilane (DMCS), Sigma
Chemical, in hexane or toluene for at least 20 minutes. The DMCS solution is toxic and
should be handled with gloves while wearing goggles. The glassware is rinsed with
hexane to remove the excess DMCS and then with methanol to exchange the residual
chlorine with a methoxy group. The glassware is rinsed a ﬁnal time with hexane and it is
then ready for use. The glassware may be slippery so care should be taken during

handling. The process of Silanization is shown below in Figure 5-7.

 

 

 

 

/- Si—OH DMCS /— Si—O—Sl(CH3)2Cl
/—'Sl-—-OH hem“ /"Si—O—SI(CH3)2CI
MeOH

/r- SI—O—SI(CH3)ZOCH3

/— Si—O—Si(CH3)2 OCH3

/‘

Figure 5-7. The Silanization of glass to derivatize the active sites on the glass

surface.

Project I

The analysis of the cell wash for leukotrienes (LTs) was accomplished by HPLC.
Standard solutions of LTB4, LTC4, D4 and E4 were a gift from J. Rokach (Merck-Frosst,

 

132

Canada). Standards were necessary to determine the HPLC retention times and detection
limits for the LTs. The development of an appropriate reverse-phase mobile phase was
important It had been shown that a fraction of peptido-LTs were permanently retained
on a ODS (octadecylsilyl) C13 column. This led to poor concentration reproducibility
and variable recovery (13). By using various EDTA sodium salts [NazEDTA (13) and
Na4EDTA (14)] this recovery problem is eliminated. EDTA also chelates the metal ions
that may be the in the mobile phase. The additive is helpful in preventing the
degradation of the peptido-LTs. LTC4 decomposition is thought to be caused by the
presence of transition metal ions (15). The problem with EDTA salts is that they
precipitate in the presence of methanol. This will clog the chromatographic system. The
system must be ﬂushed with 100% water before and after the EDTA salts are added. The
ﬂushing is usually completed overnight. Unfortunately, constant changing of the mobile
phases leads to long and frequent equilibration times. Long equilibration times made the
use of EDTA tedious. Oxalic acid was added instead of EDTA. Oxalic acid can be
added directly to the mobile phase; it is soluble in methanol. The oxalic acid works the
same way EDTA does; both are chelating agents. The actual mobile phase employed
was 70:30:0.03 MeOH/HzO/HOAC with 0.5 mM oxalic acid adjusted to pH 5.6 with
NH40H. In Figure 5-8, the chromatographic proﬁle of the standard LTs is shown as they
elute from a 10 um particle 4.6 mm X 250 mm Ultrasil ODS C18 (Altex).

After the retention times of the LTs were known, it became necessary to
determine their detection limits in the Beckman system that was used. The detection
limit of LTB4 was determined to be 2 ng and the peptido-LTs had a limit of 4 ng (range:
0.01 absorbance units full scale, 71m“: 280). The difference in detection limits is due to

the different molar absorptivities for the two types of LTs.

 

 

 

 

 

 

 

 

 

A R
1 133
LTD4 ‘
LTC4 LTB4
\ LTE4
LU
U)
2
O
a
(D W
UJ
m l
(I ‘. I
O i
S
:2 I I I I
LOU 5 10 1 5 20
TIME (MINS)
Figure 5-8. The reverse phase Ultrasil ODS C 8 HPLC chromatogram of 50 ng of

LTC4, D4, and E4, and 25 ng L 4. The flow rate was 1.0 ml/min,
range: 0.02 AUFS, and 2.: 280 nm.

In order for the project to be successful, the extraction efﬁciency of a leukotriene
standard had to be determined at each step of the sample processing. This required
methodology which would preferentially extract the LTs from the cell wash media
effectively. The recovery experiment was conducted by spiking cultured cells with
LTB4. The cells were exposed to ozone and the cell wash collected. The cell wash was
added directly to methanol so the LTB4 would not be degraded. The methanol denatured

any proteins that might decompose the LTB4.

The eicosanoid extraction was facilitated by the use of C13 Sep-Paks (Waters).
The cell medium was initially acidiﬁed to pH 3 with 1M citric acid. It was then
introduced onto a Sep—Pak which had been previously conditioned sequentially with 20
ml of ethanol and water. Methyl formate eluted the hydroxy eicosanoids (HETES and
diHETEs) preferentially (16). The methyl formate was evaporated and the residue
dissolved in the mobile phase. Representative HPLC chromatograms of the cell media

before and after LTB4 spiking are shown in Figure 5-9. Recoveries of LTB4 were

 

 

DETECTOR RESPONSE (ARBITRARY UNITS)

Figure 5-9.

134

(A)

 

LTB4

signal attenuated by a factor of 2

 

 

 

 

.4

l

1
l LTB4 spiked

signal not attenuated

 

 

 

 

5 10 15 TIME (MINS)

The reverse phase HPLC chromatograms of 100 pl of LTB4-spiked

(A) and unspiked (B) cell wash. The ﬂow rate was 1.8 ml/min, range:

0.01 AUFS and 7L: 280 nm. The area corresponding to LTB4 in (A)
is equivalent to the injection of 60 ng of standard LTB4.

 

135
calculated to be >70%.

After determining the extraction efﬁciency for the methodology that would be
used, samples that had been exposed to ozone were analyzed for LTB4. Initially the
levels of ozone were quite small, to mimic true in vivo concentrations. However, there
was no formation of LTB4 that could be detected by HPLC. More extreme conditions
were tied and ﬁnally exogeneous arachidonic acid was added along with calcium
ionphore. These ﬁnal samples did not produce LTB4, but peptido-LTs were detected.
An example of an HPLC chromatogram showing LTC4 was shown in Figure 5-1.
Further progress in this project was discontinued by the collaborator because of

instrumental availability conﬂicts.

 

136
Project II

To determine whether LTB4 would be successfully extracted from the equine
lavage and plasma samples, preliminary studies with standards were tried. These
samples had been extracted through a detailed procedure and were received in methanol.
The methanol was removed by nitrogen evaporation and the residue was derivatized to
the pentaﬂuorobenzyl (PFB) ester, trimethylsilyl (TMS) ether compounds (17). The
plasma samples were easily derivatized but the lavage samples turned cloudy and thick
during the PFB reaction. This may cause a subsequent extraction problem as well as
incomplete formation of derivatives resulting in lavage samples that may not be correctly
assessed for LTB4. It is suggested that blank bronchial lavage be spiked with 12-OH
18:0 and tested for recovery by GC—ECD before other precious samples are inadvertently

destroyed by just derivatizing the lavage as received.

A standard solution of LTB4 was also derivatized to the PFB ester, TMS ether.
The ECNI spectrum of the derivatized LTB4 is shown in Figure 5—10. The samples were
analyzed by GC-ECNI-MS. The (M-181)' ion (m/z=479) and the (M—181-90)' ion
(m/z=389) were selectively monitored for conﬁrmation of LTB4 in the plasma and
lavage. (Loss of 90 is common for TMS ethers. It corresponds to the loss of

trimethylsilanol, (CH3)3SiOH.)

Selected ion monitoring (SIM) proﬁles of spiked and unspiked plasma and lavage
are shown in Figure 5-11. The GC-MS parameters are also presented. The ions 389 and
479 were monitored for equal time intervals, but they may be monitored for different
times because their intensities are different. The intensities for 479 and 389 vary with
GC-MS conditions. Higher source temperatures and high primary electron voltages
(>200 eV) will lead to more fragmentation, i.e., more m/z= 389. The daily ﬂuctuation of
the two ion intensities makes it necessary to establish their ratio initially with standard

LTB4. This should be the accomplished before the biological samples are analyzed. The

 

137

 

.558? c.8200 co EEO. N. a 0. com- com EEO. om 8 0. com- co .828on 838an8
00 ES 3 N H .m "8335 emu 0. com I U. 08:8 .EBoEEa m2 .:2 .8m :2- <6 2: 9:82am: owm NEE 3
via 05 %a com -Cwﬁ -Sc 2: $54.29.: 0:» "NE a Vang 2:. amu.; 3854:. “68536.8 Mo 83809. Eom of.

NE.
and can ouv onv 08.. 08.. 3.. own cum onm onm can own cum onm omm 3m on... a: one

nPb——-bF——numb—nnnn_:-__-———-hb_—-——-n——--—— n—n—np

___ _

 

 

 

 

 

u. .._ E
N I
Elm IUOOQ

mwm 925 £25.

at. t a ..

 

 

 

 

HZE-rmZmHE-rw

mWQ¢BH>txJ

.28 25mm

 

 

m<lHi—3:vt*mw

KHHUJZEIl-JZH

Figure 5-11.

138
(A) LTB4 standard

50 ng
m/z= 389

 

(B) LTB4 spiked plasma sample

"20 ng 1
ml '1‘."
A I] "(1 [Tl/Z: 389
V; /“ \\\//r\\j\_ J — \Nx
l!
H
II K 179 AU
Rib/v '\~_ m/z: 479

(C) unspiked plasma sample

\\ [Tl/Z: 389

 

19 20 21 time in minutes

The HP 5985A GC—ECNI—MS SIM proﬁles of TMS, PFB derivatized:
(A) standard LTB4, (B) LTB4 spiked plasma and (C) unspiked plasma.
The GC-MS parameters are listed in Figure 5-5. AU= area units.

 

139
ion at 389 is used only as a conﬁrmation ion. The retention time of the derivatized LTB4
also can be used for conﬁrmation. Thus, it is not always necessary to monitor both m/z=
389 and m/z= 479. When 18O-labelled LTB4 is incorporated into the samples, only the

molecular ions 479 and 483 should be monitored simultaneously, for equal time intervals.

The exact amount of LTB4 present in the lavage and plasma samples was not
determined. The 1802 labelled LTB4 had not been prepared. The ﬁrst time it was
prepared, the LTB4 used was the same as the LTB4 in the spiked samples. The
chromatographic proﬁle showed that the labelled LTB4 was not suitable to be used as an
internal standard. There was more than one compound present at the known retention
time of LTB4. The presence of more than one peak could have been caused by the
degradation of the LTB4 during the labelling process or the LTB4 could have already
been degraded when it was received. The labelling of LTB4 requires at least 24 hours for
90% 180 exchange. This reaction had proceeded for 52 hours.

Labelled LTB4 is made in the following manner: H2180 (Icon Services Inc.,
Summit, NJ) and LTB4 (gift- J. Rokach) are stirred at 37°C for at least 24 hours with
pseudocholinesterase XI as the catalyst (18). Both carboxylic oxygens are exchanged
with the labelled oxygen in water. After 24 hours the reaction is acidiﬁed with HCl in
H2180. It is extracted with ethyl acetate. The % of 18O incorporation is determined by
monitoring the molecular ions of the labelled and unlabelled compounds. This reaction
should increase the molecular weight of the LTB4 by four. The reaction was conducted
again using a different source of LTB4 and was allowed to react for only 24 hours instead
of 52 hours. The incorporation was successful. The selected ion proﬁles for the PFB
ester, TMS ether derivatized LTB4 are shown in Figure 5-12. As soon as adequate 1802-
labelled LTB4 is prepared, samples will be spiked before derivatization and the recovery

of the LTB4 in plasma and lavage determined.

 

 

140

 

 

 

 

 

 

 

 

 

 

"“3 A- - 115 .z -.1.L..- .RT
5 7.4
l
l
E m/z: 483
V
e
I 3.0
ﬂ
3' m/z: 481
n
I
‘ AQ‘
t. w.
u 8.4
[Tl/Z: 479
3' 1 T ‘31'8811 ' '3288' W V V '3388' ' ' ' '34‘88' ' ﬂ Vas'eaj . YBSiBBscm
Figure 5-12. The SIM proﬁles of 18O-labelled LTB4. The 24-hour reaction

incorporated 70% of the total LTB4 with two 180 atoms (tn/2: 483)
and 29% of the total LTB4 with one 180 atom (tn/z= 481).

 

 

141
Project 111- The Birch Reduction

In order to attempt this project much research went into determining just how to
physically accomplish the organic reaction. Gratitude is expressed to Dr. William
Reusch and his graduate student, Wuyi Wang for their assistance. The physical set-up
for the reaction is shown in Figure 5-13. The apparatus was assembled in a fume hood.
The reaction was performed in an argon atmosphere; thus argon must be available in the
fume hood. The glassware used must be very dry. It is also preferable to do this reaction

on a day when the humidity was low.

Glassware with ground glass joints was required. A 25 ml three-necked round
bottom ﬂask was used for the preparation of Li/CH3NH2 solution. All glassware was
stored in a 150°C oven until use. The three-necked ﬂask was ﬁtted with a condenser.
The condenser and three-necked ﬂask are removed from the oven and allowed to cool
slightly. All joints are greased and stoppers are placed in the open holes. The greasing
was critical because it will prevent moisture from entering the ﬂask when the condenser

is ﬁlled with dry ice.

The system is purged with argon. The condenser and Dewar are then ﬁlled with
dry ice and isopropanol (—22°C). A squeeze bottle allows for the controlled addition of
isopropanol. It is also important that the dry ice be added to the condenser and Dewar
before the glass reaches room temperature (i.e. while the glass is still hot). This prevents

moisture from adhering to the glass surface that will condense when the dry ice is added.

The three-necked ﬂask where the Li/CH3NH2 reagent is produced is cooled in the
Dewar and methylamine (Aldrich) is added through a ground glass jointed ﬂow
controller. The argon is turned off momentarily to allow methylamine to bubble through
the parafﬁn. After a few minutes methylamine should start to condense in the ﬂask and

reﬂuxing will occur. The methylamine is collected until there are approximately 3 ml of

 

 

 

 

 

.5838 :83 8: E wow: 2:83am 8: 88 938 323% of. .m Tm 8:me
4.02. no. x00. no.
6:03.38.
05.5.2238 \oo. bu
xon 8:» xon .5»

 

 

 

 

 

 

 

,

142

_ocoaohomrxog bu

 

 

 

I}

 

143
methylamine for each sample. Then the methylamine is turned off but at the same time
the argon ﬂow is increased to assure that no air enters the system. The methylamine tank

is turned off and the ﬂow controller is closed.

Lithium metal (Aldrich) is cut and placed in the reﬂuxing liquid methylamine.
The procedure for cutting the lithium is as follows: using scissors cut a piece of lithium,
while it is covered with mineral oil. The metal should not be handled with bare hands.
(Since the sulﬁdopeptide leukotrienes are of very small concentration, approximately 5
mm of wire is cut; this will usually correspond to a molar excess.) The piece of Li metal
that has been out should be shiny and not dull (oxidized). Using tweezers, the metal is
placed in hexane to remove the mineral oil. It is then quickly transferred to the liquid

methylamine.

After the lithium dissolves in the liquid methylamine, the solution in the round
bottom ﬂask will turn blue. Failure to achieve a blue color may result from two causes:
a) there is water present in the closed system, or b) the liquid methylamine is too cold and
the metal has not dissolved. Case b is the more likely. If the metal does not dissolve
then the dewar surrounding the round-bottom is removed to let the system warm slightly.
Once the blue solution appears and remains for at least ten minutes it is ready to be

transferred to the sample ﬂasks.

The sample fractions are received in the mobile phase (MeOH/HZO/I'FA) in
which they eluted. They are transferred to ground glass jointed ﬂasks. Sample vessels
require at least two necks and should be 10 ml in volume. The sample vessels should be
silanized. The sample vials are rinsed twice with water to assure complete transfer of the
sample. Methanol and triﬂuoroacetic acid (TFA) are removed ﬁrst by purging with
nitrogen. The sample is then lyophilized (freeze-dried) to remove the residual water.
The Birch reaction should be completed soon after the lyophilization due to the stability

of the leukotrienes. The internal standard is added at this time. The proper internal

144
standard is ISO-labelled S-HETE. The ﬂask should contain a stir bar and the joints

should be stoppered with the proper size serum caps.

Before the transfer of the Li/CH3NH2 reagent, the sample vessel is purged with
argon and placed in a dry ice/isopropanol bath. The lithium solution is transferred
through a piece of metal tubing. The two-necked sample ﬂask will have an outlet that is
at atmospheric pressure and in the other neck will hold the transfer line. The solution
passes through the tubing because of pressure which builds up in the system. In order for
this to occur, the argon pressure is increased at the tank and the stoppers on the solution
vessel are held tightly. The outlet hole for the Ar on the parafﬁn ﬂask is plugged. About
2-3 ml of the lithium solution are added per sample ﬂask. The measurement is very
crude due to the method of transfer. The sample vessels are left to react for ten minutes
and the blue color may disappear. The reaction is quenched with a volume of methanol
that is 10% of the Li/methylamine volume (0.2-0.3 ml). When the solution is colorless
the reaction has been quenched. The sample vessels are allowed to warm to room
temperature under the hood. The septa are removed to allow the escape of methylamine

gas.

Once the methylamine has evaporated, which may be expedited with N2 purging,
1 ml of water is added to each ﬂask and the samples are treated with HCl to adjust the
pH to 3. The analyte is extracted three times with hexane/ethyl acetate (1:1). The
extracts are pooled and transferred to a silanized one dram vial. The solvent is
evaporated. The residue is then redissolved in 0.2 ml of methanol. A small amount (tip
of a spatula) of 5% Rh (Aldrich) on alumina is added. The vial is placed on ice.
Hydrogen gas is bubbled into the methanol while it is constantly stirred for ﬁfteen

minutes. The hydrogen must be bubbled into the methanol. If it is allowed to just blow

 

on the surface of the methanol, the methanol will evaporate quickly and a small ﬁre will

commence at the top of the vial. After ﬁfteen minutes the hydrogen is stopped and the

145
catalyst is removed by centrifugation. Small silanized 6 X 50 mm test tubes containing
the solution are placed in a microcentrifuge for ten seconds. The methanol can then be
separated easily from the catalyst. The vial and catalyst are washed three times with
methanol and the methanol washings are pooled. The methanol is removed with

nitrogen.

At this stage the analyte has been converted to S-hydroxyeicosanoic acid
(MW=328). The compound is derivatized with both pentaﬂuorobenzyl (PFB) bromide to
form the PFB esters and BSTFA to prepare the 5-OTMS ethers. The derivatized species
(MW=580) can then be analyzed by GC-MS following electron capture under methane

chemical ionization conditions. The carboxylate anion fragment (m/z=399) is monitored.

The outcome of these reactions was quite positive. The success of ﬁnding the
desired compound, which was present in ng to 11g quantities, after such an extensive
procedure was encouraging. The samples had been conﬁrmed to have peptido—LTs by
RIA, but the collaborating scientists wanted mass spectral conﬁrmation. Unfortunately,
GC-ECNI-MS cannot conﬁrm that peptido-LTs are present in the islet cell wash samples
that were sent. Since the S/N ratio was only equal to three, it cannot be stated that the

islet cells produce peptido-LTs when challenged. A S/N ratio of at least ﬁve is required.

In conclusion, it is apparent from these last two discussions that much more work
will be required to complete the two projects. They will be continued by other students
in our laboratory. If the procedures are followed exactly, with forethought and careful
planning, there should be no problem in passing these projects over to other people. It is
hOped that successful experiments with deﬁnite positive answers will be completed

within a year’s time.

99°39)

10.
ll.
12.
l3.

14.
15.

l6.

17.

18.

146
REFERENCES CHAPTER V

Holtzman, M. J ., Aizawa, H., Nadel, J. A.'and Goetzl, E. J ., Biochem. Biophys.
Res. Commun. 1983, 114, 1071—1076.

Holtzman, M. J., Fabbri, L. M., O’ Byme, P. M., Gold, B. D., Aizawa, H.,
Walters, E. H., Alpert, S. E. and Nadel, J. A., Am. Rev. Respir. Dis. 1983, 127,
686-690.

O’Byrne, P. M., Holtzman, M. J., Fabbri, L. M., Gold, B. D., Aizawa, H.,
Walters, E. H., Alpert, S. E. and Nadel, J. A., Am. Rev. Respir. Dis. 1984, 130,
214-219.

Scheel, L. D., Dobrogorski, O. J ., Mountain, J. T., Svirbely, J. C. and Stokinger,
H. B., J. Appl. Physiol. 1959, 14, 67-80.

Golden, J. A., Nadel, J. A. and Boushey, H. A., Am. Rev. Respir. Dis. 1978,
118, 287-294.

Menzel, D. B., J. Toxicol. Environ. Health 1984, 13, 183-204.

Balazy, M. and Murphy, R.C., Anal. Chem. 1986, 58, 1098-1101.

Birch, A.J. and Smith, H., Quart. Rev. (London), 1958, 12, 17-33.
Kaiser, E. Synthesis, 1972, 391-415.

Schon, 1., Chem. Rev., 1984, 84, 287-297.

Truce, W., Tate, D. and Burdge, D., J. Chem. Soc., 1960, 82, 2872-2876.
Truce, W. and Breiter, J., J. Chem. Soc., 1962, 84, 1621-1624.

Metz, S. A., Hall, M. B., Harper, T. W. and Murphy, R. C., J. Chromatogr.
1982, 233, 193-201.

Jubiz, W. , Nolan, G. and Kaltenborn, K. C., J. Liq. Chrom. 1985, 8, 1519-1526.

Westcott, J. Y., Clay, K. L. and Murphy, R. C., J. Allergy Clin. Immunol. 1984,
74, 363-368.

Powell, W. S. in Biochemistry of Arachidonic Acid Metabolism, 1985 , W. E. M.
Lands ed. Martinus Nijhoff Publishing: Boston, 375—403.

Westcott, J. Y., Stenmark, K. R. and Murphy, R. C., Prostaglandins, 1986, 31,
227-237.

Westcott, J. Y., Clay, K. L. and Murphy, R. C., Biomed. Mass Spec. 1985, 12,
312-317.