A ..

m... .m,

#5.. r. L.

mafia,
}-

A L}
AM. ~

.1
E

.d .

LRva‘

.a

.4.»qu
w .
mi.

3%.

..
7.

. €329 .

 

 

1:41,
I. .2
.

z ..

a.“
‘1

2.: =
.595?

L van».

 

.1"...

. a

”5.2L:

a»! u. 1
$ 4.

|§I
)

 

fun“

THESlS

r
J

(y ‘\ '1‘ 1‘- I"\\ )

 

LIBRARY mmmlillfillilll’fl‘illlluum
MIChigan State 3 1293 01766 2481
Unlverslty

 

 

 

This is to certify that the

dissertation entitled

SUBCELLULAR LOCALIZATION AND

MEMBRANE ASSOCIATION OF THE

PROSTAGLANDIN ENDOPEROXIDE H
SYNTHASE ISOZYMES

presented by
ANDREW GIBSON SPENCER

has been accepted towards fulfillment
of the requirements for

Ph . D . degree in Biochemistry

 

MM}; /.‘ fix;

Major professor

Date September 10, 1998

MS U is an Affirmative Action/Equal Opportunity Institution 042771

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

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1190 W14

SUBCELLULAR LOCALIZATION AND MEMBRANE ASSOCIATION OF THE
PROSTAGLANDIN ENDOPEROXIDE H SYNTHASE ISOZYMES

By

Andrew Gibson Spencer

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

DOCTOR OF PHILOSOPHY

Department of Biochemistry

1998

ABSTRACT

SUBCELLULAR LOCALIZATION AND MEMBRANE ASSOCIATION OF THE
PROSTAGLANDIN ENDPEROXIDE H SYNTHASE ISOZYMES

By

Andrew Gibson Spencer

Using immunoelectron microscopy and photoaffinity labeling we have determined
the subcellular locations and membrane association regions, respectively, of prostaglandin
endoperoxide H synthase-l and -2 (PGHS-l and PGHS-Z). PGHS-l and -2 are the targets
of aspirin and are integral membrane proteins of the endoplasmic reticulum and nuclear
envelope. Previous experiments in our laboratory led to the hypothesis that these enzymes
may reside, at least in part, in different subcellular compartments and that their
compartmentation may affect their access to arachidonic acid and serve to separate the
functions of the enzymes. High-resolution immunolocalization studies demonstrated that
PGHS-l and -2 were found on the lumenal surfaces of the ER and nuclear envelope in
murine NIH 3T3 cells, human monocytes, and human umbilical vein endothelial cells.
Within the nuclear envelope, PGHS-l and -2 were present on both the inner and outer
nuclear membranes and in similar proportions. Western blotting data showed a similar
distribution of PGHS-l and —2 in subcellular fractions. Thus, we are unable to attribute the
independent functioning of PGHS-l and -2 to differences in their subcellular locations. A
further conclusion of importance from a cell biological perspective is that membrane proteins
of the ER such as PGHS-l and PGHS-Z, which are located on the lumenal surface of the ER,

are able to diffuse freely among the ER and the inner and outer nuclear membranes of the

nuclear envelope.

PGHS-l and -2 are integral membrane proteins but do not possess transmembrane
domains. Instead, X-ray crystallographic studies led to the hypothesis that these proteins
associate with membranes through a novel, monotopic membrane binding domain (Picot, D.,
et a1. (1994) Nature, 367:243-249). Early studies from our laboratory demonstrated that the
membrane binding domain of ovine PGHS-l resides between amino acids 25-166 (Otto, IQ
and Smith, W.L. (1996) J. Biol. Chem, 271:9906-9910). Using the hydrophobic,
photoactivatible reagent 3-trifluoro-3-(m-['251]iodophenyl)diazirine ([‘”I]-TID) to label the
membrane associated domains of PGHSs, we have extended this work and provided
biochemical evidence for the membrane association domains of PGHS-l and PGHS-Z. The
PGHS-l and PGHS-Z membrane binding domains reside within residues 74-140 and 59-125,

respectivley.

For my family...

Dad for paving the way and everything else.
Mom for giving love and life.
Molly, a true kindred spirit.

Matt, my idol and friend.

I miss my lovely mother, and I
1 love my busy father (sic).
I know I owe my brother one thing or another.
I hear my sister singing. . .

James Taylor, Terra Nova

iv

ACKNOWLEDGMENTS

First of all a big thanks to Bill Smith for teaching me the persistence required to do
good science. He was patient with me when I was ready to quit early on and was supportive
of my repeated trips to New Jersey to do experiments. Dr. Smith demonstrates by example
every day what it takes to be successful in this business. There is probably not another lab
in the country that could have provided a better environment for me personally as a graduate
student, and I truly believe that. So thanks a lot to Dr. Smith and all current and former
Smith/DeWitt lab members for making it great.

Special thanks to Dave DeWitt for being a second mentor to me. Dave is always
great to talk to about life in general, experimental problems, computers, food, wine, beer, etc.
Thanks also for keeping me going when the experiments in New Jersey weren’t cooperating.

Another special thanks goes to John Woods at Merck & Co. In John’s lab I first
learned how to get things to work. I also became more streetwise to the ways of science and
Tri-State trains. My experience in John’s lab was very positive and I thank him for
everything. Other people at Merck whom I owe big are Jeanne for being like a big sister and
Marty Springer for stories, wisdom, and the “best pizza in Brooklyn” at Patsy’s.

Okay, find yourself: Katie (the best), the Millers (home and great meals), Stebbins
and KQ (FOOD, music, everything else), Otto (wisdom, ski trips), Mike and Tim (Halloween
and declared holidays), Shelby, Pete, Canfield, Mindy, Tucker, Piper, Strait, Luke, Voskuil,
G, Stain, Leland, Hough, Kirchen, Jane, Powell, Uncle Jon. Thanks to all!

Finally a word to the memory of those I lost in 1997-98: My grandfather Paul
Johnson, friend and stellar human Scott Bishop, and former Hornet Ty Barber. You are

missed.

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................................ xiii

LIST OF FIGURES ............................................................................................................ ix

LIST OF ABBREVLATIONS ............................................................................................. xi

CHAPTER 1

LITERATURE REVIEW
Introduction .............................................................................................................. 1
Historical perspective .................................................................................................... 7
Physical characteristics of PGHS-l and PGHS-2 ....................................................... 14
Regulation of PGHS—1 and PGHS-2 expression ........................................................ 19
Aberrant regulation of PGHS-2 and implications for cancer ...................................... 22
Mobilization of arachidonic acid and its availability to PGHS-l and PGHS-2 .......... 24
Prostaglandin signaling via G-protein coupled receptors ............................................ 30
Evidence for nuclear prostaglandin signaling systems .............................................. 32
The nuclear envelope: structure and membrane protein trafficking ............................ 35
PGHS catalysis and inhibition by NSAIDS ................................................................. 37

CHAPTER 2

SUBCELLULAR LOCALIZATION OF PROSTAGLANDIN ENDOPEROXIDE H
SYNTHASE-1 AND -2 BY IMMUNOELECTRON MICRSOCOPY

Introduction ................................................................................................................. 43
Methods ....................................................................................................................... 45
Results ......................................................................................................................... 52
Discussion ................................................................................................................... 62
Acknowledgement ....................................................................................................... 68

vi

CHAPTER 3
MEMBRANE ASSOCIATION OF PROSTAGLANDIN
ENDOPEROXIDE H SYNTHASE-1 AND -2

Introduction ................................................................................................................ 69
Methods ....................................................................................................................... 72
Results ......................................................................................................................... 78
Discussion .................................................................................................................... 89
APPENDIX .......................................................................................................................... 92
BIBLIOGRAPHY .................................................................................................................... 96

vii

LIST OF TABLES

Table 1. Analysis of PGHS-1 and PGHS-2 distribution within the
nuclear envelope ........................................................................................ 55

viii

Figure 1.
Figure 2.
Figure 3.
Figure 4.
Figure 5.
Figure 6.

Figure 7.

Figure 8.

Figure 9.

Figure 10.
Figure l I.

Figure 12.

Figure 13.

Figure 14.

Figure 15.

LIST OF FIGURES

The eicosanoids ............................................................................................ 4
Enzymatic activities of PGHS-l and PGHS-2 ............................................. 5
Model of two independent prostaglandin biosynthetic systems ................... 6
Structures of some common N SAIDS ....................................................... 13
Comparison of the PGHS-1 and PGHS-2 proteins .................................... 17
X-ray crystal structure of ovine PGHS-1 ................................................... 18

Schematic diagram of the biphasic prostaglandin synthesis in
activated inflammatory cells ...................................................................... 29

The continuous membrane systems of the ER

and nuclear envelope .................................................................................. 38
Peroxidase and cyclooxygenase catalysis .................................................. 41
Immunogold labeling of PGHS—2 in human monocytes ............................ 53
Immunogold labeling of PGHS-1 in murine NIH 3T3 cells ...................... 57

Western blot analysis of PGHS-l and PGHS-2 in
subcellular fractions of murine NIH 3T3 cells ........................................... 61

Diffusion of integral membrane proteins
in the nuclear envelope .............................................................................. 67

Mechanism of [‘ZSI]-TID labeling of membrane proteins .......................... 71

Predicted proteolysis products of oPGHS-l
with Lys C and Glu C ................................................................................ 79

ix

Figure 16.
Figure 17.

Figure 18.

Figure 19.

Figure 20.

Figure 21.

Figure 22.

[”SIJ-TID labeling of ovine seminal vesicle membrane proteins ............... 8O

[‘251]-TID labeling of an N-terminal peptide of oPGHS-l ......................... 81
A 7.4 kDa peptide of oPGHS-l containing the predicted

membrane binding domain is labeled by ['ZSH-TID .................................. 82
Primary structure of oPGHS-l ................................................................... 83
Predicted major proteolysis product of hPGHS-2

with Glu C .................................................................................................. 84
An 8 kDa peptide of hPGHS-2 containing the predicted

membrane binding domain is labeled by ['ZSH-TID ................................... 86
Primary structure of hPGHS-2 ................................................................... 88

15d-PGJ2
5-LO
cPLA2
EFA
EGF

ER
FLAP
Glu C
IL-lB
HET Es
HUVEC
LPS

Lys C
MBD
NE

NPC
NSAIDS
PBMCs
PG
PGHS
PPAR
PVDF
SDS-PAGE
sPLA2
SREBPI
Tx
[‘251]-TID

ABBREVIATIONS

15-deoxy-A‘2’M-Prostaglandin J2
5-lipoxygenase

cytosolic phospholipase A2

essential fatty acid

epidermal growth factor

endoplasmic reticulum

5-lipoxygenase activating protein
endoproteinase Glu C

interleukin-l, [3 isoform
hydroxyeicosatetraenoic acids

human umbilical vein endothelial cells
bacterial lipopolysaccharide
endoproteinase Lys C

membrane binding domain

nuclear envelope

nuclear pore complex

non-steroidal anti-inflammatory drugs
peripheral blood mononuclear cells
prostaglandin

prostaglandin endoperoxide H synthase
peroxisome proliferator activated receptors
polyvinylidene fluoride

sodium dodecyl sulfate-polyacrylamide gel electrophoresis
secretory phospholipase A2

sterol regulatory element binding protein 1
thromboxane
3-trifluoro-3-(m-[‘25I]iodopheny1)diazirine

xi

CHAPTER I

LITERATURE REVIEW

Introduction

The widespread use of aspirin, ibuprofen, and other non-steroidal anti-inflammatory
drugs (N SAIDS) has called for a more thorough understanding of the physiological systems
in which they work. NSAIDS exert their anti-inflammatory and analgesic effects through
the inhibition of prostaglandin endoperoxide H synthase-1 and -2 (PGHS-1 and PGHS-2),
two biochemically similar enzymes that catalyze the committed step in the formation of
prostaglandins from arachidonic acid [1]. Prostaglandins, along with leukotrienes, belong
to a family of oxygenated, 20-carbon lipid signaling molecules called eicosanoids (Fig. 1).
These molecules act as local (autocrine or paracrine) hormones to regulate a broad range of
physiological processes including vascular smooth muscle dilation, platelet aggregation,
renal water reabsorption, parturition, and tumor formation. The prostaglandin biosynthetic
activity of PGHS-1 was first isolated over 20 years ago, and the enzyme has been
thoroughly characterized biochemically [2, 3]. In the early 1990's, the discovery of PGHS-2
as a v-src and tumor promoter inducible immediate early gene precipitated renewed vigor
in prostaglandin research and an attempt to determine the physiological need for two
PGHSs [4]. Research into the biology of these two enzymes has increased in our

understanding of how aspirin works and how prostaglandins act throughout the body to

2

Both PGHSs are integral membrane proteins of the endoplasmic reticulum (ER) and
nuclear envelope (NE). Their product, PGHZ, which results from the addition of two moles
of molecular oxygen to arachidonic acid through the sequential cyclooxygenase and
peroxidase activities of PGHS-1 and -2, is the precursor to all prostaglandins (Fig. 2). The
biologically active prostaglandins (PGEZ, PGan, PGDZ, PGIZ, and thromboxane) are then
synthesized from PGH2 by specific downstream synthases [2]. After their synthesis,
prostaglandins are exported from the cell by one or more plasma membrane transporters
where they become available to specific G-protein linked receptors. It is these receptors
through which prostaglandins exert their effects by triggering second-messenger responses
within cells.

Our laboratory uses chemical, cell biological, and biochemical approaches to gather
basic information on the signaling mechanisms of prostaglandins. Structural and biophysical
studies of PGHS-1 and -2 have led to important advances in our understanding of aspirin
action and prostaglandin biology. An important realization has been that despite their
similar biochemical properties, , PGHS-1 and PGHS-2 mediated prostaglandin synthesis can
lead to different physiological outcomes (Fig. 3)[1]. One line of evidence for the “two
systems” model stems from the vastly different expression patterns of the enzymes: PGHS-1
is expressed constitutively in nearly all mammalian cells and is likely an important
housekeeping gene. PGHS-2 is normally absent from cells but can be induced by grth
factors, inflammatory cytokines, and bacterial endotoxin. Furthermore, even when both
enzymes are present in the same cells simultaneously they seem to be differentially available
to their substrate, arachidonic acid [Reddy, 1997 #342]. Recently created knockout mice

deficient in PGHS-1 and PGHS-2 have markedly different phenotypes, clearly

3

demonstrating the different physiological roles of the enzymes [Morham, 1995 #341;
Langenbach, 1995 #339].

My dissertation research has focused on the subcellular localization and membrane
association of PGHS-1 and PGHS-2. Using high-resolution immunocytochemical techniques
I have described the subcellular localization of the two enzymes and provided information
important to the determination of the mechanisms by which cells separate their activities.
Further, my work on the membrane association of PGHS-1 and -2 has provided
biochemical evidence for a novel membrane binding domain that traverses a single leaflet

of the lipid bilayer.

  
 

 

__ __ / COO'
____/\______/\/\/
Arachidonic acid
PGHS-ll-Z 5-LO, FLAP Cytochrome P450s
v
Prostaglandins Leukotriencs Hydroxy acids
0 _ coo: 0H _ H COO. _ __ coo:
/ —— =/\/\/\OH
no on
PGE, LTB. HETEs

Figure 1. The eicosanoids. Arachidonic acid (20:4) is a component of cellular membranes
and is obtained through the diet or via biosynthesis from linoleic acid (18:2). Upon its
release from phospholipid stores, this essential fatty acid can be oxygenated and thereby
converted to one of several hormones of the eicosanoid family. Members of this family of
20-carbon, lipid-derived hormones include the prostaglandins (e. g., PGEZ), the leukotrienes
(e.g., LTB4), and the hydroxy acids (e.g., HETEs).

5

 

 

 

PHOSPHOLIPID
PHOSPHOLIPASE A2 ‘
ARACHIDONIC ACID
CYCLOOXYGENASE
PGH
o ' svmnases
\
PEROXIDASE
o— “h
\
PGH2 __
SYNTHASES /‘ OH\
PROSTAG LANDINS
PROSTACYCLIN THROMBOXANE A2

Figure 2. Enzymatic activities of PGHS-1 and PGHS-2. PGHS-1 and PGHS-2 each
catalyze the conversion of arachidonic acid to PGH2 by consecutive cyclooxygenase and
peroxidase activities. The cyclooxygenase activity, which is inhibited by aspirin, adds two
moles of molecular oxygen per mole of arachidonate to give PGGZ. The peroxidase activity
then converts the 15-hydroperoxy group to the corresponding alcohol to yield PGH2 [1].

 

 

lngS-IIPGHS-Z Gd
mm

    
 
   
   

t.

   

Nuclear
pGD PGHS-l/PGHS—l (?)
systau
Nucleus /

.. _’—‘

12@

§ 1......

Figure 3. Model of two independent prostaglandin biosynthetic systems. PGHS-1 and
-2 are not functionally redundant and may synthesize prostaglandins that subserve different
cellular functions. Both PGHS-1 and -2 are known to synthesize prostaglandins that exit the
cell and act hormonally via G-protein linked receptors. Due in part to the correlation of
PGHS-2 expression with nuclear events like cellular differentiation and mitogenesis, some
have hypothesized that a nuclear prostaglandin signaling system exists. Some studies
suggest that a prostaglandin may act to modulate transcription by acting as a ligand for the
PPAR family of orphan nuclear receptors. AA = arachidonic acid; PGs = prostaglandins;
R~ plasma membrane prostaglandin receptor; G = heterotrimeric G protein; E = effector
protein; PPAR = peroxisome proliferator activated receptor; PGR? = putative unknown
nuclear prostaglandin receptor.

  

7
Historical perspective

Among the landmark studies in vertebrate physiology were those by George and
Mildred Burr in 1929 and 1930 demonstrating the dietary essentiality of fatty acids [5, 6].
At the time of their studies, there was general uncertainty regarding the necessity for fats in
the diet. During the late 1920's, the Burrs transported a cadre of Long-Evans rats in a Model
T Ford from the hills of Berkeley, California to their new laboratory at the University of
Minnesota where George was appointed as a plant pathologist in the Department of Botany.
When he consulted with the department head — whom he feared might question the need for
a plant pathologist to have an animal colony in the Medical school — he was told, “I do not
care what field you work in, just do good work in some field.” Accordingly, the Burrs began
to rigidly exclude fat fi'om the diets of rats over periods of several months, feeding them only
protein, carbohydrate, and vitamins A and D. The first obvious effect of dietary fat exclusion
was the development of scaly skin, dandruff, and necrosis of the tail. Burr observed that “the
early outward signs of an unhealthy condition of the animal are soon followed by a cessation
of growth when the animal is about 25% underweight in comparison with the controls
receiving fat” [5]. After several months, the rats typically began to lose fur around the face
and head and developed sores of the skin. Blood was normally observed in the urine of fat
deprived rats. Without exception, rats kept on a fat fiee diet lost large amounts of weight and
died. Autopsies on the dead animals revealed that “the most marked and uniform pathology
(was) observed in the urinary tract and the kidney. There seems little doubt that this is an
important factor in the death of the animal”[5]. Further study of the fat deprived rats
demonstrated a striking lack of fertility when compared to control animals. Interestingly,

both the kidney pathology and reproductive problems would be experienced 60 years later

8

by mice lacking functional PGHS or the appropriate prostaglandin receptors [7, 8, 167].

All symptoms of the fatty acid deficiency disease could be rescued by very small
doses of lard, corn oil, linseed oil, or several other “remedies” containing unsaturated fatty
acids. Each of these remedies were later found to contain small amounts of arachidonic acid
and large amounts of linoleic acid, the metabolic precursor to arachidonate. When the Burrs
showed that purified linoleic acid was just as curative of the fat deficiency disease symptoms
as the lard or oils, they concluded that “linoleic acid (and possibly other acids) therefore is
an essential fatty acid” [6]. Today we fully appreciate the essentiality of plant-derived
unsaturated fatty acids in the mammalian diet. These fatty acids and their derivatives are not
only important for cellular membrane structure, they serve as precursors to the
prostaglandins and other important signaling molecules throughout the body. The kidney
failure of Burr’s rats was undoubtedly due in part to the lack of substrate for prostaglandin
biosynthesis; endogenous PGE2 production in fatty acid deficient animals is severely reduced
[9].

Elsewhere in the early 1930's, while the Burrs contemplated the significance of their
work, Ulfvon Euler began his postgraduate work in London. He focused on studying what
now may seem the inordinately general topic of “naturally occurring, biologically active
compounds” [10]. Later, von Euler would be awarded the Nobel Prize for his work on
neurotransmitters and the storage and release of norepinephrine. But in the early 1930's, he
was busy making extracts from various animal organs and testing their effect on rabbit blood
pressure. After many fruitless excursions, von Euler injected small amounts of human semen
and saw a dramatic drop in blood pressure [11]. Similar results were observed in extracts

from sheep seminal vesicles and using this tissue von Euler continued on this track,

9

eventually demonstrating that the active component was soluble in lipid solvents. Regarding
this observation, von Euler wrote that the finding of a lipid-soluble signaling factor “was at
first rather surprising and actually became the turning point of the work, and as much as one
can speak of a discovery this finding was clearly the most important one, since it introduced
an entirely new kind of pharmacodynamically active substance occurring naturally in the
body” [10]. Prior to von Euler’s work two American physicians at Columbia University,
Kurzrok and Lieb, were trying out artificial insemination in women. In some cases, injection
of the semen into the uterus would cause a violent contraction of the uterine muscle which
could be reproduced in vitro [12]. Although this was the first demonstrated effect of what
would later become known as prostaglandins, Kurzrok and Lieb attributed this contractile
activity to acetylcholine and left the door open for von Euler to correctly describe the nature
of the active compounds.
Despite having passed through the same Swedish laboratory in 1933, Burr and von
Euler did not meet and remained oblivious to the strong connections between their lines of
investigation in pre—World War 11 Europe. The subsequent years would forever link these
two areas of basic research and describe a hormonal signaling system that conceivably effects
every cell in a mammalian organism. As George Burr wrote of von Euler in 1982, “If we had
met (in the 19305) it is likely that we would have discussed our then current work on
essential fatty acids and prostaglandins without noting any significant relationship between
the two. Now, 50 years later, these compounds are being treated as a unit for the solution of
their biological functions”[13].
Due in large part to World War II, progress in the years following the landmark

studies of Burr and von Euler was slow. Both pioneers pursued other avenues of research

10

more vigorously and a group of scientists at the Karolinska Institute in Stockholm began to
gather steam in their quest to isolate and describe the first prostaglandin. By the early 1950's
the essentiality of dietary unsaturated fatty acids was becoming well understood. Perhaps
the greatest advances in the 1950's were chemical and technical in nature. Several of the
essential fatty acids including arachidonate, are present in very small quantities, and not until
A.T. James and others pioneered the technique of gas-liquid chromatography was it possible
for the metabolic pathways traveled by EFAs after ingestion to be mapped [14]. Early
radioisotope incorporations were also useful in these studies.

It was more than 20 years after von Euler coined the term “prostaglandin” (a
misnomer, the compounds do not originate in the prostate), that Sjovall and Bergstrom at
the Karolinska Institute isolated the first prostaglandins in crystalline form [15]. This study
predicted the structures of PGE1 and PGFm and preceded the many years of work required
to determine the structures of the vast number of arachidonic acid-derived hormones called
eicosanoids. The 1960's also bore witness to the belated bridge between the early works of
Burr and von Euler as Van Dorp and Bergstrom independently showed that these compounds
called prostaglandins were synthesized enzymatically from the nutritionally essential
arachidonic acid [16, 17].

Research on the biosynthesis of prostaglandins truly took 03 in the early 1970's. The
number of publications related to prostaglandin biology began to skyrocket [13]. The
enzyme implicated in its biosynthesis, prostaglandin endoperoxide H synthase, was first
purified to homogeneity from bovine and sheep seminal vesicles in 1976 [3, 18, 19]. Five
years earlier, John Vane in London had demonstrated that humankind’s most common

remedy, aspirin, exerted its action through the inhibition of PGHS [20, 21]. This finding, for

11

which Vane was later awarded a Nobel Prize along with Bergstrom and Samuelsson,
eventually led to an adequate explanation of how drugs such as aspirin and ibuprofen work
to reduce inflammation, pain, and fever.

Vane’s discovery had a dramatic impact on the medical community. After all,
aspirin-like drugs had been in use since the time of Hippocrates 2300 years ago when the
great Greek physician had patients chew on white willow leaves to alleviate pain and fever.
These willow leaves contained salicylic acid, now a known anti-inflammatory agent.
Salicylic acid has a tendency to cause upset stomach in frequent users, and it was this side
effect that inspired Felix Hoffman, a German chemist employed by the Bayer Corporation
whose arthritic father used salicylic acid, to synthesize aspirin fi'om the original willow
compound in 1897. Aspirin is simply an acetylated form of the compound found in the
white willow tree. Even today, 100 years after Hoffman’s original synthesis, our
understanding of aspirin and the physiology associated with its action is continually
expanding.

With the 19803 came the molecular biology revolution. Genes for PGHS were cloned
[22-24], and site directed mutagenesis studies provided information on which amino acids
were important for PGHS catalysis, substrate binding, and aspirin acetylation [2, 2527].
Much of the enzymatic mechanism was hashed out during this period.
Immunocytochemistry was used to localize PGHS-1 to the endoplasmic reticulum and
nuclear envelope [28]. In the mid-1980's, Kennedy and others provided the first
comprehensive classification scheme for the diverse set of prostaglandin receptors, now
known to be members of the G-protein linked receptor family [29, 30]. Each prostaglandin

(e.g., PGEz, PGDZ) binds to one at least one type of specific receptors whose expression

12

varies according to (a) cell type and (b) the associated second messenger system to which it
is coupled.

Much of what was known about pro staglandin signaling mechanisms was challenged
in 1991 when the second known prostaglandin synthase gene was cloned. PGHS-2 was
discovered as an immediate early gene whose expression is induced by phorbol esters and
the expression of v-src in chicken fibroblasts. Others had previously suggested the existence
of two pools of PGHS [31], but proof of the existence of two genes did not come until 1991
in the labs of Harvey Herschman and Dan Simmons [32, 33]. Since then it has become
apparent that PGHS-1 and PGHS-2 are not functionally redundant. Instead, they mediate two
independent prostaglandin biosynthetic systems in mammalian organisms. Crystallographic
analysis of the structures of PGHS-1 and PGHS-2 indicate that these are globular proteins
with an associated four helix domain thought to be used for membrane anchoring [34, 35].
Although their crystal structures are virtually superimposable on a global level, important
amino acid substitutions around the active site enable the enzymes to use various substrates
to different extents and have led to the development of specific inhibitors of each enzyme
[26, 36-38].

Perhaps it shouldn’t be surprising that mice lacking either PGHS gene exhibited some
of the same pathologies as the fat-deprived rats of Burr and Burr 65 years earlier. PGHS-2
knockouts lived only a few weeks before succumbing to severe kidney pathologies and are
severely compromised in their reproduction [7]. Other knockout experiments have been
performed on genes for prostaglandin receptors. Mice lacking a functional gene encoding

the receptor for PGF20L are able to harbor babies in their womb but are unable to deliver the

13

 

11
0H 0 —C — CH 3
}
COOH COOH
Salicylic acid Aspirin
(white willow bark) (Bayer, Excedrin, Bufferin)
CH3
/ CH 3
F cm I
\COOH C<
COOH

Ibuprofen

Flurbiprofen . .
(Motrm, Advrl)

Figure 4. Structures of some common non-steroidal anti-inflammatory drugs

(N SAIDS).

14
pups (parturition) without induction [167].

The fields of prostaglandin biology and eicosanoids in general are glowing examples
of how basic research in the biological sciences can eventually lead to a general
understanding of important physiological systems with great clinical relevance. The passage
of time lends a perspective unattainable through even the most thorough review of the

literature. And as Dr. Burr wrote in 1982, “it also pans the sand and gravel from the gold”

[13].

Physical Characteristics of PGHS-1 and PGHS-2

Since PGHS-1 was first purified to homogeneity fiom seminal vesicles, the genes for
PGHS-1 and -2 have been cloned from several species including human, mouse, and rat.
Deduced amino acid sequences predict isozymes that share 60% of their primary structure
and, when conservative amino acid changes are considered, are 75% similar. The residues
required for enzymatic activity are conserved between PGHS-1 and -2 in all species. Both
enzymes contain a heme group. PGHS-1 has a molecular weight of 70 kDa, while PGHS-2
exists in both 70 and 72 kDa forms due to differences in glycosylation (see below).

Two clear differences in the primary structures of PGHS-1 and -2 are observed at the
N and C termini. Signal peptides in PGHS-1 and -2 differ in their length and share only 25%
identity in primary structure. Although these signal peptides are likely to mediate the
insertion of PGHS-1 and -2 into the ER during their synthesis, they appear not to confer

differential localization nor functional differences of the two enzymes within the ER‘.

 

lD.L. DeWitt and W.L. Smith, unpublished results.

15

Another obvious difference between the PGHSs is the 18-amino acid cassette comprising
residues 5 80-598 of PGHS-2. The function of this cassette is unknown. It does not appear
to play a role in the subcellular targeting of PGHS-2 but may be involved in targeting
PGHS-2 for degradation by cellular proteosomes.

Structures of both PGHS-l and -2 have been solved by X-ray diffraction of crystals
of purified, soluble enzyme [34, 35]. Globally, the structures are essentially superimposable
and exhibit a largely a-helical secondary structure. As predicted from early crosslinking and
ultracentrifugation studies [18, 39], the crystal structures indicate that both enzymes exist as
head-to-tail homodimers. Three major domains are clearly visible in each monomer (Fig.
6). An N-terminal epidermal growth factor (EGF)-like domain is tethered by three disulfide
bridges and may play a role in the dimerization of the monomers. The globular catalytic
domains of PGHS-1 and -2 encompass both the cyclooxygenase and peroxidase active sites
which neighbor the heme site on opposite sides.

The third domain is a novel region predicted to be the membrane binding domain
(MBD) of the PGHSs and consists of four amphipathic a helices. These helices form the
opening to the cyclooxygenase active site channel where arachidonate is assumed to bind
prior to catalysis. From the MBD protrude several hydrophobic amino acid side chains that
are likely to interact with the hydrophobic environment of the membrane bilayer [3 4].
PGHS-1 and 2 are integral membrane proteins, but the MBD is not a classical
transmembrane domain. It is predicted that the four helices traverse a single leaflet of the
membrane bilayer. Recently, the crystal structure of a bacterial squalene cyclase was solved
revealing a PGHS-like putative membrane binding domain [40]. Despite the lack of

sequence homology between the squalene cyclase and the PGHSs, it seems probable that this

l6

structural motif may be used by many integral membrane proteins.

Over 8% of the mass of the PGHSs is carbohydrate. Studies by Otto and Smith
showed that of four consensus N-glycosylation sites (Asn-XS/T), three (Asn 68, Asn 144,
and Asn 410) are glycosylated in ovine PGHS-1 [41]. The fourth, Asn-104, is not
glycosylated most likely due to its presence in the putative membrane binding domain where
it is sterically unavailable to the glycosylation machinery. The three consensus sites in
PGHS-1 are conserved in PGHS-2 and are presumed to be glycosylated in the second
isozyme based on electrophoretic mobility. In addition, PGHS-2 has two auxiliary consensus
N-glycosylation sites. One of these, Asn-580, is glycosylated in approximately 50% of
PGHS-2 molecules and results in PGHS-2 running as a doublet on SDS-PAGE. The
firnction of the PGHS glycosylations are unknown, but mutations in any of the three
conserved glycosylation sites in PGHS-1 and PGHS-2 result in less active or inactive
enzymes. In contrast, Asn-580 in PGHS-2 can be mutated without loss of activity. All
known N-glycosylations take place in the lumen of the ER. Taken together with the
mutational analysis of the N-glycosylation sites, this suggests that PGHS-1 and -2 are present
on the lumenal surface of the ER.

The possibility of a lumenal orientation was tested directly using
immunocytochemistry [42]. Using detergents to selectively permeabilize cellular
membranes, immunostaining with multiple antibodies showed that PGHS-1 and -2 are
located within the ER lumen. Despite some published evidence to the contrary , the fact that
both enzymes are N-glycosylated in cells coupled with the immunofluorescence work
provide strong evidence that both PGHSs are present on the lumenal surface of the ER and

nuclear envelope.

l7

 

PGHS-1 Nil [EGF MBDl Wide KER

59% 38% 70%

PGHS-2 N{(\[ 50121le Catalytic I Lil:

 

 

 

Figure 5. Comparison of the PGHS-1 and PGHS-2 proteins. Percentages denote the
sequence identity shared between PGHS-l and PGHS-2 in the various structural domains.
Shaded regions depict short inserts in PGHS-1 and -2 that do not appear in the other isozyme.
EGF = endothelial growth factor domain; MBD = putative membrane binding domain;
Catalytic = catalytic domain, which includes the cyclooxygenase and peroxidase active sites
along with the aspirin acetylation region.

Catalytic domains

 

Heme site

Flurbiprofen w

Putative membrane
binding domain

Figure 6. Crystal structure of ovine PGHS-1. PGHS-1 exists as a head to tail homodirner
with three major domains: the EGF domain, catalytic domain, and putative membrane
binding domain [34]. The structure of PGHS-2 is similar from this View. However, there
areseveral non-conservative amino acid substitutions near the active site that result in differ-
ential interactions with various non-steroidal anti-inflammatory drugs and fatty acid
substrates. Both PGHS-1 and PGHS-2 contain four amphipathic helices (A, B, C, and D
above) that are hypothesized to interact with a single leaflet of the lipid bilayer. The four
helices of the membrane binding domain form the opening of a hydrophobic channel that
leads directly toward the heme site where the enzymatic chemistry takes place. Modified
from Ref. 1.

19

Regulation of PGHS-1 and PGHS-2 expression

The most striking difference in the biology of PGHS-1 and PGHS-2 can be found in
their patterns of expression. PGHS-l is expressed constitutively in nearly all mammalian
tissues and forms prostanoids central to several housekeeping functions including water
reabsorption in the kidney, vascular homeostasis, and platelet aggregation via the production
of thromboxane. PGHS-2 protein, while absent from most cells, can be rapidly and
dramatically induced in many cell types upon treatment with inflammatory cytokines, growth
factors, and tumor promoters. Thus, PGHS-1 and -2 are often referred to as the constitutive
and inducible PGHSs, respectively. However, the “constitutive” and “inducible” labels for
PGHS-l and PGHS-2 do not always apply.

The promoter in the PGHS-1 gene, like many constitutively expressed
“housekeeping” genes, does not contain a TATA box, and the 5' region of the gene cannot
dramatically activate the transcription of some reporter genes in transfection assays [43].
Culturing NIH 3T3 cells express PGHS-1 at relatively constant levels under many different
stimulation conditions. In living mammals, however, the developmental regulation of
PGHS-1 is important to normal physiology. For instance, castrated male sheep will not
develop functional seminal vesicles expressing PGHS-1, the major membrane protein in this
tissue. This effect can be reversed with testosterone treatment, suggesting a role for
androgens in the developmental regulation of PGHS-1 in this tissue [44]. In addition,
newborn lambs up regulate PGHS-1 at birth to produce a bolus of PGI2 in their pulmonary
artery, presumably to aid in opening the airways to oxygen for the first time [45].

Another tissue in which PGHS-1 is up regulated developmentally is in inflammatory

cells like monocytes and macrophages. PGHS-1 is present in monocytes at very low levels.

20

As they differentiate into macrophages, PGHS-1 is significantly unregulated. A common
feature of the developmental up regulation of PGHS-1 in hematopoetic cells is its
permanence. The enzyme is unregulated once and presumably becomes a constitutive
component of the mature cell or tissue. Basal levels of PGHS-l expression may be
modulated as cells progress through differentiation pathways.

Despite consistent observations that PGHS-1 protein levels do not change
dramatically under different culture conditions, it is now clear that its expression is actively
regulated in vivo in cell and tissue-specific manners. For instance, endothelial cells up
regulate PGHS-l transcription upon treatment with PMA under the likely control of an Spl
binding site [46, 47].

In contrast to PGHS-1, the up regulation of PGHS-2 is most often rapid and transient
in nature. PGHS-2 is classified as an immediate early gene; i.e., its expression does not
require the synthesis of new protein. Although generally considered an inducible gene,
PGHS-2 is present constitutively in some tissues, most notably the brain and kidney.
Immunohistochemical staining of normal rat brain has shown discrete depots of PGHS-2
protein in various regions of the brain, including neurons [48]. Based on the regions of the
brain that exhibited PGHS-2 immunostaining, these studies suggest that PGHS-2 may play
a role in special sensory input and in the elaboration of behavioral responses. Experiments
in knockout mice have shown that the expression of PGHS-2 in the kidney is crucial to
mammalian development as PGHS-2 knockout mice die early from acute renal failure [7].

Most tissues do not express PGHS-2 until some stimulus is provided causing a rapid
and transient period of PGHS-2 expression. Since PGHS-2 expression is often associated

with the inflammatory response, several studies have examined the regulation of this

21

enzyme in inflammatory cells like neutrophils, monocytes, and mature macrophages.

Human monocytes, when treated with low doses of inflammatory cytokines ( IL-lB) or

bacterial lipopolysaccharide (LPS) begin to up regulate PGHS-2 and synthesize
prostaglandins after 12-14 hours [163-165]. This lag period is long compared to other cell
types. For example, PGHS-2 expression can be induced within 3 hours in serum-starved
NIH 3T3 fibroblasts by the addition of 10% fetal calf serum [49]. The delay in monocytes
may indicate the requirement of a secondary autocrine response in addition to the primary

IL-lB stimulation.

Up regulation of PGHS-2 expression is likely to occur through several signal
transduction pathways. Stimuli that induce the expression of the enzyme vary depending on
the cell type being examined. For example, PGHS-2 mRN A accumulates in fibroblasts upon
treatment with growth factors and tumor promoters [49]. In macrophages, inflammatory
cytokines and LPS trigger its induction. Mast cells require the aggregation of IgE receptors
at the cell surface [50]. These and other stimuli are likely to converge on the regulatory
regions of the PGHS-2 promoter via several signal transduction pathways including JAK-
STAT, p38 map kinase, and protein kinase C pathways [4]. The cyclic AMP response
element (CRE) located between nucleotides -56 and -48 upstream of the PGHS-2
transcriptional start site has been shown to play an important role in the transactivation of
the PGHS-2 gene. This was shown by transfecting a deletion series of the PGHS-2
promoter/luciferase reporter construct along with a v-src expression vector into NIH 3T3
cells [51]. Gel shift analysis of mutations made in this CRE demonstrated that it is indeed

the CRE, and not a neighboring E-box, which is responsible for v-src mediated PGHS-2

22

induction. Similar results have been obtained through the analysis of the human PGHS-2
promoter in monocytic cells [52]. Many other putative response elements exist in the PGHS-
2 promoter (e.g. AP-l, NF-kB, and SRE elements) but have not been shown to be of
functional significance.

Many patients suffering from chronic inflammation are prescribed corticosteroids if
common NSAIDS like aspirin fail to alleviate their symptoms. At least part of the anti-
inflammatory activity of corticosteroids is derived from their ability to inhibit the expression
of PGHS-2 protein. Serum- or phorbol ester-induced transcription of the PGHS-2 gene in
NIH 3T3 cells is blocked by dexamethasone [49]. Removal of the physiological source of
glucocorticoids (the adrenal gland) in rats results in increased levels of PGHS-2 (but not
PGHS-l) mRNA and protein in peritoneal macrophages [53]. Treatment of rats lacking
adrenal glands with corticosteroids suppressed this accumulation of PGHS-2, demonstrating

that adrenal steroids modulate the expression of PGHS-2 in vivo.

Aberrant Regulation of PGHS-2 and Implications for Cancer

Failure to properly regulate PGHS-2 expression may play a role in the development
of certain cancers. The early evidence for this was circumstantial. For instance, normal
colonic epithelia does not express PGHS-2, but a significant number of cancerous lesions in
the same tissue do express the enzyme. Epidemiologic data show that patients who
consume aspirin regularly are reportedly 50% less likely to die from colon cancer [54].
Recently, clues to the molecular basis for these observations have come from studying
apoptosis in colonic epithelial cell lines and from in vivo genetic studies using knockout

mice. Tsujii, et a1. observed that the overexpression of PGHS-2 in colonic epithelial cell

23

lines impairs the apoptotic response and increases cell adhesion to extracellular matrix
proteins. This suggested that the overexpression of PGHS-2 can lead to phenotypic changes
that may enhance cells’ potential for tumorigenesis [55].

The congenital human disease familial adenomatous polyposis (FAP) is the result of
mutations in the APC gene [56]. Patients with this disease develop hundreds of colonic
polyps, many of which progress to carcinoma. A mouse model of this disease was recently
used to demonstrate the role played by PGHS-2 in neoplastic transformation. Like FAP
patients, mice lacking a functional APC gene spontaneously accumulate hundreds of
precancerous polyps in their colonic epithelium. In the recent study, APC-deficient mice
were crossbred with mice lacking functional PGHS-2. The striking result was a loss of 86%
of the precancerous lesions in the APC/PGHS-2 double knockout compared to the APC
knockout [5 7]. When the APC-deficient mice were treated with specific inhibitors of PGHS-
2 a similar reduction in precancerous polyps was observed providing genetic and
pharmacological evidence that PGHS-2 activity is involved in the neoplastic transformation
of colonic epithelium.

Similar results have been observed in humans. FAP patients experience a drastic
reduction in the size and number of cancerous polyps when treated with the NSAID sulindac
[58]. It seems likely that a significant number of available NSAIDS may be broadly
effective in preventing colon cancer, although controlled human studies are lacking. In any
event, aberrant regulation of PGHS-2 has clearly been implicated in the early steps of APC-

mediated neoplastic transformation.

24

Mobilization of Arachidonic Acid and its Availability
to PGHS-l and PGHS-2

An important step in the biosynthesis of prostaglandins is the release of arachidonic
acid from phospholipid stores. Arachidonic acid is preferentially found at the sn-2 position
of glycerophospholipids and is cleaved by one of a number of phospholipases A2 (PLAzs) to
generate “fiee” arachidonate and lysophospholipids. Whether or not “fiee” arachidonate
exists within a cell is questionable. Several fatty acid binding proteins (FABPs) have been
cloned, and thus it is conceivable that there are proteins that shuttle arachidonate within the
cell after its release from phospholipid stores. The number of identified PLAzs is growing
steadily (now over ten), and all are classified according to molecular weight, amino acid
sequence, and structural homology [59, 60]. In addition to providing arachidonate to the
prostaglandin biosynthetic machinery, PLAz's also provide this substrate to the other major
arm of the arachidonic acid cascade, the leukotriene biosynthetic system. Despite their
common firnction of releasing arachidonic acid from membranes, the physiological roles of
different PLAz's appear distinct and different PLAzs may be independently coupled to PGHS-
1 or PGHS-2.

Cytosolic PLA2 (cPLAz) is an 85-kDa enzyme that when activated preferentially
hydrolyzes arachidonate from the sn-2 position of phosphoglycerides. In resting cells cPLA2
is soluble and inactive. Agonist-induced increases in intracellular calcium levels (from 50

nM to 0.3-1.0 uM) activate cPLA2 by promoting its association with membranes,

presumably the ER and nuclear envelope [61, 62]. A further level of cPLA2 regulation is
achieved through phosphorylation. Ser-505 pho sphorylation in vivo, which may be mediated

by members of the MAP kinase family [63-65], results in an increase in cPLA2 activity.

25

Expression of cPLA2 is nearly ubiquitous, but it is enriched in cells involved in the
inflammatory and allergic responses.

cPLA2 is active at the nuclear envelope and was recently shown to be expressed in
the nucleoplasm of endothelial cells under certain conditions [59, 66]. Several eicosanoid
biosynthetic enzymes, including PGHS-1, PGHS-2, 5-lipoxygenase, and 5-lipoxygenase
activating protein have a demonstrated association with the nucleus [6 7-70]. The subcellular
locations of these proteins strongly suggest that the functional components required for
eicosanoid biosynthesis can assemble on the nuclear envelope.

The other major group of PLAzs are the 14 kDa secretory PLAzs (sPLAzs). These
enzymes also require calcium but use it for the catalytic mechanism of arachidonate release
rather than for membrane association. Five mammalian 14 kDa PLAzs have been
characterized (Group IB, group IIA, group IIC, group V, and the calcium-independent group
VI) and are distinguished by their number of disulfide bridges, expression patterns, and
physiological roles [60]. Although they are not specific for arachidonic acid at the sn-2
position like the 85-kDa cPLA2 (e. g., sPLAzs will hydrolyze arachidonic and linoleic acid
from the sn-2 position with equal efficiency), the 14 kDa PLAzs hydrolyze significant
amounts of arachidonate from inflammatory cells, lung, and synovial fluid. Depending on
the cell type, the sPLAz's seem to play important roles in the release of arachidonic acid by
activated inflammatory cells (see below).

The 14kDa PLAzs must interact with the plasma membrane for activity in vivo.
Recent work by Mike Gelb’s group at the University of Washington used biophysical
techniques to show that a patch of hydrophobic residues conserved in all sPLAZ's forms a

binding surface through which the enzymes interact with membranes [71]. Interestingly, the

26

sPLAZ's apparently require a perturbation of cellular membranes for activity, as their activity
on intact membranes is low [72, 73]. Consistent with these observations, the ability of
sPLAzs to release arachidonate from the plasma membrane is highly dependent on the
activation state of the cells being studied. Exogenous addition of sPLA2 to macrophages
does not result in prostaglandin synthesis unless the cells have been pretreated with
lipopolysaccharide and platelet activating factor [74].

The simple presence of PGHS-l/—2 and a phospholipase A2 in the cell is not
sufficient in many cases to achieve prostaglandin synthesis. In fact, PGHS-1 and -2 are
often differentially available to arachidonic acid when expressed simultaneously in the same
cells. For example, arachidonate can be released into RAW 264.7 cells expressing both
PGHSs and only PGHS-l can use it to make prostaglandins [75]. In addition, cell lines from
which the gene for PGHS-1 has been deleted express high levels of PGHS-2, but cannot
synthesize prostaglandins from arachidonic acid added exogenously. Conversely, PGHS-2
knockout cell lines express PGHS-1 and can make prostaglandins fiom arachidonic acid
added to the cells [76]. The observation that not all arachidonate is indiscriminately
available to the PGHSs led to the idea that there were two cellular “pools” of arachidonate,
one for PGHS-1 and one for PGHS-2. A more likely explanation for these observations is
that distinct PLAz's are functionally coupled to either PGHS-1 or PGHS-2.

The possibility of PLA2 coupling to PGHS-1 or -2 is being studied in several
laboratories using activated inflammatory cells, which can release arachidonate and
synthesize prostaglandins in distinct early and late phases. The early phase releases
arachidonate within the cell and occurs during the first 10-20 minutes of stimulation

resulting in prostaglandin synthesis via PGHS-1 . The late phase, extracellular release of

27

arachidonate persists for several hours and is mediated by PGHS-2 [50, 77]. Attempts to
determine the PLAz's responsible for the early and late phases of arachidonate release have
revealed a complex and often contradictory picture. For example, Reddy, et al. found that
early and late phase arachidonate release was mediated by type IIA sPLA2 and cPLAz,
respectively [78]. Earlier, Bingham et a1. had observed the converse [79]. Conflicting
results such as these may be explained by differences in both the condition of the cells and
the stimulation used. Other labs have since performed similar experiments, and is most cases
it seems that cPLA2 releases arachidonate intracellularly for early phase prostaglandin
synthesis while sPLA2 mediates the extracellular, late phase burst of arachidonate release and
prostaglandin synthesis. This model is consistent with the cell biological properties of cPLA2
and the sPLAz's, which bind to intracellular and extracellular membranes, respectively, when
activated.

The early studies mentioned above indicated that the type IIA sPLA2 played a
central role in PGD2 synthesis by stimulated mast cells [77, 79]. However, recently
discovered null mutations in the type IIA sPLA2 gene were shown to have no effect on the
patterns of arachidonic acid release and PGD2 synthesis. Instead, the newly cloned type V
sPLA2 appeared to be the relevant enzyme in late phase mast cell arachidonate release [80].
A similar reassessment of the roles of the type IIA and type V sPLAzs in mouse
macrophages was performed in Ed Dennis’ lab at the University of California at San
Francisco [81, 82]. The misidentification of type IIA sPLA2 in inflammatory cell
arachidonate release was likely due to activity assays and antibodies that were unable to
distinguish between the structurally similar type IIA and type V enzymes. Nevertheless, type

IIA sPLA2 is obviously important in some cell types. For instance, rat fibroblasts that do not

28

express the type V sPLA2 display the typical biphasic prostaglandin production mediated by
cPLA2 and type IIA sPLA2 [83]. It may be that under certain conditions the type IIA and type
V PLAzs are functionally redundant [84].

There may be crosstalk between cPLA2 and type V sPLA2 that is essential to the
coordination of early and late phase prostaglandin synthesis in inflammatory cells. As
mentioned above, a membrane perturbation event is required before sPLA2 is able to release
arachidonate from cellular membranes [72, 73]. One model of this crosstalk system
implicates cPLAz-derived arachidonate in the generation of the perturbation event.

Most of the late phase PGE2 released from stimulated mouse macrophages is derived
from PGHS-2. Interestingly, one can prevent this late phase, sPLAz-mediated PGE2 release
by specific inhibition of cPLA2 activity [74]. This effect cannot be reversed by exogenously
added sPLAz. However, it can be reversed by artificially increasing the intracellular
arachidonate levels in the cell prior to activation. These studies, along with recent
observations by Kuwata, et al,. suggest that early phase cPLAz-derived arachidonic acid
generates a signal that is required for the late phase, sPLAz-mediated release of
prostaglandins (Fig. 7) [74, 83].

One rather speculative possibility to account for the independent operation of
PGHS-l and PGHS-2 in cells where both isozymes are expressed is the existence of
accessory proteins which differentially affect the rate of prostaglandin endoperoxide
formation by PGHS-l versus PGHS-2. Although no such protein(s) has been identified in
the prostanoid biosynthetic system, there is a precedent for an accessory protein in the
leukotriene pathway. Leukotrienes synthesized through 5-lipoxygenase arise from

arachidonic acid apparently delivered to the 5-lipoxygenase by an activating protein,

 

 

Figure 7. Schematic diagram of the biphasic prostaglandin synthesis in activated
inflammatory cells.

30

FLAP [85, 86].

Prostaglandin Signaling via G-protein Coupled Receptors

Prostaglandins act as short-lived local hormones to maintain homeostasis in the cells
from which they were released and their neighbors. Immediately after their synthesis,
prostaglandins are released from the cell via at least one known prostaglandin transporter
[87, 8 8]. The family of G-protein coupled receptors through which prostaglandins exert their
signaling messages was classified by Kennedy and Coleman in 1983 [29]. Specific receptors
exist for PGEZ, PGIZ, PGD2, PGFm, and thromboxane. The corresponding receptors are
known as EP, IP, DP, FP, and TP receptors, respectively, and some subclasses exist (EPl,
EP2, etc.) These receptors are coupled to heterotrimeric G-proteins which in turn modulate
the activities of effector proteins to control cellular levels of CAMP, Ca2+, and phosphatidyl
inositol [89]. Nearly all mammalian tissues possess at least some prostaglandin receptors.

Structurally, all the known prostanoid receptors are thought to resemble the
rhodopsin family of seven transmembrane domain receptors based on similarities to known
rhodpsin-like structures and hydrophobic sequence analysis [89]. The N-terminal
extracellular regions contain several N-glycosylation sites and about 35% of the receptor
mass is carbohydrate [90]. Two cysteines in the first and second intracellular loops may
provide structural stability through a disulfide bridge. The seventh transmembrane domain
of all prostanoid receptors contains a conserved arginine. This residue probably participates
in binding a part of the prostaglandins present in all species. Funk, et al. demonstrated in
1993 that a point mutation at this arginine residue eliminates the ability of the thromboxane

receptor (TP) to bind its ligand [166]. Intracellularly, a series of Ser and Tyr

31

phosphorylation sites are thought to serve as desensitization switches like those found in the
adrenergic and muscarinic receptors of the rhodopsin family, and some work has been done
to demonstrate this [91-95]. Prostaglandin receptors, like PGHS-1 and -2, are important
clinical targets and prostaglandins and their analogues are frequently used clinically to
combat cardiovascular disease, induce labor in pregnant women, and relieve erectile
dysfuntion in impotent men [96].

Prostaglandins are regulatory hormones that often work in opposition to one another
to maintain homeostasis in a given tissue. For example, thrombin stimulates platelets to
release thromboxane A2 which causes platelets to aggregate. Balancing this action,
thrombin-induced PGI2 release fiom endothelial cells inhibits platelet aggregation, thus
preventing unnecessary clotting. A firrther example of the opposing actions of
prostaglandins can be observed in vascular smooth muscle. PGE2 can cause the contraction
or relaxation of vascular and nonvascular muscles through EPl and EP2 receptors,
respectively, because the two receptors are coupled to different G proteins that mediate
opposing actions. Thus, the transcriptional regulation of prostaglandin receptors can
significantly affect a tissue’s response to a given stimulus.

All the prostanoid receptors bind their respective ligands with binding constants of
1.3-40 nM [89]. At concentrations higher than this, cross reactivity is observed. For
instance, PGI can activate the EPl receptor as potently as PGE2 [97]. It is therefore important
to take into account the possibility for receptor cross reactivity when studying cells that

express more than one PG receptor.

32

Evidence for Nuclear Prostaglandin Signaling Systems.

The physiological effects of prostaglandins are best understood through their role at
the plasma membrane where they bind to specific G-protein coupled receptors. Recently, the
idea that prostaglandins act exclusively through extracellular mechanisms has become the
subject of debate. Due in part to the correlation of PGHS-2 expression with nuclear events
like increased gene transcription and mitogenesis [4], it has been hypothesized that a subset
of prostaglandins enter the nucleus and modulate transcription through a group of orphan
nuclear receptors. This section will review the evidence for and against the possibility of
nuclear prostaglandin signaling mediated by the orphan nuclear receptor, PPARy.

Peroxisome proliferator activated receptors (PPAROL, PPARy, and PPAR8) are
members of the nuclear hormone receptor superfamily [98]. Upon ligand binding, these
receptors form a heterodimer with the retinoid X receptor (in the case of PPARy) which in
turn activates the transcription of genes involved in lipid metabolism and biosynthesis. Like
other members of the nuclear receptor superfamily, PPARs are comprised of a ligand
binding domain, a hinge domain, and a DNA-binding domain. In the case of PPAR’Y the
DNA binding domain binds to a specific 6-nucleotide sequence dubbed DR-l [99]. The
endogenous ligands for the PPARs are unknown, but several synthetic anti-diabetic drugs of
the thiazolidinedione family (e.g., troglitazone and BRL49653) can act as ligands in vivo and
in vitro [100]. Despite appreciable sequence homology, the PPAR’s exhibit very different
tissue distribution and are known to bind ligands with differing affinities.

The biological role of PPARy is best understood in the context of the differentiation

of adipose tissue. This process is controlled by the interplay between PPARy and other

33
transcription factors [101]. Since its discovery, the endogenous ligand for PPARy has been
unknown. Recently, some have speculated that a prostaglandin or other eicosanoid may be
the endogenous ligand for the orphan nuclear receptor, PPARy. This hypothesis is based on
the ability of certain prostaglandins and unsaturated fatty acids to activate transcription via
PPARY under experimental conditions [102, 103]. Activation of PPARy with ligands is
sufficient to induce adipogenesis in several fibroblast lines [101]. The likelihood that
adipose tissue itself has prostaglandin biosynthetic capacity is suspect, but circumstantial
evidence is beginning to gather in support of a role for nuclear prostaglandin signaling in
other cell types. For instance, stimuli that induce the expression of PGHS-2 in monocytes

can also trigger the expression of PPARy.

Many of the recent studies positing a role for prostaglandins in PPARy—mediated
nuclear signaling have been conducted using cells transfected to express PPAR‘y. Treatment
of PPARy-transfected NIH 3T3 mouse fibroblasts with the prostaglandin lS-deoxy-A‘Z'M-
prostaglandin J2 causes their dramatic differentiation to adipocytes [104]. Similarly, genes
known to be regulated by PPARy can be co-transfected into cells along with PPAR‘Y and

turned on with exogenous treatment with 15d-PGJ2. lSd—PGJZ, is a naturally-occurring
metabolite of PGDZ, the major prostanoid product in monocytes and macrophages.
Prostaglandins can also modulate transcription in cells that express PPARy
endogenously. 15 d—PGJ2 can prevent the usual transcription of at least three inflammatory
genes in activated mouse peritoneal macrophages [105]. Further, the normal production of
inflammatory cytokines by activated monocytes can be inhibited by treatment with lSd-PGJ2

[106]. In all of the studies mentioned above, the involvement of PPARy in the observed

34

phenomena was confirmed by treatment with the known PPARy ligands troglitazone and
BRL, members of a family of anti-diabetic drugs that were some of the first identified
PPARy ligands. However, the case for nuclear action by lSd-PGJ2 is far from closed.

Prostaglandins bind to their respective plasma membrane receptors with Kd values
between 1.3 and 40 nM [89]. Known endogenous ligands for members of the nuclear
hormone receptor superfamily like the retinoid X receptor (PPARy’s heterodimerization
partner) have Kd’s that are typically <1 nM. In contrast, all of the studies purporting 15d-
PGJ2 to be the ligand for PPARy have used ligand concentrations of at least 100 nM and
often higher. At these concentrations of ligand one must take into account at least three
considerations: (a) lSd-PGJ2 may non-specifically activate one of the known plasma
membrane receptors (e. g., EPl , EP2, or FPl ); (b) there may be cell surface receptors for 15 d-
PGJ2 itself; and (c) exogenously administered 15d-PGJ2 would have to cross the plasma
membrane to reach PPAR'y. The latter point refers to the fact that prostaglandins are
generally not able to diffuse across cellular membranes. Exogenously administered
prostaglandins may require a transporter to reach the nucleus.

Bruce Speigelman at Harvard Medical School is currently using a slightly different

approach in his search for the endogenous ligand for PPARy. Instead of treating cells with
suspected natural ligands of PPARy, his group is collecting lipid-soluble molecules in tissue
culture supematants and testing these extracts for the ability to modulate PPARy -mediated
transcription. PPARy is known to interact functionally with SREBP-l, a protein implicated

in fatty acid and cholesterol metabolism [101]. Spiegelman’s group recently showed that

the expression of SREBPl specifically increases the transcriptional activity of PPARy

35
through its ligand binding domain [107]. This stimulation of PPARy by SREBPl does not
require co-expression in the same cells; lipids extracted from the supematants of cells

expressing SREBPl can stimulate the transcriptional activity of PPARy in separate

cultures. Biochemical experiments showed that cells expressing SREBPI secrete lipid
molecules that bind directly to PPARy.

From these studies, Speigleman’s group concluded that SREBPl stimulates the
production of an unknown lipid mediator which can then enter the nucleus and modulate

transcription via PPARy. Here again one should consider that prostaglandins are generally

impermeable to cellular membranes. However, some studies of prostaglandin metabolism
in the lung have suggested that prostaglandins can indeed be transported into the cell in a
specific manner [108]. A similar transport system could conceivably act to facilitate
prostanoid access to the nucleus. Although the existence of eicosanoid signaling in the
nucleus has not been rigorously described, the localization of the required biosynthetic
enzymes and prostaglandin activation of PPARy provide important leads. Spiegelman’s
work is interesting and his group is currently attempting to determine the nature of the
SREBP-l regulated factor that activates PPARy. Preliminary results suggest that it is not

warm, [107].

The Nuclear Envelope: Structure and Membrane Protein Trafficking
As integral membrane proteins of the ER and nuclear envelope, PGHS-1 and PGHS-
2 are interesting in terms of their cell biology and the strategies they use to reach the cellular

membranes where their activities are necessary. Immunolabeling studies of PGHS-1 and -2

36

discussed in Chapter II of this dissertation provide the first evidence that an integral
membrane of the ER can reach the inner membrane of the nuclear envelope, presumably by
lateral diffusion.

The membranes of the nuclear envelope serve to compartmentalize the nucleus of
eukaryotic cells. The inner nuclear membrane, outer nuclear membrane, nuclear pore
membranes, and the ER form a continuous membrane system (Fig. 8) [109]. The outer
nuclear membrane is essentially part of the ER; any membrane proteins present in the ER are
therefore expected to be in the outer nuclear membrane by extension. The inner nuclear
membrane contacts the outer nuclear membrane only at the nuclear pore complexes (N PC).
Thus, any protein able to reach the inner nuclear membrane must first circumvent the large
proteinaceous NPC. The inner nuclear membrane is functionally and biochemically distinct
from the outer membrane. Several proteins have now been localized exclusively to the inner
nuclear membrane, including the lamin B receptor (LBR), LAPlC, LAP2, and emerin [110-
114]. All these proteins play a role in the structural integrity of the nuclear envelope or
interphase chromosomes through interactions with chromatin and the nuclear matrix.

A model has emerged to describe how integral membrane proteins become localized
to the inner nuclear membrane after their synthesis in the ER [115]. Proteins are thought
diffuse laterally in the membrane bilayer from the outer nuclear membrane through the
nuclear pore membranes until they reach the inner nuclear membrane. Ellenberg, et al. used
fluorescence recovery techniques to provide evidence for the lateral diffusion model using
LBR/GFP fusion proteins [116]. This study demonstrated that recently synthesized LBR in

the ER could freely diffuse in the lipid bilayer until it reached the inner nuclear membrane

37

where it became anchored to its ligand. Due to the physical barrier of the NPC, certain types
of membrane proteins are prevented from reaching the inner nuclear membrane via lateral
diffusion. For example, membrane proteins with large (>70 kDa) cytoplasmic/nucleoplasmic
domains fail to reach the inner nuclear membrane (see the Chapter II Discussion) [1 14]. This
observation is likely due to a size constraint imposed by the lateral channel diameter of the
NPC.

PGHS-1 and -2 are interesting in this regard. Both proteins have large, globular
extramembrane domains, but due to their lumenal orientation in the membrane, can
apparently diffuse laterally past the NPC and to the inner nuclear membrane where they may
serve to synthesize prostaglandins that enter the nucleus. The PGHSs are (a) the first ER
membrane proteins with a demonstrated presence on the inner nuclear membrane, and (b)
the first known proteins of the inner nuclear membrane with no known role in the structural
integrity of the nucleus or nuclear envelope. Further studies in this area will provide
fimdamental evidence pursuant to the properties of membrane proteins and their trafficking

patterns within cells.

PGHS Catalysis and Inhibition by NSAIDS
PGHSs contain two enzymatic activities. The cyclooxygenase activity catalyzes the
bis-oxygenation of arachidonic acid to give PGGZ. The peroxidase activity then reduces the
15-hydroperoxyl group of PGG2 to form PGH2 (Fig. 2). The existence of two activities in
a single enzyme make the kinetic properties of PGHS-l and -2 interesting and complex [2].
Crystallographic analysis reveals that although these active sites neighbor each other, they

are on opposite sides of the heme and appear to be distinct [34].

38

Outer nuclear
membrane

Nuclear pore 2 ER L N
‘ ‘ Q
l

NUCLEOPLASM Inner nuclear
membrane

 

Figure 8. The continuous membrane systems of the ER and nuclear envelope. The
endoplasmic reticulum is attached to the nuclear envelope by virtue of its continuity with
the outer nuclear membrane. The inner and outer nuclear membranes are separated by the
nuclear pore membranes which surround the nuclear pore complexes (NPC). The NPC
serves as a barrier to free diffusion of membrane proteins from the outer nuclear
membrane to the inner nuclear membrane, and thus keep the inner and outer membranes
functionally and biochemically separate. Modified from Ellenberg, et al. 1997.

39

Following the binding of substrate, cyclooxygenase activity is triggered by a
hydroperoxidedriven oxidation of the heme iron at the peroxidase active site (Fig. 9). With
its iron atom now in a Fe4+ oxidation state, the heme can oxidize Tyr3852 to create a tyrosyl
radical. This radical is thought to perform the rate-limiting abstraction of the l3-pro S
hydrogen of arachidonic acid. Once this occurs, molecular oxygen is able to attack the
substrate at position 1 l and form a bridged endoperoxide with carbon 9. A second molecule
of oxygen is then added at carbon 15 to give PGGZ, which is quickly reduced at the
peroxidase active site to form PGH2.

How the PGG2 moves from the cyclooxygenase channel to the peroxidase active site
is poorly understood. Structural biologists generally agree that it is unlikely PGG2 moves
from the proximal to the distal side of the heme through the protein. Instead, the newly
oxygenated substrate probably exits the hydrophobic channel of the cyclooxygenase active
site through an unknown route and diffuses to the peroxidase site. The lower region of the
active site channel contains one possible exit route between Helix A and the EGF domain
(see Fig. 6). Conformational changes in the membrane associated region of the protein may
help to facilitate PGG2 diffusion, and the PGHSs are generally flexible in this region
suggesting other exit routes may exist.

Several other amino acids in both PGHS isozymes are required for activity. At the
mouth of the cyclooxygenase channel, arginine 120 interacts with the carboxyl group of

arachidonic acid to position the substrate for hydrogen abstraction by Tyr385. Similar

 

2Amino acid numbers correspond to the sequence of ovine PGHS-1.

4O

interactions are important in PGHS-2, but to a lesser extent3 . The heme group is coordinated
by two histidine residues, with His388 and HisZO7 acting as the proximal and distal ligands,
respectively [34, 117]. Mutations at these positions result in a loss of both enzymatic
activities.

Both PGHS-1 and -2 undergo a poorly understood phenomenon called suicide
inactivation. Oxygen uptake assays of preparations of PGHS-1 or -2 show an initial burst
of O2 consumption as PGG2 is formed that tapers off within two minutes. On average, each
enzyme molecule can turnover about 400 times before becoming suicide inactivated. How
this inactivation occurs is unknown, but it probably results from the decay of an unstable
enzyme-substrate intermediate. Interestingly, some mutations (His386Ala and ArngOGlu)
result in an enzyme that does not undergo suicide inactivation [118]. The His 386 residue
may be involved in the transport of the radical from neighboring Y385 that may somehow
trigger an aberrant cross linking reaction that kills the enzyme.

All common NSAIDS (e. g., aspirin, ibuprofen) inhibit prostaglandin synthesis by

binding to or covalently modifying the cyclooxygenase active site of the PGHSs. Aspirin
exerts its pharmacological action by acetylating serine 530 in PGHS-1 [25, 119]. PGHS-2
is also acetylated by aspirin but unlike PGHS-1 is not completely prevented from
oxygenating the arachidonate and can make 15 -HETE [120]. This suggests that the PGHS-2

active site is larger

 

3C.J. Rieke and W.L.Smith, unpublished results.

 

 

 

 

 

 

 

B
SUICIDE INACTIVATION CYCLOOXYGENASE
A
I 7 W \
, no2
PPlX-Fo 4+.-.o
I “1*“ IPPIX-F- +=01 [PPlX-FOMEO] 3‘5 Tyr
LS OZAAOO
:;:w A" P66 2
[WW-“3 1 Lgamut-Flo “=01 «c—r [PPlx-Fo4+=0]3351- '
>:00 “/o', [7] " ’ II
[PPlX-Fo “:01
L J
Y
PEROXIDASE

Figure 9. Peroxidase and cyclooxygenase catalysis. A, model of the cyclooxygenase and
peroxidase active sites of ovine PGHS-1. An alkyl hydroperoxide is shown bound to the
heme group at the peroxidase active site, and arachidonate is shown bound to the
cyclooxygenase active site. His-388 and His-207 are the proximal and distal heme ligands,
respectively. Tyr-3 85 is likely the residue that abstracts the (l3S)-hydrogen fi'om
arachidonate via a tyrosyl radical, thereby initiating cyclooxygenase catalysis. Ser-530 is the
site of aspirin acetylation. Arg-120 is present at the opening of the hydrophobic substrate
binding channel and is the countering for the carboxylate group of arachidonate. B, A model
of PGHS-1 and PGHS-2 catalysis. A two-electron oxidation of the heme group of PGHS by
a hydroperoxide yields a peroxidase spectral Intermediate I containing an oxyferryl form of
iron (Fe(IV)) and a protoporphyrin radical cation. The oxidized heme group in turn oxidizes
a neighboring tyrosine residue, probably Tyr-385 to yield peroxidase Intermediate 11 having
a tyrosyl radical and an oxyferryl Fe(IV). This protein radical is likely the species that
abstracts the (l 3S)-hydro gen from arachidonate. PPIX, protoporphyrin IX; AA, arachidonic
acid.

42

and more accommodating than that of PGHS-1. Crystallography supports this view, and the
relatively large compounds that specifically inhibit PGHS-2 take advantage of the space
created by a valine to isoleucine substitution in PGHS-2 [35 , 121].

Covalent modification of the enzymes is just one of three different inhibition
mechanisms employed by common NSAIDS. Ibuprofen, the active ingredient in Advil®
and Motrin®, inhibits the cyclooxygenase chemistry by simple reversible, competitive
inhibition. Flurbiprofen and indomethacin are only slowly reversible; they are classified as
time-dependent, reversible inhibitors. These drugs initially bind reversibly, but soon drive
the formation of an El* complex from which their dissociation is very slow. Currently
available NSAIDS inhibit both PGHS-1 and -2 with essentially equal potency.

When PGHS-2 was discovered it was almost immediately recognized by the
pharmaceutical industry as a prime target to reduce pain and inflammation. A new group of
NSAIDS are currently in clinical trials that were developed with the goal of inhibiting
prostaglandin synthesis involved in inflammation while avoiding the gastrointestinal
ulceration often associated with the inhibition of PGHS-1 . These PGHS-2 specific inhibitors
are all time-dependent, reversible inhibitors and have shown initial promise as non-
ulcerogenic anti-inflammatory and anti-pyretic agents. Although their capacity to inhibit
prostaglandin synthesis may be no better than that of common NSAIDS like aspirin, a
significantly safer side-effect profile may allow PGHS-2 specific inhibitors to carve out an

important clinical niche.

43

CHAPTER II

SUBCELLULAR LOCALIZATION OF PGHS-1 AND PGHS-2 BY
IMMUNOELECTRON MICROSCOPY
Introduction

Prostaglandin endoperoxide H synthases-1 and -2 (PGHS-1 and -2) are the major
targets of nonsteroidal anti-inflammatory drugs like aspirin and ibuprofen. These enzymes
catalyze the committed step in the formation of prostanoids from arachidonic acid. Although
PGHS-1 and -2 are similar biochemically, a number of studies suggest that PGHS-1 and
PGHS-2 function independently to form prostanoids which subserve different cellular
functions. We have hypothesized that these isozymes may reside, at least in part, in different
subcellular compartments and that their compartmentation may affect their access to
arachidonic acid and serve to separate the functions of the enzymes. To obtain high-
resolution data on the subcellular locations of PGHS-1 and -2, we employed immunoelectron
microscopy with multiple antibodies specific to each isozyme. Both PGHS-l and -2 were
found on the lumenal surfaces of the endoplasmic reticulum (ER) and nuclear envelope of
human monocytes, murine NIH 3T3 cells, and human umbilical vein endothelial cells.
Within the nuclear envelope, PGHS-l and -2 were present on both the inner and outer
nuclear membranes and in similar proportions. Western blotting data showed a similar
distribution of PGHS-l and -2 in subcellular fractions, and product analysis using isozyme-
specific inhibitors suggested that both enzymes generate the same products in NIH 3T3 cells.
Thus, we are unable to attribute the independent firnctioning of PGHS-1 and PGHS-2 to

differences in their subcellular locations. Instead, the independent operation of these

44

isozymes may be attributable to subtle kinetic differences such that at low concentrations of
arachidonate (ca. 50-1000 11M) this substrate is preferentially utilized by PGHS-2. A further
conclusion of importance from a cell biological perspective is that membrane proteins such
as PGHS-l and -2, which are located on the lumenal surface of the ER, are able to diffuse

freely among the ER and the inner and outer membranes of the nuclear envelope.

45

Methods
Materials. All materials were purchased from Sigma Chemical Company unless

otherwise noted.

Isolation and preparation of human monocytes for electron microscopy. Peripheral
blood mononuclear cells (PBMCs) were isolated from several healthy human volunteers.
Heparinized blood was diluted 1:1 in Hanks Balanced Salts Solution (HBSS, GIBCO BRL,
Grand Island, NY) and spun against a Ficol-Hypaque (LSM - Organon Teknika) gradient for
20 min at 500 x g. PBMCs were isolated as a band from the interface and washed 2 x with
HBSS before resuspension in 10 ml of RPMI medium (GIBCO BRL, Grand Island, NY)
containing 5% autologous serum and penicillin-streptomycin (100 U/ml). Cells (106/ml)
were incubated in the presence or absence of 1 ng/ml of lipopolysaccharide (LPS; from Jayne
Chen at Merck) and incubated with gentle agitation in a water-saturated 5% CO2 atmosphere
at 37 °C for 24 h. Fixation was performed after LPS- or vehicle (buffer alone)-stimulated
cells were collected by centrifugation and resuspended in appropriate media (see below).
Monocytes comprise 5-8% of PBMCs. Most of the other cells in PBMC preparations are T
and B lymphocytes which do not normally express PGHSs and, therefore, provide a
convenient internal negative control against PGHS immunostaining. Because ultratlrin
sectioning makes possible the observation of more than one section from a given cell, care
was taken to use several different samples in the interest of gathering data from many
different cells. PBMCs from each volunteer were separated into between five to ten separate
pellets per experiment. Between every few grids of thin sections taken from a single pellet,
several thick (1 pm) sections were cut to maximize the number of different cells observed

per experiment.

46

Preparation of NIH 3T 3 cells and H U VEC cells for electron microscopy. Murine
NIH 3T3 cells express PGHS-l constitutively [49]. Serum-starved, quiescent 3T3 cells were

cultured and then stimulated by the addition of 10% fetal calf serum (HyClone Laboratories,

 

Inc., Logan, UT) for 3 h to induce PGHS-2 as described previously [49]. After serum
stimulation, cells from five 100 mm culture dishes were removed from the growing surface
with a rubber policeman, resuspended in DMEM (GIBCO BRL, Grand Island, NY) and
immediately fixed (see below). HUVECs (Cell Systems, Seattle, WA) were thawed fiom
stocks frozen at passage 1 and expanded through passage 3 - 4 as suggested by the
manufacturer. After stimulation for 20 h with 10 ng/ml IL-lB, cells were removed from
culture dishes with a rubber policeman, immediately resuspended in CS-C media (Cell
Systems, Seattle, WA) and fixed (see below). Previous PGHS immunolocalization
experiments on both adherant and removed cells gave similar results, suggesting that
removal of cells from the growing surface does not affect the localization of PGHS-l or
PGHS-2 [28, 67].

Fixation and cryoprotection for electron microscopy. Cells (PBMCs, NIH 3T3,
HUVEC) were obtained in a 2.5 ml suspension of an appropriate medium and immediately
fixed by the addition of 2.5 ml of 2X fixative. Two different fixation methods were
employed to allow for differences in the fixation sensitivity of PGHS-1 and -2 antigens.
Microwave/glutaraldehyde fixed cells were fixed exactly as described previously [69].
Cells fixed with Nakane fixative [122] alone were treated by adding 2X Nakane fixative (1X
Nakane = 0.1 M NaIO4, 0.75 M Lysine, 0.0375 M phosphate buffer, 2% paraformaldehyde

(Fisher)) to an equal volume of cell suspension and incubated for two h at room temperature.

47

These cells were then washed twice with 1X Nakane fixative and resuspended in fresh 1X
Nakane before an overnight (or shorter) incubation at 4 °C. NIH 3T3 cells immunolabeled
for PGHS-2 were only fixed at 4 0C for an additional 2 h after the 2 h room temperature
fixation. After the overnight fixation step, both microwave/glutaraldehyde— and Nakane-
fixed cells were collected by centri fugation, resuspended in 0. 1 M sucrose/PB S, and pelleted
in 2% low gelling temperature agarose/0.l M sucrose/PBS. Cryoprotection was performed
for 2 h (25 0C) or overnight (4 0C) in polyvinylpyrrolidone/2.3 M sucrose/phosphate buffer
pH 7.2 (PVP-sucrose) [69]. Cell pellets were mounted on bullseye specimen pins (Ted Pella,
Inc.), frozen by plunging into liquid propane, and stored under liquid nitrogen until use.
Immunogold labeling. Immunogold labeling was performed essentially as described
previously [69]. Briefly, ultrathin (75-80 mm) sections of cells were cut at -106 °C on a
Reichert Ultracut S ultramicrotome fitted with a Reichert FCS cryoattachment. Sections
were collected on drops of 2.3 M sucrose and placed on glow discharged, formvar-ooated
nickel grids (Ted Pella, Inc.). After a minimum of one h in blocking solution (5% milk/1%
BSA/PBS/0.02% sodium azide) and washing, grids were placed section side down on 25 ul
drops of primary antibody solutions of various concentrations and incubated for various
times depending on the primary antibody. All antibody solutions were cleaned by filtration
through 0.2 pm filters prior to use. To control for antibody specificity, primary antibodies
were incubated with a 50-fold molar excess of either cognate peptide (for anti-PGHS-l or
anti-PGHS-2 peptide directed antibodies) or an approximately 10-fold molar excess of
purified ovine PGHS-l or -2 (for antibodies prepared against either whole protein).

Preadsorption of whole protein antibodies with purified PGHS-1 or -2 was performed with

48

constant agitation at 4 °C for 24 h. The samples were centrifuged at 12,000 rpm in a
Beckrnan microcentrifuge for l h at 4 °C, the supematants were removed, filtered through
a 0.2 pm syringe filter, and used as primary antibody solutions in immunogold labeling
experiments. Purified ovine PGHS-l was from RM. Garavito at Michigan State University;
purified ovine PGHS-2 was from Cayman Chemical Co., Ann Arbor, MI. After incubation
with primary antibody, each grid was washed eight times for three min on drops of 1%
BSA/PBS/0.02% sodium azide before a l h incubation at room temperature on drops of gold-
conjugated secondary antibody (Amersham GAR-G5 diluted 1:75 in 1% BSA/PBS/0.02%
sodium azide). After a washing step, sections were post-fixed and stained for contrast by
floating the grids sequentially on drops of 2% glutaraldehyde, 2% osmium tetroxide, and 2%
uranyl acetate. Polyvinyl alcohol (2%) was used to embed the grids before observation by
transmission electron microscopy [123].

Antibodies specific for PGHS-I and PGHS-2. All primary antibodies were raised in
rabbits as described previously [42]. For PGHS-2 staining of human monocytes and
HUVECs, a polyclonal antibody raised against ovine PGHS-2 was used. This antibody
(fiom Dr. Jilly Evans, Merck Frosst) cross-reacts with human PGHS-2 but not with ovine
or human PGHS-l [124]. PGHS-l immunostaining of NIH 3T3 mouse fibroblasts was
performed with an affinity purified anti-peptide antibody directed against amino acids
Leu274 - Arg288 of murine PGHS-1 [125]. This antibody does not cross react with PGHS-2
in Western blotting experiments. Another antibody, an IgG fraction from a polyclonal
antibody raised against PGHS-1 [126], recognizes only PGHS-1 and was used for

immunolabeling of NIH 3T3 cells and HUVECs. PGHS-2 labeling of NIH 3T3 cells was

49

performed using an affinity purified anti-peptide antibody directed against an 18 amino acid
cassette near the C-terrninus of PGHS-2 [67, 127]; this antibody does not cross react with
any known PGHS-1 but recognizes human and murine PGHS-2 on Western blots.

Distribution of PGHS-l and PGHS-2 between inner and outer membranes of the
nuclear envelope. We determined the distribution of PGHS-l and PGHS-2 between the
inner and outer nuclear membranes essentially as described previously [69]. Gold particles
lying on or within one 5 nm particle diameter of the inner nuclear membrane were designated
as being on the inner nuclear membrane. Particles lying on or within one 5 nm particle
diameter of the outer nuclear membrane were designated as being on the outer nuclear
membrane. When a gold particle was observed in the nuclear envelope and not within one
particle diameter of either membrane, it was designated as lumenal. Analysis of the
distribution of PGHS-1 and PGHS-2 was performed on a total of 24 and 32 cells,
respectively. Distribution analysis was limited to well-preserved sections of nuclear
envelope. Only regions of cells in which the inner and outer nuclear envelope were clearly
distinguishable were used for our analyses.

Statistical Analysis. The distribution analysis described above was performed using
at least three experimental groups of well-preserved cells for both PGHS-l in NIH 3T3 cells
and PGHS-2 in monocytes. The analysis of PGHS-2 in NIH 3T 3 cells was performed on
seven cells taken from two separate experiments. Analysis of the mean distribution
percentages using Student’s t test showed that there was no significant difference between
the inner membrane/outer membrane distribution of PGHS-l and -2 in the cell types

analyzed. It is possible for a protein of the nuclear envelope to be present in significantly

50

different amounts, as evidenced by previous work on 5-lipoxygenase and FLAP [69].
Subcellularflactionation and Western analysis. NIH 3T3 cells were prepared as
whole cell lysates, microsomes, or isolated nuclei. For whole cells, harvests of three 100 mm
culture dishes of serum-stimulated NIH 3T3 cells in PBS were collected by centrifugation
and resuspended in Hanks Balanced Salts Solution (HBSS). The resuspended cells were then
sonicated and dounce homogenized to produce whole cell lysates. Microsomes were
prepared as described previously [27, 117, 128] from ten plates of cells except that the
200,000 x g pellet was resuspended in HBSS. Isolation of cell nuclei [129, 130] began with
the harvest of twenty dishes of 3T3 cells in PBS followed by centrifugation at 1000 x g.
Nuclei were isolated in the presence of 0.2% or 1% saponin as described below. Isolation
of nuclei in the absence of detergent was begun by resuspension in 10 ml of cold Buffer A
(10 mM Tris, 10 mM NaCl, 1 mM EDTA, 0.5 mM EGTA, 1 mM PMSF, 1 pg leupeptin/ml,
and 3 mM MgC12’ pH 8.0) and incubation on ice for 5 min, followed by 5 seconds of gentle
vortexing. After passage ten times through a 20-gauge needle, the crude nuclei were again
collected by centrifugation at 1000 x g. The pellet was resuspended in 6 ml of cold Buffer
A and incubated on ice for 5 min. After an additional ten passages through a 20-gauge
needle, the suspension was subjected to gentle homogenization in a Teflon homogenizer.
Half of the suspension was placed in each of two Beckrnan 5 ml UltraClear ultracentrifirge
tubes. A nuclear spin cushion was prepared by dissolving 6.16 g sucrose in 10 ml of Buffer
B (60 mM KCl, 15 mM NaCl, 15 mM Tris, 0.15 mM spermine, 0.5 mM sperrnidine, and 0.5
mM DTT) to give a 1.8 M sucrose solution. After underlaying 2 ml of the sucrose solution

beneath the crude nuclear suspension, the tubes were centrifuged for 19 min at 4 °C at 13,500

51

rpm in a Beckrnan SW50.1 rotor. Isolated nuclei were visible at the bottom of the tubes.

The liquid was removed from the tubes by aspiration, and the nuclear pellet was resuspended
in 400 pl of HBSS. Nuclei were washed 4 times with HBSS and collected by centrifugation
at 1800 rpm for three min in a microcentrifuge. Purified nuclei and whole cell lysates were
disrupted by sonication before protein concentrations were determined using a BioRad
Protein Assay solution. Aliquots of broken cells, microsomes, or nuclei (20 pg) were
separated by SDS-PAGE and analyzed by Western blotting as described previously [41].
Densitometric quantitation of immunoreactive PGHS-1 or -2 was performed using a BioRad
GS-505 Molecular Imaging System and Molecular Analyst software. For a given isozyme,
densities are expressed as the ratio of nuclear to microsomal immunoreactivity. Student’s
t test was utilized to determine if these ratios for PGHS-l and PGHS-2 were significantly

different for each experimental condition.

52

Results
Immunogold labeling of PGHS-2 in human monocytes and HUVECs. Peripheral

blood mononuclear cells (PBMCs) were isolated from healthy human volunteers, treated for
20 h with bacterial lipopolysaccharide (LPS), and processed for electron microscopy.
Western blotting of samples of these cells prior to fixation established that, as expected,
PGHS-2 expression was induced by LPS, but that vehicle-treated cells lacked detectable
levels of this isozyme [131]. Monocytes in LPS-treated samples, when incubated with an
antibody raised against ovine PGHS-2, exhibited perinuclear staining for PGHS-2 (Fig.
10A). PGHS-2 labeling was completely eliminated by preadsorption of the anti-PGHS-2
antibody with purified ovine PGHS-2 (Figure 108). No comparable gold label was seen in
vehicle-treated cells (data not shown). PGHS-2 labeling was observed on both the inner and
outer membranes of the nuclear envelope in monocytes present in LPS-treated PBMC
preparations (Fig. 10C). The distribution of PGHS-2 within the nuclear envelope of 26 LPS-
treated monocytes was determined by counting the gold particles associated with well-
preserved regions of the inner and outer nuclear membranes (Table I). In performing these
distribution analyses, only those segments of the nuclear envelope in which both the inner
and outer nuclear membranes were clearly distinguishable were used. PGHS-2 labeling was
approximately equally distributed between the inner and outer membranes of the nuclear
envelope.

To determine if the distribution pattern of PGHS-2 was the same in cell types other
than monocytes, we performed additional studies with (a) serum treated murine NIH 3T3
cells and (b) IL-12-treated human umbilical vein endothelial cells (HUVEC) both of which

are known to express PGHS-2 [49, 132]. An anti-peptide antibody raised against the

53

.81 :6 u :0: 030m
.8233»: 20:02:38 :0 ”320:: . Z 30:89:08 320:: .050 0:0 0:2 05
:0038: 8:392: ”$29: :0 20:25 0:: E :00: 0:02:08 :0 035800050:
0:20>:0 320:: .20 80:00: 00300892203 030:2: 030.5: :0:0 600030
AB 5 0:05:08 320:: 5 0:2 05 :0 3:2 3:820 33:30:: 0m:2 :5
0:05:88 320:: A3 00:0 0:: :0 0225: 20m 328:0: 83:30.5 :08: :23
0:20>:0 320:: a :0 5:0: ADV 0:00:305 :08: h: 080:0: 20:2 20w ::3
800E030 0:: 320:: 0:: m0 :0E0: a w:m30:m 00:80:08 «0 0:0U00m Am .3
.0Ew:_.:00 Eowéwm 2:58.250 3:00:83 0 ::3 E0830: 3: 003220.: 003 2:8
.20 3:0: 8:5: as, 88882: as: ea 6:3 3:0: .3 Beam
.3028: SEC: 0 :0 0.3 «-mmcm :0: 0:02“: .8026: SEE: a 05:0 :23
020:2 0:03 80:88::0 2:305 .08: 0:: E 02:30: 0: 30088:: :05020
:0: 0000002: 0:0 :05: 02202 0:03 0:00 :020::0:0:: :02: aggro:
583: 600.028.: 5:5: 5 «.mmcm «0 mam—0:2 Bog—SEE: .3 gr:

54

 

55

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

23221}, $126) Total Gold Particles Percentage of Total Gold Particles (%)
Inner Outer Lumeml Inner Outer Lumenal

PGHS-2 123 165 52 37i2 491-2 15:1:3
(monocyte)

lPGHS-Z 27 20 8 473:7 3916 14i3

(3T3 cells)

PGHS-1 82 83 34 41i2 43:I:4 172t3

I(3T3 cells)

! Ii

 

 

Table 1. Analysis of PGHS-1 and PGHS-2 distribution within the nuclear envelope.
The distributions of PGH S-1 and PGHS-2 within the nuclear envelope of LP S-treated human
monocytes and serum-treated murine NIH 3T3 cells were determined as detailed in Materials
and Methods. In this analysis gold particles were counted as being associated with either the
inner or outer nuclear membrane if they were touching or within one particle diameter of that
membrane. Other particles, not within one particle diameter of either membrane, were
designated here as lumenal. Only well-preserved regions of nuclear envelopes (i.e. regions
in which both the inner and outer nuclear membranes were visible) were used for this
analysis. For monocytes, particles representing PGHS-2 from a total of 26 cells were
counted. For NIH 3T3 cells, a total of 27 and 7 cells were analyzed for particles
corresponding to PGHS-1 and PGHS-2, respectively. Analysis of the mean distribution
percentages using Student’s t test showed that there was no significant difference between
the inner membrane/outer membrane distribution of PGHS-1 and -2 in the cell types
analyzed.

56

C-terminal 18 amino acid cassette of PGHS-2 was used to label PGHS-2 in serum-treated
NIH 3T3 cells. This antibody labeled serum-treated NIH 3T3 cells along the nuclear
membrane and in the ER (not shown). PGHS-2 immunoreactivity was distributed equally
between the inner and outer nuclear membranes in NIH 3T3 cells (Table I). PGHS-2
labeling of serum-treated NIH 3T3 cells was eliminated by preincubation of the antibody
with its cognate peptide. Finally, PGHS-2 staining in IL-lB- treated HUVECs was present
on the inner and outer nuclear membranes and in the ER, a pattern similar to that seen in
LPS-treated monocytes and serum-treated NIH 3T3 cells (data not shown).

Immunogold labeling of PGHS-I in murine NIH 3 T3 cells. Serum-starved or serum-
treated NIH 3T3 cells were fixed and processed for immunoelectron microscopy. The
expectation that PGHS-1 would be expressed at similar levels in both starved and treated
cells [49, 133] was confirmed by western blotting (not shown). Serum-treated (Fig. 11A)
or serum-starved (not shown) NIH 3T3 cells exhibited perinuclear and ER staining for
PGHS-1. PGHS-1 labeling was eliminated by preincubation of the anti-PGHS-l antibody
with its cognate peptide (Fi g. 118). The immunostaining experiments depicted in Fig. 11
were performed using an anti-PGHS-l antibody raised against a peptide corresponding to
amino acids Leu274-Arg288 of murine PGHS-1. Identical experiments performed with
another anti-PGHS-l antibody, this one raised against whole ovine PGHS-1 [126], resulted
in a pattern of staining indistinguishable from that seen with the anti-peptide antibody (not
shown). PGHS-1 staining by the antibody prepared against the whole protein was eliminated
by preadsorption of the antibody with purified ovine PGHS-1. Several experiments were

performed using a total of 24 different NIH 3T3 cells in which well-preserved sections

57

.81 :d u :0: 030m 82:28: 2802:005
.:0 ”0:20:: .:2 0805808 :020:: :0::0 0:0 :08: 0:: :003:0: 80:28.20
Tmmwm :0 020308: 0:: 2 000: 0:02:80 :0 0>::0::000::0: 0:20>:0 :020:: :0 0:02:80
00>:000::-=03 0:00:02 030::0 :0:0 .0000::0 AU: :: .0:0::808 :020:: 3 :08:
0:: :0 3:0: w:::0:00 0000:3080 0mg 0:0 0:0::808 :020:: A0: :0::0 0:: :0 022::0:
0:0w 0:00:00 0000:3080 =080 :23 0:20.80 :020:: 0 :0 :200M 6: 0000:3080
=080 a: 00:0:00 :0:2 Row :23 800::020 0:0 0:20:: 0:: :0 0:280: w230:0
0:00 mhm IE :0 0:02:00m Am J: .0:0m:.::00 20:63 2::0:.2::0 b00:0000 0 :23
::08:00:: a: 00323: 003 02:. Am: 000:0: 20:: 00 02 :23 00:0::020:: 30280
0800 0:: :0 AU .5 70mm: 0:.C:8 :0 03:24:83.: 00:0:000 0:: 0: :00280 000:0:
-280 2::0: 00:28: .3850 :0 :0:20 :23 020:0: 0:03 0:02:0000b0 2:00:23 .80:
0:: 2 00:20:00 00 3000828 :0::020 :0: 0000008: 0:0 000: .00:00::-8::00 0:03 0:00
20 82 0:8 2.0 82 8.5:: a: 0080: 0 802:0. 0.88.85: .: 8:00:

58

 

59

of the nuclear envelope were analyzed to determine the distribution of PGHS-1 between the
inner and outer nuclear membranes (Fig. l 1 C). Gold particles representing PGHS-l staining
along the NE were present in approximately equal abundance on the inner and outer nuclear
membranes (Table 1). Our results confirm that PGHS-l is a membrane protein of the ER and
nuclear envelope [28] and establish that PGHS-l is distributed equally between the
membranes of the nuclear envelope. Collectively, the data demonstrate that PGHS-l and -2
reside in the same subcellular membranes.

Western blotting of PGHS-1 and PGHS-2 in subcellular fractions. Earlier
immunofluorescence studies from our laboratory indicated that PGHS-1 and PGHS-2 are in
the ER and nuclear envelope but that the concentration of PGHS-2 in the nuclear envelope
is roughly twice that of the ER while the concentration of PGHS-1 is the same in the nuclear
envelope and the ER [67]. In contrast, our quantitative immunoelectron microscopy
indicated that PGHS-1 and -2 are present in the same subcellular locations in approximately
equal proportions. There are significant differences between the fixation and staining
protocols used for immunofluorescence and electron microscopy any one of which could
conceivably cause subtle differences in the patterns of staining. One difference involves the
use of detergents to permeabilize the cells for immunofluorescence staining. Accordingly,
we determined the effect of saponin, the detergent used in our earlier immunofluorescence
work [67], on the distribution of PGHS-1 and PGHS-2 in nuclear and ER fractions from NIH
3T3 cells. When nuclei were isolated in the presence of 1% saponin from murine NIH 3T3
cells, immunoreactive PGHS-2 was relatively more abundant in nuclear membranes than in

microsomal membranes while PGHS-l immunoreactivity was equally distributed between

60

nuclear and microsomal membranes in the same experiments (Fig. 12). Similar results were
obtained when nuclei were isolated in the presence of 0.2% saponin, although the difference
between PGHS-l and PGHS-2 distribution were not statistically significant. However, when
Western blotting experiments were conducted on subcellular fractions isolated from serum-
stimulated murine NIH 3T3 cells in the absence of detergent (Fig. 12), PGHS-1 and PGHS-2
were similarly distributed between nuclear and microsomal membranes. We measured the
ratios of immunoreactive PGHS-1 and -2 in nuclei versus microsomes using densitometry.
In three experiments in the presence of saponin, the ratios of nuclear to microsomal
immunoreactive PGHS-l and -2 were 1.1 +/- 0.1 and 1.9 +/- 0.2, respectively. In the
presence of 0.2% saponin, the same ratios for PGHS-1 and-2 were 1.14 +/- 0.1 and 1.26 +/-
0.1. Exclusion of detergent resulted in nuclear to microsomal immunoreactivity ratios for
PGHS-l and -2 of 1.08 +/- 0.1 and 1.10 +/- 0.2, respectively. These data suggest that
PGHS-1 and PGHS-2 can be differentially solubilized from nuclear membranes by high
concentrations of detergents such as saponin. This may account for the differential patterns
of PGHS-1 and PGHS-2 staining that have been observed in immunofluorescence studies

[67].

O\
—A

Microsomes
Microsomes

Nuclei
Cells

to
=

d.)
U

   
   

0.2% Saponin .

1 % Saponin

Ratio of nucleizmicrosomes

 

0.2% Saponin 1.14 +/- 0.06 1.26 +/- 0.06
1% Saponin 1.1 +/- 0.1 1.9 +/- 0.2
No Detergent 1.10 +/- 0.20 1.08 +/- 0.14

PGHS-1 PGHS-2

Figure 12. Western blot analysis of PGHS-1 and PGHS-2 in subcellular fractions of
murine NIH 3T3 cells. Whole cells, microsomes or nuclei were prepared from murine NIH
3T3 cells in the presence of 0.2% saponin, 1% saponin, or no detergent as described in the
text. A, After separation of the protein fractions by SDS-PAGE and transfer to nitrocellulose,
isozyme-specific antibodies were used as indicated to visualize PGHS-1 and PGHS-2
immunoreactivity in each membrane fraction. B, Densitometric analysis was used to
determine the ratio of immunoreactive PGHS-1 or PGHS-2 in nuclear versus microsomal
membranes. Ratios obtained using 1% saponin were significantly different from each other
according to Student’s t test.

62

Discussion
PGHS-1 and PGHS-2 have structurally homologous membrane binding domains

containing four amphipathic helices which anchor the proteins to one leaflet of the lipid
bilayer [28, 34, 35, 131, 134-136]. Previous work has established that both isozymes are
located on the lumenal surfaces of the ER and nuclear envelope [28, 42]. However,
immunocytofluorescence studies suggested that PGHS-2 was more concentrated in the
nuclear envelope than the ER whereas PGHS-l was equally distributed in both compartments
[67]. One possibility to account for these latter findings was that PGHS-2 was uniquely
localized on the inner membrane of the nuclear envelope. The major aim of the present
studies was to examine this possibility using immunoelectron microscopy. The issue of
PGHS-1 versus PGHS-2 localization is of particular interest because of the potential
relationship between PGHS-Z-derived products generated at the nuclear envelope and nuclear
signaling associated with cell replication or differentiation.

The major finding of our EM study is that in NIH 3T3 cells both PGHS-1 and
PGHS-2 are present in equal proportions on both the inner and outer membranes of the
nuclear envelope. Human monocytes and umbilical vein endothelial cells stained for PGHS-
l or PGHS-2 exhibited identical patterns. Our results make it clear that both isozymes are
present in the same subcellular compartments and at comparable concentrations. Of course,
we cannot rule out the possibility that PGHS-1 and PGHS-2 are associated with different
microdomains within these compartments.

Earlier results from immunofluorescence studies which employed low concentrations
of saponin to permeabilize cell membranes and had suggested that PGHS-2 is preferentially

localized to the nuclear envelope [67] can be explained on the basis of differential

63

solubilization of the two proteins from the nuclear envelope. When ER and nuclear
membranes were prepared in the presence of 1% saponin, PGHS-2 was more concentrated
in the nuclear fraction while PGHS-1 was found in equal abundance in the ER and nuclear
membranes; in the absence of saponin, PGHS-2 was present at the same concentration in the
ER and nuclear envelope. Presumably these differences in affinities of the two proteins for
the nuclear envelope are a result of the significant differences in the amino acid composition
of the membrane binding domains of the two isozymes [1, 43].

A model has emerged suggesting that PGHS-1 and PGHS-2 act independently and
that, at least in part, the inducible isozyme, PGHS-2, provides prostaglandins for a nuclear
eicosanoid signaling system [1, 104, 137]. While this model may still be correct, our results
imply that any specific connection between PGHS-2 and the generation of products which
function in the nucleus would have to result from differences in the expression of the
activities of PGHS-l and PGHS-2 and not from gross differences in the subcellular
distributions of PGHS-2 versus PGHS-1 . That is, because both PGHS-1 and PGHS-2 appear
to be present in the same membranes, factors other than compartmentation account for the
separation of their activities into two independent systems. The most likely factors are
differences in interactions with different phospholipases and/or differences in enzyme
kinetics. Arm, Austen and Herschman and their colleagues [78, 79, 138-140] have
demonstrated that two separate phases of PGD;2 synthesis in mast cells are independently
coupled to PGHS-l (early phase) and PGHS-2 (late phase) by different phospholipases A2.
Kinetic mechanisms for separating the actions of PGHS-1 and PGHS-2 have also been

described. For example, PGHS-2 has a significantly lower threshold for hydroperoxide

64

activation than PGHS-1 thereby enabling PGHS-2 to oxygenate arachidonic acid in the
presence of lower peroxide concentrations [141, 142]. In addition, negative allosteric
regulation of PGHS-1 by arachidonic acid, at concentrations between 0.5 nM and 1 uM, has
the overall effect of causing a 2-4 fold greater rate of PGHS-Z-mediated prostanoid formation
[143]. These kinetic differences between PGHS-1 and PGHS-2 have been identified with
purified or partially purified enzyme preparations. However, it may be possible to test for
kinetic differences between the two isozymes in intact cells using histochemical assays of
enzyme activity [67].

Another more speculative possibility to account for the independent operation of
PGHS-1 and PGHS-2 in cells where both isozymes are expressed is the existence of
accessory proteins which differentially affect the rate of prostaglandin endoperoxide
formation by PGHS-l versus PGHS-2. Although no such protein(s) has been identified in
the prostanoid biosynthetic system, there is a precedent for an accessory protein in the
leukotriene pathway. Leukotrienes synthesized through S-lipoxygenase arise from
arachidonic acid apparently delivered to the 5-lipoxygenase by an activating protein, FLAP
[85, 143-147].

The observation that both PGHS-1 and PGHS-2 are located on the inner nuclear
membrane is of interest from a cell biology perspective. To our knowledge, no other
endogenous integral membrane proteins of the ER have been demonstrated to be present on
the inner nuclear membrane. FLAP, another integral membrane protein involved in
eicosanoid biosynthesis has been localized to the inner and outer membranes of the NE [69,

70]. However, in contrast to PGHS-l and -2, FLAP is predominantly localized to the nuclear

65

envelope; the orientation of FLAP in the membrane is not well characterized. One model
of how integral membrane proteins synthesized in the ER reach the inner nuclear membrane
involves lateral diffusion through the membrane bilayers of the NE [69, 70, 1 14, l 15, 148].
According to this model, membrane proteins are subject to a size constraint imposed by the
lateral channel diameter of the nuclear pore complex, which serves to separate the inner and
outer nuclear membranes. Both PGHS-1 and PGHS-2 are targeted initially to the ER by a
KDEL-like C-terminal targeting signal [149]. We propose that both isozymes are then able
to bypass the nuclear pore complex and reach the inner nuclear membrane by lateral
diffusion. Our reasoning is based on the nature of their interaction of PGHS-1 and PGHS-2
with membranes (i.e. via a monotopic membrane binding domain) and the fact that the
proteins are on the lumenal surface of the membrane (Fig. 13). Consistent with this concept
are studies with the lamin B receptor (LBR), an integral membrane protein that is targeted
exclusively to the inner nuclear membrane via specific targeting signals [114]. When the
nucleoplasmic/cytoplasmic-oriented extramembrane domain of LBR is artificially enlarged,
the protein is retained in the ER (and outer nuclear membrane). Other ER membrane
proteins with large extramembrane domains oriented towards the cytoplasm are also thought
to be restricted to the outer membrane of the nuclear envelope and the ER [1 14, 150]. P4508
such as thromboxane synthase and prostacyclin synthase are integral membrane proteins of
the ER and have relatively large cytoplasmic domains. Thus, these proteins are not likely to
be present on the nucleoplasmic face of the inner nuclear membrane and would be unable to

metabolize efficiently PGHZ generated by PGHS-1 or PGHS-2 present on this membrane.

66

We conclude that PGHS-l and PGHS-2 are present in similar proportions on the
endoplasmic reticulum, outer nuclear membrane, and inner nuclear membrane of NIH 3T3
cells, human monocytes, and HUVECs. Fatty acid substrates for PGHSs appear to be
supplied via both an sPLA2 that functions on the phospholipids on the cell surface and a
cPLA2 that undergoes a Ca21-dependent translocation to the cytosolic face of the ER and
outer membrane of the nuclear envelope [59, 61, 65, 151, 152] and perhaps the inner
membrane of the nuclear envelope [66]. However, the issue of whether there is preferential
coupling of different phospholipases A2 to PGHS-l versus PGHS-2 is currently unresolved.
PGH2 formed through both PGHS-l and PGHS-2 can apparently diffuse readily through
membranes [153]. That PGH2 which diffuses from the ER lumen into the cytoplasm is likely
to be metabolized by enzymes such as TxA synthase or PGIZ synthase located on the
cytoplasmic surface of the ER and nuclear envelope [154]. The fate of the PGH2 which

diffuses into the nucleoplasm is presently unknown.

67

 

W = lamin B receptor
= PGHS-l or PGHS-2

 

@ = nuclear pore complex

Figure 13. Diffusion of integral membrane proteins in the nuclear envelope.
Membrane topography of integral membrane proteins affects their ability to migrate through
the nuclear envelope [114]. The inner and outer nuclear membranes are biochemically and
functionally distinct, and are physically separated by the nuclear pore membranes and pore
complexes (N PC). Lumenally-oriented proteins like PGHS-1 and -2 are able to reach the
inner nuclear membrane presumably by passive diffusion. Likewise, the
cytoplasmically-oriented lamin B receptor, which is found exclusively on the inner nuclear
membrane, has a small extramembrane domain that is not held up at the NPC. One can
speculate that membrane proteins with large cytoplasmic domains will be physically
prevented from reaching the inner nuclear membrane due to the barrier of the NPC.

68

Acknowledgment
This work was previously published in the Journal of Biological Chemistry as
Spencer, AG, Woods, JW, Singer, 11, Arakawa, T, and Smith WL. (1998) “Subcellular
localization of prostaglandin endoperoxide H synthases-1 and -2 by immunoelectron

microscopy” J. Biol. Chem. 273, 9886-9893., and is reprinted with permission.

CHAPTER III
THE MEMBRANE ASSOCIATION OF PGHS-1 AND PGHS-2
Introduction

PGHS-1 and PGHS-2 are integral membrane proteins by definition since detergents,
and not just chaotropic salts, are required to disrupt their association with cellular membranes
[1 8]. X-ray crystallographic studies have demonstrated that neither isozyme contains
structural motifs that are well-known to confer membrane association to integral membrane
proteins (e.g., transmembrane domains or prenylation sites). Instead, PGHS-1 and -2 are
thought to associate with cellular membranes through a novel, monotopic membrane binding

domain that traverses a single leaflet of the lipid bilayer.
Blobel speculated in 1980 on the existence of monotopic integral membrane proteins
[155], but the first structural evidence for such a protein came from the X-ray
crystallographic description of ovine PGHS-l by Picot, et al. [34]. The tertiary structure of
PGHS-1 is predominantly globular. Projecting from the globular catalytic domain in each
monomer are four amphipathic alpha helices (labeled A, B, C, and D in Fig. 6).
Hydrophobic and aromatic residues protrude from these helices and away fi'om the
hydrophilic surface of the catalytic domain to create a hydrophobic patch. These helices also
form the opening to the cyclooxygenase active site and were predicted by Picot et al. to
facilitate interaction with the lipid bilayer. It is important to our understanding of PGHS
biology to know the mechanism by which PGHS-l and -2 interact with cellular membranes.

The substrate of PGHSs, arachidonic acid, is released from the bilayer and presumably

69

70

travels through the core of helices A-D en route to the cyclooxygenase active site. As
discussed in Chapter II, the topology of membrane proteins plays a significant role in their
subcellular localization. In addition, a thorough description of the membrane binding
domains of PGHS-1 and -2 may provide new information regarding a recently-discovered
membrane association strategy that may be used by other lipid biosynthetic enzymes.

We have aimed to characterize biochemically the regions of PGHS -1 and PGHS-2
that interact with cellular membranes using 3-trifluoro-3-(m-[‘251]iodophenyl)diazirine ([1251]-
TID). [‘251]-TID has been used as a tool to selectively label the hydrophobic regions of
integral membrane proteins [156, 157]. The compound is hydrophobic, photoactivatable,
and partitions into lipid bilayers where, following activation with UV light, it cross-links
to membrane-associated regions of proteins (Fig. 14). In 1996, Otto and Smith used [”51]-
TID to show that an N-terminal region ofovine PGHS-l associates with membranes between
amino acids 25-166 [135]. Here, we extend this work and provide similar evidence in
PGHS-2 that the putative membrane binding domain predicted by Picot et al. is indeed

associated with membranes in the endoplasmic reticulum.

 

 

 

125] 1251 [251
._T.W .@.
I
N2
CF3 — C

 

 

 

 

 

 

 

 

 

 

CF3 _ C— CF; — C:
I f
N Cabana
[‘”I]-TID ['”I]-TID/protein linkage
PGHS-1b:
H
1111

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Incubate withl '“I - TID (0)

 

Danaturationland proteolysis

[_:1 g J [ ] l 1
m 3'2'2'9

 

 

Figure 14. Mechanism of [‘“Il-TID cross linking to membrane proteins. [‘zsfl-TID is
a hydrophobic, photoactivatable compound that can be used to label the membrane-
associated regions of proteins. Due to its extreme hydrophobicity, [mI]-TID labels only
hydrophobic regions of membrane proteins and does not label regions of proteins exposed
to an aqueous environment [156]. Top: Photoactivated release of the diazirine group of
[‘25I]-TID as dinitro gen leads to the formation of a highly reactive carbene that
nonspecifically cross-links to constituents of cellular membranes, including both proteins and
lipids. This chemistry allows one to covalently label the membrane-associated region of a
protein with a radioactive marker. Bottom: Schematic diagram of a [1251]-TID labeling
experiment.

72

Materials and Methods

Preparation of PGHS-I from ovine seminal vesicles. Microsomal ovine PGHS-
1(oPGHS-l) was obtained from sheep seminal vesicles essentially as described previously
[135]. Briefly, 5 g of seminal vesicles were sliced into 1 mm strips with a razor blade and
suspended in 30 m1 HEPES buffer (20 mM HEPES, 20 mM glutamic acid, 2 mm magnesium
acetate, 200 mM sucrose, pH 7.5). After three 30 second blasts with a Polytron
homogenizer, heavy tissue debris was pelleted by centrifugation at 10,000 x g. Microsomal
membranes in the supernatant of the low speed spin were collected by ultracentrifugation at
200,000 x g in a Beckman SW50.1 rotor. The microsomal pellet was resuspended in 2 ml
HEPES buffer, dounce homogenized, cleared by centrifugation at 10,000 x g, and assayed
for protein using the BioRad Bradford protein assay reagent.

Preparation of PGHS-2 from baculovirus-infected Sf21 cells. Sf21 insect cells were
grown in 500 m1 spinner flasks or a 12 L bioreactor in Sigma TC-IOO medium containing
10% fetal calf serum. When the cell density reached 1.8 x 106 cells/ml, cells were infected
with baculovirus containing a gene for human PGHS-2 (hPGHS-Z) that had been engineered
to contain a hexa-Histidine tag at its N-terminus. Infection was allowed to proceed for three
days. Approximately 9.0 x 108 Sf21 cells were harvested by centrifiigation and resuspended
in 10 ml 0.1 M Tris, pH 7.4. Cells were then disrupted by vigorous sonication using a
microtip attachment to a Misonix Sonicator. After a low speed (10,000 x g) spin to pellet
cell debris, microsomal membranes were collected by centrifugation at 200,000 x g in a
Beckman SW50.1 swinging bucket rotor. The microsomal pellet was resuspended in 2 ml

0. l M Tris, pH 7.4, dounce homogenized, and assayed for protein using the BioRad Bradford

73

protein assay reagent.

[’25I]- TID labeling of PGHS-1 and -2. This procedure was carried out at 4°C unless
otherwise noted. Microsomal preparations of oPGHS-l or hPGHS-2 were prepared as
described above and resuspended to a final protein concentration of 15-20 mg/ml in the
appropriate buffer. Aliquots (400 ul) were preincubated in the presence of 80 [4M
flurbiprofen to block the ability of [‘ZSI]-TID to label the PGHS active site. In some
experiments, 200 mM reduced glutathione was included to scavenge ['251]-TID exposed to
the aqueous phase. [‘251]-TID was added to a final concentration of approximately 5 uM.
After incubation at 4°C for 10 min., resuspended microsomes were transferred to a 1 cm
diameter plastic culture dish and irradiated for 10 min. with a UV illuminator (366 nm) held
at a distance of 5 cm. Labeled membrane proteins were then solubilized from cellular
membranes using the detergent CmE6 (Anatrace, Inc.) at a concentration of 1% (v/v) for 1
hour with gentle shaking. oPGHS-l or hPGHS-2 were purified from the solubilized
membrane protein fraction using immunoprecipitation or Ni-Agarose, respectively.

Immunoprecipitation of PGHS-1 from solubilized microsomes. Following treatment
with ['251]-TID, solubilized microsomal membrane proteins were transferred to a new tube.
A monoclonal antibody directed against oPGHS-l (50 pl, concentration: 0.1 mg/ml) [158]
was added to the solubilized proteins and the sample allowed to incubate with shaking for
30 min at 4°C. PGHS-1 was precipitated by the addition of 100 pl of Protein A-Sepharose
(50% slurry in 0.1 M Tris/0.1% Tween 20/pH 8.0) and incubation for 30 min. at 4°C. The
immunocomplex was pelleted at 500 x g in a tabletop microcentrifuge and washed three

times with 0.1 M Tris/0. 1% Tween 20/pH 8.0. To elute immunoprecipitated oPGHS-l, the

74

immunocomplex was resuspended in 0.5% recrystallized SDS and boiled for 5 min with
periodic vortexing. The Protein A-Sepharose was then pelleted at high speed, and the
supernatant was removed. The supernatant containing the immunopurified, [USU-TID
labeled oPGHS-l was then (a) subjected to immediate proteolysis or (b) denatured, reduced,
and alkylated prior to proteolysis. Immediate proteolysis required the dilution of 0.5% SDS
to 0.05% SDS by diluting lO-fold in an appropriate buffer and reconcentrating the [‘251]-TID
labeled protein using a Millipore UltraFree concentrator. Denaturation, reduction, and
alkylation were performed as described below.

Purification of His-tagged PGHS-2 by nickel chromatography. Following labeling
with [‘251]-TID, solubilized microsomal hPGHS-2 was placed in a 15 m1 tube and diluted
threefold with Buffer A (10 mM phosphate/100 mM NaC1/20 mM imidazole/pH 7.2). Ni-
NTA (Qiagen; 1 ml) was added and the tube was incubated at 4°C with gentle shaking for
1 h. Ni-NTA was collected by centrifugation at 1200 rpm in a Beckman table top centrifuge
for 2 min and the resin was washed three times for 5 min in wash buffer (10 mM
phosphate/500 mM NaCl/30 mM imidazole/O.l% CloEg/pH 7.2). Photoaffinity-labeled
hPGHS-2 was eluted fiom the Ni-NTA by incubation in elution buffer (10 mM
phosphate/200 mM NaCl/200 mM imidazole/pH 7.2) for 1 hour at 4 °C. Eluted protein was
concentrated using a Millipore Ultra-Free concentrator to a volume of approximately 400
pl. Elution salts were removed by dialysis against 10 mM Tris, pH 8.2 prior to denaturation,
reduction, and alkylation.

Denaturation, alkylation, reduction, and proteolysis of oPGHS-I and hPGHS-Z.

Prior to denaturation, ['251]-TID labeled protein samples were concentrated to a volume of

75

25 ul. Samples were denatured by incubation in 6M (final concentration) guanidium
hydrochloride at 55 0C for three hours. Freshly prepared 1M dithiothreitol was then added
to a final concentration of 10 mM. Afier covering with argon, the sample was incubated at
55 °C for 2 h. Reduced disulfide bridges were prevented from reforming by the addition of
freshly prepared N-isopropyl-iodoacetic acid (Molecular Probes, Inc.) and incubation at room
temperature in the dark for 1 h. After the alkylation step, samples were dialyzed against 4
L of either 0.05 M NH4HCO3/pH 7.9 ( for Glu C digestion) or 0.2 M Tris/pH 8.0 (for Lys
C digestion) for 12-15 hours in a Fisher 0.5 ml Slide-A-Lyzer (MWCO = 10,000 Da). Prior
to proteolysis, samples were concentrated in a SpeedVac to a volume of approximately 50
ul. Proteolytic enzymes endproteinase Lys C or endoproteinase Glu C were added such that
the final ratio of protease to PGHS- l /-2 was approximately 1 :50. Di gestions were performed
at 37 °C for 20-24 h, with additional endoproteinase Glu C being added every 5 h in the case
of hPGHS-2. Proteolytic fiagments were analyzed by SDS-PAGE as described below.
SDS-PA GE and analysis of proteolytic products. The following precautions were
employed to prevent the N-terminal blockage of peptides separated by SDS-PAGE: (a)
highly pure and fresh acrylamide reagents and buffers were used to prepare 16% slab gels,
(b) the gels were allowed to polymerize overnight at room temperature before use, and (c)
recrystallized SDS was used in all buffers and gels. Proteolytic peptides or untreated
controls were separated on the slab gels and transferred to PVDF membranes for Western
analysis or autoradiography. In some experiments, portions of gels were silver-stained to
observe total protein staining using the BioRad Silver Stain kit. Autoradiography was

performed by (a) sandwiching the PVDF membrane between two intensifying screens and

76

exposing to Amersham Hyperfilm-MP for 4 days at -80°C or (b) exposing the PVDF
membrane to a phosphor screen (Molecular Dynamics, Inc.) for two days. In the former
case, Stratagene Glogos II fluorescent markers were used to orient the film to the PVDF
membrane prior to excision of [‘251]-TID labeled peptides. Immunoblot visualization of
PGHS-1 or -2 and their proteolytic products was done as described previously. Antibodies
against hPGHS-2 were directed against a C-tenninal l8-arnino acid cassette or amino acids
20-40 at the N-terminus (Santa Cruz Biotechnology, Inc.). An antibody recognizing the N
terminus of oPGHS-l was directed against amino acids 25-35.

Secondary digestion of the PGHS-1 Lys C fragment with Glu C. Using the
autoradiograph as a template, the [ '251]-TID-labeled oPGHS-l Lys C fragments were excised
from the PVDF membrane using a new, clean razor blade. The excised band was cut into
1 mm2 squares and placed in a 1.5 ml screw top microcentrifuge tube that had been
prewashed once with methanol and twice with water. Glu C digestion buffer (50 pl of 100
mM Tris-Cl/l% reduced TritonX-lOO/pH 8.0 ) was added to the squares. The tube was
vortexed for 20 s and allowed to incubate at room temperature for 30 min. Endoproteinase
Glu C was added to an approximate proteasezprotein ratio of 1:25. The digestion was
performed at 37°C for 24 hours. To recover the proteolytic peptides from the PVDF
membrane, sample tubes were vortexed for 10 seconds and placed in a water bath sonicator
for 5 min. The tubes were centrifuged for 2 min at 4000 rpm in a Beckrnan microcentrifuge.
The resulting supernatant was transferred to a new, clean microcentrifuge tube. An
additional 50 pl aliquot of digestion buffer was added to the squares and subjected to

vortexing, sonication, centrifugation, and the supernatant was removed to the new tube. This

77

was repeated a third time. The sample containing the eluted peptides was concentrated in
a SpeedVac to a volume of approximately 50 pl prior to analysis by SDS-PAGE.
Microsequencing of radiolabeled proteolytic peptides. Using the autoradiograph as
a template, [‘251]-TID labeled peptides were excised from PVDF membranes and cut into
1mm2 squares. Amino terminal amino acid sequence analysis was performed in
collaboration with Dr. Joe Leykam of the Michigan State University Macromolecular
Structure Facility using automated Edman degradation in an Applied Biosystem 477A gas
phase amino acid sequencer. Phenylthiohydantoin amino acid derivatives were identified

by HPLC.

78

Results

Photolabeling of the PGHS-1 membrane binding domain. Ovine seminal vesicle microsomes
containing oPGHS-l were prepared, preincubated with 80 mM flurbipro fen, and labeled with
the hydrophobic, photolabile reagent [‘251]-TID. Flurbiprofen prevents the labeling of the
oPGHS-l active site by [‘35I]-TID [135]. As expected, oPGHS-l was the major membrane
protein in this tissue (Fig. 16). Immunoprecipitation of oPGHS-l followed by treatment with
endoproteinase LysC resulted in an N-terminal fragment with an observed molecular mass of
20 kDa (Fig. 17). This band had been previously observed by Otto and Smith [135] and
represents a glycosylated form of the predicted, 16 kDa Lys C product (Fig. 15). The 20 kDa
band contains amino acids 25-166 and harbors the predicted membrane binding domain.
Similar photolabeling patterns were observed in the presence and absence of glutathione,
demonstrating that ['251]-TID labeled only those regions of oPGHS-l located within the lipid
bilayer. Several attempts were made to further digest the 20 kDa band in order to generate
a minimum fragment containing the putative membrane binding domain; however, we were
never able to obtain enough yield to facilitate sequencing.

The primary structure of oPGHS-l predicts endoproteinase Glu C products that
include a 7.4 kDa peptide containing amino acids 74-140 (Fig. 15). This region includes the
predicted membrane binding domain. To determine the regions(s) of oPGHS-l that contained
the [‘251]-TID label, photoaffinity—labled PGHS-1 was denatured in guanidium hydrochloride
and subjected to exhaustive proteolysis with endoproteinase Glu C. Separation of the Glu C
proteolysis products by SDS-PAGE, transfer to PVDF membranes, and subsequent

autoradiography revealed a 7.4 kDa, ['25I]-TID labeled peptide (Fig. 18).

79

2‘5 MBD "[36 PGHS -1 600

var—Illllll l

Endoproteinase Lys C, SDS-RAGE

25 166 74 140
.‘ MBD , g I MBD ]
Endoproteinase Glu C,
SDS-PAGE

Figure 15. Predicted proteolysis products of oPGHS-l with Lys C and Glu C. The
predicted membrane binding domain is symbolized by a black rectangle and abbreviated

MBD. The grey rectangle represents the binding site for the peptide directed antibody against
residues 25-35 of oPGHS-l.

80

72 kDa- -

 

 

Lipid -

 

 

Figure 16. Pull-TID labeling of ovine seminal vesicle membrane proteins. Seminal
vesicle microsomes were labeled with [mu-TID. One sample was subjected to
immunoprecipitation with a monoclonal antibody against PGHS-1. Proteins were separated
by SDS polyacrylamide gel electrophoresis, transferred to PVDF membranes, and [mm-TID
labeling was detected by autoradiography.

81

 

207 kDa—
176 -

80' w -
42.4 —

32. 6 :

18‘

7. 2-

 

 

 

 

 

 

 

 

Figure 17. [‘“Il-TID labeling of an N-terminal peptide of oPGHS-l.
Alter photolabeling ovine seminal vesicles with [‘”I]-TID, oPGHS-l was
immunoprecipitated from solubilized membrane proteins. Samples were
incubated in the presence (right lane) or absence (left lane) of endoproteinase
Lys C for 20 h at 37°C. Proteolytic peptides were separated by SDS
polyacrylamide gel electrophoresis, transferred to PVDF membranes, and
exposed to a phosphor screen to detect [‘”I]-TID—1abeled peptides. A and B,
Autoradio graphs of untreated and Lys C treated oPGHS-l labeled with [‘51]-
TID in the presence and absence of glutathione, respectively. C, Untreated
and Lys C treated oPGHS-l visualized by immunoblot using an antibody
against amino acids 25-35 [42].

82

 

72kDo —

7.4 kDa—

 

 

 

 

 

 

 

oPGHS-I

Figure 18. An 7 .4 kDa peptide of oPGHS-l containing the predicted membrane
binding domain is labeled by [lull-TID. [WU-TID labeled oPGHS-l samples were
incubated in the presence or absence of endoproteinase Glu C for 20 h at 37 °C. Proteins were
separated by SDS polyacrylamide gel electrophoresis and transferred to PVDF membranes.
Right panel, Immunoblot of separated proteins with an antibody directed against amino acids
20-40 of hPGHS-2. Note: endoproteinase Glu C reacts nonspecifically with the secondary
antibody used in the immunoblot. Left panel, Autoradiograph of the same PVDF membrane
used for the immunoblot.

10 20
MSRQSISLRF

60
GLDRYQCDCT

110
DFVNATFIRD

160
ILPSVPRDCP

210
FAQHFTHQFF

260
KLKYQMLNGE

310
YATIWLREHN

360
QLSGYFLQLK

410
DYSYEQFLFN

460
VIKESRVLRL

510
LEFYPGLLLE

560
GGEVGFNLVK

Figure 19. Primary structure of oPGHS-l. The predicted membrane binding domain
predicted by Picot, et al. (34), residues 74-117, is depicted in red typeface. The region of
PGHS-1 that corresponds to the 7.4 kDa, [mm-TID labeled fragment from the
autoradiograph in Fig. 17 is underlined. Italicized residues indicate the region of the 7.4 kDa

30

PLLLLLLSPS

70
RTGYSGPNCT

120
TLMRLVLTVR

170
TPMGTKGKKQ

220
KTSGKMGPGF

270
VYPPSVEEAP

320
RVCDLLKAEH

370
FDPELLFGAQ

420
TSMLVDYGVE

470
QPFNEYRKRF

520
KCHPNSIFGE

570
TATLKKLVCL

83

40
PVFSADPGAP

80
IPEIWTWLRT

50
APVNPCCYYP

90
TLRPSPSFIH

CQHQGICVRF

100
FMLTHGRWLW

 

130
SNLIPSPPTY

180
LPDAEFLSRR

230
TKALGHGVDL

280
VLMHYPRGIP

330
PTWGDEQLFQ

380
FQYRNRIAME

430
ALVDAFSRQP

480
GMKPYTSFQE

530
SMIEMGAPFS

580
NTKTCPYVSF

140
NIAHDYISWE

190
FLLRRKFIPD

240
GHIYGDNLER

290
PQSQMAVGQE

340
TARLILIGET

390
FNQLYHWHPL

440
AGRIGGGRNI

490
LTGEKEMAAE

540
LKGLLGNPIC

590
HVPDPRQEDR

peptide that was confirmed by Edman degradation.

150
SFSNVSYYTR

200
PQGTNLMFAF

250
QYQLRLFKDG

300
VFGLLPGLML

350
IKIVIEEYVQ

400
MPDSFRVGPQ

450
DHHILHVAVD

500
LEELYGDIDA

550
SPEYWKASTF

600
PGVERPPTEL

84

2.318036 PGH3'2 Si"

Endoproteinase Glu C, SDS-PAGE

59 126
I MBD I

Figure 20. Predicted major proteolysis product of hPGHS-Z with Glu C. The predicted
membrane binding domain is symbolized by a black rectangle and abbreviated MBD. The
grey rectangle represents the binding site for the anti-peptide antibody directed against amino
acids 20-40 of hPGHS-Z.

 

85

This observation suggested that amino acids 74-140 of oPGHS-l are associated with the ER
in seminal vesicle gland microsomes. Small amounts of incompletely digested PGHS-1
would be predicted to contain the N-terminus of the protein and were recognized by an
antibody directed against amino acids 25-35 (Fig. 18).

Peptide mapping of the membrane binding domain of hPGHS—Z with [ ’ 25I]- T ID.
hPGHS-2 was engineered to contain a six-histidine tag and expressed in Sf21 insect cells

using a baculovirus expression system. The Km, V and subcellular localization of His-

tagged PGHS-2 are indistinguishable from those of the native enzyme'. To determine which
regions of PGHS-2 were associated with the endoplasmic reticulum, microsomal membranes
from Sf21 cells were prepared and photoaifinity labeled with [WU-TID. ['251]-TID-labeled
hPGHS-2 was then solubilized and purified by nickel affinity chromatography. Analysis of
the purified protein by SDS-PAGE revealed the expected [mm-TID labeled doublet at 72 and
74 kDa that was recognized by an antibody directed against the N-terminus of PGHS-2 (data
not shown).

The primary structure of PGHS-2 predicts endoproteinase Glu C proteolytic products
that include an 8 kDa fragment containing amino acids 59-126 (Fig. 20). This region
includes the predicted membrane binding domain. To determine the region(s) of PGHS-2
that contained the [‘2SI]-TID label, photoaffinity-labeled hPGHS-2 was denatured in
guanidium hydrochloride and subjected to exhaustive proteolysis with endoproteinase Glu

C. Separation of the Glu C proteolysis products by SDS-PAGE and subsequent

autoradiography revealed an 8 kDa fragment (Fig. 21). This observation suggests that

 

lT.Smith, D. DeWitt, unpublished results.

 

 

5’55: ¢ PGHS-2
42.8— A

H {Glu C protease
32.6- ‘

17.6-
4—15 kDa peptide

7.5- f y. +8 kDa peptide

 

 

 

Immunoblot Autorad

 

Figure 21. An 8 kDa peptide of hPGHS-Z containing the predicted membrane binding
domain is labeled by [‘Z’II-TID. [‘”I]-T ID labeled hPGHS-Z samples were incubated in
the presence or absence of endoproteinase Glu C for 20 h at 37°C. Proteins were separated
by SDS polyacrylamide gel electrophoresis and transferred to PVDF membranes. A,
Immunoblot of separated proteins with an antibody directed against amino acids 20-40 of
hPGHS-Z. Note: endoproteinase Glu C reacts nonspecifically with the antibodies used in the
immunoblot. B, Autoradio graph of the same PVDF membrane used for the immunoblot.

87

amino acids 59-126 of PGHS-2 are associated with the ER membrane in microsomal
preparations. The 8 kDa fragment was not recognized by an antibody directed against amino
acids 20-40 of hPGHS-Z. Incompletely digested hPGHS-2 peptides (e. g., at 15 kDa) would
be predicted to contain the N-terminus of the proteins and were recognized by this antibody
(Fig. 21).

To confirm the identity of the 8 kDa photoaffinity labeled fragment from Fig. 21, the
radioactive peptides were excised from the PVDF membrane and subjected to 15 rounds of
Edman degradation to determine their N-terminal sequence. The N—terminal sequence of this
peptide was FLTRIKLFLKPTPNT, which corresponds to amino acids 59-73 of hPGHS-2.
As shown in Fig. 22, the 8 kDa band corresponds to amino acids 59-126 in PGHS-2. Thus,
amino acids 59-126 of PGHS-2 contain the membrane binding domain. This domain of
PGHS-2 corresponds to the four amphipathic helices of PGHS-1 that was predicted by Picot

et al. to serve as the membrane association domain [34].

10
MLARALLLCA
70
TRIKLFLKPT
130
GYKSWEAFSN
190
NMMFAFEAQH
250
QIIDGEMYPP
310
VLKQEHPEWG
370
NRIAAEFNTL
430
AGGRNVPPAV
490
YGDIDAVELY
550

GFQIINTASI

STEL

Figure 22. Primary structure of hPGHS-z. The predicted membrane binding domain
predicted by Picot, et a1. (34), residues 58-108, is depicted in red typeface. The region of
PGHS-2 that corresponds to the 8 kDa, [‘ZSU-TID labeled fragment from the autoradiograph
in Fig. 19 is underlined. Italicized residues indicate the region of the 8 kDa peptide that was

20
VLALSHTANP
80
PNTVHYILTH
140
LSYYTRALPP
200
FTHQFFKTDH
260
TVKDTQAEMI
320
DEQLFQTSRL
380
YHWHPLLPDT
440
QKVSQASTDQ
500
PALLVEKPRP
560

QSLICNNVKG

30
CCSHPCQNRG
90
FKGFWNVVNN
150
VPDDCPTPLG
210
KRGPAFTNGL
270
YPPQVPEHLR
330
ILIGETIKIV
390
FQIHDQKYNY
450
SRQMKYQSFN
510
DAIFGETMVE
570

CPFTSFSVPD

confirmed by Edman degradation.

88

40
VCMSVGFDQY

100

160
VKGKKQLPDS
220
GHGVDLNHIY
280
FAVGQEVFGL
340
IEDYVQHLSG
400
QQFIYNNSIL
460
EYRKRFMLKP
520
VGAPFSLKGL
580

PELIKTVTIN

50
KCDCTRTGFY
110
YVLTSRSHLI
170
NEIVEKLLLR
230
GETLARQRKL
290
VPGLMMYATI
350
YHFKLKFDPE
410
LEHGITQFVE
470
YESFEELTGE
530
MGNVICSPAY
590

ASSSRSGLDD

60
GENcsrpsgg
120
DSPPTYNADY
180
RKFIPDPQGS
240
RLFKDGKMKY
300
WLREHNRVCD
360
LLFNKQFQYQ
420
SFTRQIAGRV
480
KEMSAELEAL
540
WKPSTFGGEV
600

INPTVLLKER

89

Discussion

PGHS-1 and PGHS-2 are integral membranes of the endoplasmic reticulum and
nuclear envelope. The major aim of this study was to determine which regions of these
proteins interact with the membrane bilayer. Determination of the membrane associated
regions of PGHS-l and -2 is of interest because the enzymes do not have classical
transmembrane domains. Instead, these proteins were hypothesized by Picot, et al. to
associate with the ER through four amphipathic alpha helices [34]. We have used the
hydrophobic, photoactivatable compound [ 125I]-TID to show that the membrane binding
domains of oPGHS-l and hPGHS-2 reside within amino acids 74-140 and 59-126,
respectively. The present study extends previous work from our laboratory demonstrating
that the membrane binding domain of oPGHS-l is between amino acids 25-166. In
addition, these studies provide the first biochemical identification of the membrane
association domain of hPGHS-2. Our results indicate that both isozymes interact with the
ER through the four-helix region predicted by X-ray crystallographic analysis of ovine
PGHS-1 (Fig 6).

Recent crystallographic analysis of a bacterial squalene cyclase has suggested that
membrane proteins other than PGHS-1 and -2 may interact with cellular membranes in a
similar fashion [40]. The overall protein folds of PGHS-l and bacterial squalene cyclase are
quite different. However, the overall shapes of the two proteins are similar. Both proteins
are homodimeric and possess hydrophobic patches that protrude away from the hydrophilic
catalytic domains.

It is interesting that the only two known protein structures with apparent monotopic

membrane domains are lipid biosynthetic enzymes. Monotopic membrane binding domains

90

like those of the PGHSs and squalene cyclase appear to be positioned in such a way to allow
direct substrate access to the active site. This would be thermodynamically favorable since
hydrophobic substrates like arachidonate and squalene are not likely to be released directly
into an aqueous environment. Thus, the membrane binding domains of PGHS and squalene
cyclase appear to function as (a) membrane anchors, and (b) continuous hydrophobic tunnels
from the site of substrate release to the active site. It is conceivable that other enzymes
whose substrate comes directly from the lipid bilayer may use a similar membrane
association strategy.

We have shown previously that PGHS-l and -2 are both present in the ER and the
inner and outer membranes of the nuclear envelope in similar proportions [68]. This work
suggested that the enzymes are not physically separated by compartmentalization in different
cellular membranes, but did not rule out the possibility that PGHS-1 and PGHS-2 are
associated with different phospholipid microdomains within the ER and nuclear envelope.
The catalytic domains of PGHS-l and -2 share 70% of their primary structure, while the
membrane binding domains of the enzymes are only 38% identical. Several non-
conservative amino acid substitutions exist in PGHS-2 at positions that may interact with
polar phospholipid heads of the membrane bilayer. An interesting question is whether the
sequence differences in the membrane binding domains of PGHS-1 and -2 may promote
association of each isozyme with different phospholipid microdomains within cellular
membranes.

Interactions between membrane proteins and the polar head groups of phospholipids
have been shown to be of biological importance [159-161]. In addition, the induction of

phospholipid microdomains upon exposure of biological membranes to physiological

91

concentrations of calcium imply that phospholipid microdomains may form in vivo [162].
The phospholipid microenvironment immediately surrounding an integral membrane
protein may affect its biological activity by promoting or restricting access of the protein to
various components of the cellular membrane. In the case of PGHS-1 and -2, the significant
sequence differences at the protein-phospholipid interface may confer differential
interactions with phospholipid microdomains with which the proteins associate in vivo.

We conclude that PGHS-1 and -2 interact with the ER through a novel domain
consisting of four amphipathic alpha helices. This region was previously hypothesized by

Picot, et al. to serve as the membrane binding domain.

APPENDIX

92

APPENDIX A

Future Directions

One interesting question arising from the work presented in Chapter II concerns the
topological properties and trafficking of integral membrane proteins in the nuclear envelope.
Our irnmuno gold localization experiments led us to speculate that the lumenal orientation
of PGHS-1 and -2 allows them to reach the inner nuclear membrane by lateral diffusion. We
plan to use immunoelectron microscopy to examine other ER membrane proteins in the
nuclear envelope. Of particular interest will be the distribution of cytoplasmically-oriented
ER proteins like the sterol regulatory element binding proteins (SREBPs) and thromboxane
A;2 synthase (TxS). One may predict that these proteins, which have large globular domains
on the cytoplasmic surface of the ER, are prevented from reaching the inner nuclear
membrane due to the physical barrier of the nuclear pore complex (Fig. 13). Determining
the nuclear envelope distributions of SREBPs, TxS, and other proteins with similar
membrane topography will provide important information on two fronts. First, we will have
an improved understanding of the cell biology of integral membrane proteins and the
topological properties of proteins needed to navigate the different membranes of the nuclear
envelope. Secondly, these studies may provide some insight into the possibile existence of
a nuclear prostaglandin signaling system. We have shown that PGHS-l and -2 are present
on the inner nuclear membrane; however, their product, PGHz, must be converted to active

prostaglandins by downstream synthases like TxS and PGIz-synthase. The distribution of

93

these prostaglandin synthases in the nuclear envelope is currently unknown but are likely to
play an important role in eicosanoid biosynthesis that may take place at the nuclear
envelope.

Much of the speculation surrounding the separation of the PGHS-l and -2 activities
has focused on the cell biological aspects of the enzymes. Differential localization, coupling
to specific PLAzs, FLAP-like arachidonate shuttling proteins, and localization to different
phospholipid microdomains have all been discussed as possible ways to separate the
activities of the PGHSs. However, it is possible that cell biological properties have little to
do with separating the activities of PGHS—l and -2 in living cells. There is some evidence
that biochemical properties of the two enzymes allow them to be exclusively active under
different conditions. Specifically, substrate-dependent allosteric effects have been shown to
produce a 2-4-fold greater rate of prostaglandin production by PGHS-2 compared to PGHS-1
at arachidonate concentrations below 0.5 pM (143). In effect, PGHS-2 could be
preferentially activated by a stimulation that released arachidonate at levels below the activity
threshold of PGHS-l. This would account for the ability of the enzymes to be exclusively
active despite the seemingly ubiquitous release of substrate.

The work hypothesizing different substrate thresholds for PGHS-l and -2 was
performed using purifed proteins. It may be possible to examine the substrate thresholds
required for activity in living cells. A fluorescent histochemical assay has been used by our
laboratory to visualize the synthesis of PGH2 by PGHS-1 and -2. This assay works by adding
a fluorescent peroxidase co-substrate that serves as a reducing substrate for the peroxidase
activity of the PGHSs. One is therefore able to visualize the activities of PGHS-1 and -2

together and individually through the use of isozyme-specific inhibitors. This histochemical

94

technique would allow one to determine which PGHS enzyme was active under various
stimulation conditions and at different concentrations of exogenously-added arachidonic
acid. Therefore, these studies would enable one to test the “substrate threshold” hypothesis
in a living experimental system.

As discussed above, the release of arachidonic acid and its subsequent availability to
PGHS-l and -2 is a complex process. The sheer number of mammalian PLAzs (over ten)
makes sorting out their individual physiological roles very difficult. However, the
relationships between the various PLAzs and the PGHS isozymes are clearly important to
eicosanoid signaling. It may be useful to develop an experimental system that would permit
the use of genetics to analyze the roles of proteins involved in eicosanoid biosynthesis.
Genetic analysis allows one to dissect complex signaling pathways by screening for
organisms deficient in a particular portion of the pathway. Furthermore, crosstalk between
a pathway of interest and other signaling pathways can be detected by genetic crossing of two
or more mutant strains. Most experimental systems of genetic analysis employ the use of
model organisms.

The nematode Caenorhabditis elegans may present a model system that would be
useful to genetically analyze prostanoid-like signaling pathways. This organism is a 959-cell
roundworm that has been used for over 20 years to study many aspects of developmental
biology. Its complete genomic sequence will be available in late 1998. C. elegans contain
both arachidonic (20:4) and eicosapentanoic acid (20:5), which are precursors for eicosanoids
in mammalian systems (168). They also contain DNA sequences that show sequence
homology to phospholipases A2. If one could show that a prostanoid-like signaling pathway

existed in C. elegans, then its genetic tractability might prove to be useful in disecting the

95
complex pathways controlled by lipid mediators. The existence of such a pathway in C.
elegans has neither been demonstrated nor looked for. However, the infastructure is present
in this organsism and the initial studies to determine the presence of lipid oxygenases are
straightforward biochemical assays using radiolabeled substrates and product analysis. If
successful, this sytem could provide insight into prostanoid biology as well as the role of

lipid mediators in development of metazoan organisms.

BIBLIOGRAPHY

10.

ll.

12.

13.

96

BIBLIOGRAPHY

Smith, W.L., R.M. Garavito, and D.L. DeWitt, Prostaglandin endoperoxide H
synthases (cyclooxygenases)-1 and -2 Prostaglandin endoperoxide H synthases-1 and
-2. J. Biol.Chem., 1996. 271(52): p. 33157-60.

Smith, W.L. and LJ. Marnett, Prostaglandin endoperoxide synthase: structure and
catalysis. Biochim. Biophys. Acta, 1991. 1083(1): p. 1-17.

Hemler, M. and WE. Lands, Purification of the cyclooxygenase that forms
prostaglandins. Demonstration of two forms of iron in the holoenzyme. J. Biol.
Chem, 1976. 251(18): p. 5575-9.

Herschman, H.R., Prostaglandin synthase 2. Biochim. Biophys. Acta, 1996. 1299(1):
p. 125-40.

Burr, G. and Burr, M.M., A new deficiency disease produced by the rigid exclusion
of fat from the diet. J. Biol. Chem, 1929. 82(1): p. 345-367.

Burr, G. and Burr., MM, On the Nature and role of the fatty acids essential in
nutrition J. Biol. Chem, 1930. 86(1): p. 587-619.

Morham, S.G., et al., Prostaglandin synthase 2 gene disruption causes severe renal
pathology in the mouse. Cell, 1995. 83(3): p. 473-82.

Langenbach, R., et al., Prostaglandin synthase 1 gene disruption in mice reduces
arachidonic acid-induced inflammation and indomethacin-induced gastric ulceration
Prostaglandin synthase 2 gene disruption causes severe renal pathology in the mouse.
Cell, 1995. 83(3): p. 483-92.

Dunham, E.W., Balasingham, M., Privett, 0.8., and Nickell, E.C., Effects of
essential fatty acid deficiency on prostaglandin synthesis and fatty acid composition
in renal medulla Lipids, 1978. 13: p. 892-897.

von Euler, U., Prostaglandins: Historical remarks, in Prog. Lipid Res. 1982. p. xxxi-
xxxv.

von Euler, U.S., On the specific vasodilating and plain muscle stimulating substances
from accessory genital glands in man and certain animals. J. Physiol., 1936. 88: p.
213-234.

Kurzrok, R.a.L., C.C., Biochemical studies on human semen. The action of semen
on the human uterus. Proc. Soc. Exp. Biol. Med, 1930. 28: p. 268-272.

Willis, A.L., Prosaglandins and Related Lipids. 1 ed. CRC Handbook of Eicosanoig.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

97

Vol. I. 1987, Boca Raton: CRC Press, Inc. 314.

Sinclair, H.M., Essential Fatty Acids. 1 ed. 1958, London: Butterworths Scientific
Publications. 264.

Bergstrom, S., Rhyhage, R., Samuelsson, B., and Sjovall, J., The structures of
prostaglandins El, Flu and F ,3. J. Biol. Chem, 1963. 238: p. 3555-3564.

Bergstrom, S., Danielsson, H., and Samuelsson, B., The enzymatic formation of
prostaglandin E2 from arachidonic acid Biochimica Biophysica Acta, 1964. 90: p.
207-210.

Van Dorp, D.A., Beerthuis, RIC, Nugteren, DH, and Vonkeman, H., The
biosynthesis of prostaglandins. Biochimica Biophysica Acta, 1964. 90: p. 204-297.

Van der Ouderaa, F.J., et al., Purification and characterisation of prostaglandin
endoperoxide synthetase from sheep vesicular glands. Biochim Biophys Acta, 1977.
487(2): p. 315-31.

Miyamoto, T., et al., Purification of prostaglandin endoperoxide synthetase from
bovine vesicular gland microsomes. J. Biol. Chem, 1976. 251(9): p. 2629-36.

Vane, J.R., Inhibition of prostaglandin synthesis as a mechanism of action for
aspirin-like drugs. Nat. New Biol, 1971. 231(25): p. 232-5.

Ferreira, S.H., S. Moncada, and JR. Vane, Indomethacin and aspirin abolish
prostaglandin release from the spleen. Nat. New Biol, 1971. 231(25): p. 237-9.

DeWitt, D.L. and W.L. Smith, Cloning of sheep and mouse prostaglandin
endoperoxide synthases. Methods Enzymol., 1990. 187: p. 469-79.

Smith, W.L., et al., Structure-function relationships in sheep, mouse, and human
prostaglandin endoperoxide G/H synthases. Adv. Prostaglandin T hromboxane
Leukot. Res, 1990. 20: p. 14-21.

Hla, T. and K Neilson, Human cyclooxygenase-2 cDNA. Proc. Natl. Acad. Sci.
USA, 1992. 89(16): p. 7384-8.

Shimokawa, T. and W.L. Smith, Prostaglandin endoperoxide synthase. The aspirin
acetylation region. J. Biol. Chem, 1992. 267(17): p. 12387-92.

Laneuville, O., et al., Fatty acid substrate specificities of human prostaglandin-
endoperoxide H synthase-1 and -2. Formation of 12-hydroxy-(9Z, 13E/Z, 15Z)-
octadecatrienoic acids from alpha-linolenic acid. J. Biol.Chem., 1995. 270(33): p.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

98

19330-6.

DeWitt, D.L., et al., The aspirin and heme-binding sites of ovine and murine
prostaglandin endoperoxide synthases. J. Biol. Chem, 1990. 265(9): p. 5192-8.

Rollins, TE. and W.L. Smith, Subcellular localization of prostaglandin-forming
cyclooxygenase in Swiss mouse 3T3 fibroblasts by electron microscopic
immunocytochemistry. J. Biol. Chem, 1980. 255(10): p. 4872-5.

Kennedy, 1., et al., Studies on the characterization of prostanoid receptors. Adv.
Prostaglandin T hromboxane Leukot. Res, 1983. 11: p. 327-32.

Coleman, R.A., et al., Characterisation of the prostanoid receptors mediating
contraction of guinea-pig isolated trachea: Studies on the characterization of
prostanoid receptors. Prostaglandins, 1985. 29(3): p. 363-75.

Fu, J.Y., et al., The induction and suppression of prostaglandin H2 synthase
(cyclooxygenase) in human monocytes. J. Biol. Chem, 1990. 265(28): p. 16737-40.

Xie, W.L., et al., Expression of a mitogen-responsive gene encoding prostaglandin
synthase is regulated by mRNA splicing. Proc. Natl. Acad. Sci. U S A, 1991. 88(7):
p. 2692-6.

Kujubu, D.A., et al., TISIO, a phorbol ester tumor promoter-inducible mRNA from
Swiss 3T3 cells, encodes a novel prostaglandin synthase/cyclooxygenase homologue.
J. Biol. Chem, 1991. 266(20): p. 12866-72.

Picot, D., P.J. Loll, and RM. Garavito, The X-ray crystal structure of the membrane
protein prostaglandin H2 synthase-1. Nature, 1994. 367(6460): p. 243-9.

Kurumbail, R.G., et al. , Structural basis for selective inhibition of cyclooxygenase-2
by anti-inflammatory agents. Nature, 1996. 384(6610): p. 644-8.

Meade, E.A., W.L. Smith, and D.L. DeWitt, Differential inhibition of prostaglandin
endopero xide synthase (cyclooxygenase) isozymes by aspirin and other non-steroidal
anti-inflammatory drugs. J. Biol. Chem, 1993. 268(9): p. 6610—4.

Laneuville, O., et al., Differential inhibition of human prostaglandin endoperoxide
H synthases-1 and -2 by nonsteroidal anti-inflammatory drugs. J. Pharmacol. Exp.
T her., 1994. 271(2): p. 927-34.

DeWitt, D.L., et al., PGH synthase isoenzyme selectivity: the potential for safer
nonsteroidal antiinflammatory drugs. Am. J. Med, 1993. 95(2a): p. 403-44s.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

99

Roth, G.J., C.J. Siok, and J. 02013, Structural characteristics of prostaglandin
synthetase from sheep vesicular gland J. Biol. Chem, 1980. 255(4): p. 1301-4.

Wendt, K.U., K Poralla, and GE. Schulz, Structure and function of a squalene
cyclase. Science, 1997. 277(5333): p. 1811-5.

Otto, J.C., D.L. DeWitt, and W.L. Smith, N-glycosylation of prostaglandin
endoperoxide synthases-l and -2 and their orientations in the endoplasmic reticulum.
J. Biol. Chem, 1993. 268(24): p. 18234-42.

Otto, J .C. and W.L. Smith, The orientation of prostaglandin endoperoxide synthases-
1 and -2 in the endoplasmic reticulum. J. Biol. Chem, 1994. 269(31): p.19868-75.

Smith, W.L. and D.L. Dewitt, Prostaglandin endoperoxide H synthases-1 and -2. Adv.
Immunol., 1996. 62: p. 167-215.

Silvia, W.J., Brockman, J.A., Kaminski, M.A., DeWitt, D.L., and Smith, W.L.,
Prostaglandin endoperoxide synthase in seminal vesicles. Mol. Androl., 1994. 6: p.
197-207.

Brannon, T.S., et al., Prostacyclin synthesis in ovine pulmonary artery is
developmentally regulated by changes in cyclooxygenase-l gene expression. J. Clin.
Invest, 1994. 93(5): p. 2230-5.

Xu, X.M., et al., Involvement of two Spl elements in basal endothelial prostaglandin
H synthase-1 promoter activity. J. Biol. Chem, 1997. 272(11): p. 6943-50.

Xu, X.M., et al., Transcriptional regulation of endothelial constitutive PGHS-l
expression by phorbol ester. Am. J. Physiol, 1996. 270(1 Pt 1): p. C259-64.

Breder, C.D., DeWitt, D.L., Kraig, R.P., Characterization of inducible
cyclooxygenase in rat brain. Journal of Comparative Neurolog, 1995 . 355: p. 296-
315.

DeWitt, D.L. and EA. Meade, Serum and glucocorticoid regulation of gene
transcription and expression of the prostaglandin H synthase-l and prostaglandin H
synthase-2 isozymes. Arch. Biochem. Biophys., 1993. 306(1): p. 94-102.

Murakami, M., et al., IgE-dependent activation of cytokine-primed mouse cultured
mast cells induces a delayed phase of prostaglandin D2 generation via prostaglandin
endoperoxide synthase-2. J. Immunol., 1995. 155(9): p. 4445-53.

Xie, W., Fletcher, B.S., Andersen, RD, and Herschman, H.R., v-src induction of the
T13 10/PGS2 prostaglandin synthase gene is mediated by an ATP/CRE transcription

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

62.

63.

100

response element. Mol. Cell Biol, 1994. 14: p. 6531-6539.

Inoue, H., Nanayama, T., Hara, 8., Yokoyama, C., and Tanabe, T., The cyclic AMP
response element plays an essential role in the expression of the human
prostaglandin-endoperoxide synthase 2 gene in differentiated U937 monocytic cells.
FEBS Lett., 1994. 350: p. 51-54.

Masferrer, J.L., et al., In vivo glucocorticoids regulate cyclooxygenase-2 but not
cyclooxygenase-1 in peritoneal macrophages. J. Pharmacol. Exp. Ther., 1994.
270(3): p. 1340-4.

Marnett, L.J., Aspirin and the potential role of prostaglandins in colon cancer.
Cancer Res, 1992. 52(20): p. 5575-89.

Tsujii, M. and RN. DuBois, Alterations in cellular adhesion and apoptosis in
epithelial cells overexpressing prostaglandin endoperoxide synthase 2. Cell, 1995 .
83(3): p. 493-501.

Kinzler, KW. and B. Vogelstein, Lessons from hereditary colorectal cancer. Cell,
1996. 87(2): p. 159-70.

Oshima, M., et al., Suppression of intestinal polyposis in Apc delta716 knockout
mice by inhibition of cyclooxygenase 2 (COX-2). Cell, 1996. 87(5): p. 803-9.

Giardiello, F.M., et al., Treatment of colonic and rectal adenomas with sulindac in
familial adenomatous polyposis. N. Engl. J. Med, 1993. 328(18): p. 1313-6.

Leslie, C.C., Properties and Regulation of Cytosolic Phospholipase A2. J. Biol.
Chem, 1997. 272: p. 16709-16712.

Tischfield, J .A., A reassessment of the low molecular weight phospholipase A2 gene
family in mammals. J. Biol. Chem, 1997. 272(28): p. 17247-50.

Schievella, A.R., et al., Calcium-mediated translocation of cytosolic phospholipase
A2 to the nuclear envelope and endoplasmic reticulum J. Biol. Chem, 1995 . 270(51):
p. 30749-54.

Glover, S., et al. , Translocation of the 85-kDa phospholipase A2 from cytosol to the
nuclear envelope in rat basophilic leukemia cells stimulated with calcium ionophore
or IgE/antigen [published erratum appears in J Biol Chem 1995 Sep
1;270(35):20870]. J. Biol. Chem, 1995. 270(25): p. 15359-67.

Borsch Haubold, A.G., et al., Identification of the phosphorylation sites of cytosolic
phospholipase A2 in agonist-stimulated human platelets and HeLa cells. J. Biol.

64.

65.

66.

67.

68.

69.

70.

71.

72.

73.

74.

75.

101

Chem, 1998. 273(8): p. 4449-58.

Lin, L.L., et al., cPLA2 is phosphorylated and activated by MAP kinase. Cell, 1993.
72(2): p. 269-78.

Nemenoff, R.A., et al., Phosphorylation and activation of a high molecular weight
form of phospholipase A2 by p42 microtubule-associated protein 2 kinase and protein
kinase C. .1. Biol. Chem, 1993. 268(3): p. 1960-4.

Sierra Honigmann, M.R., J .R. Bradley, and J .S. Pober, "Cytosolic" phospholipase A2
is in the nucleus of subconfluent endothelial cells but confined to the cytoplasm of
confluent endothelial cells and redistributes to the nuclear envelope and cell junctions
upon histamine stimulation. Lab. Invest., 1996. 74(3): p. 684-95.

Morita, 1., et al., Different intracellular locations for prostaglandin endoperoxide H
synthase-1 and -2. J. Biol. Chem, 1995. 270(18): p. 10902-8.

Spencer, A.G., Woods, J.W., Arakawa, T ., Singer, 1., and Smith, W.L., Subcellular
localization of prostaglandin endoperoxide H synthases-1 and -2 by immunoelectron
microscopy. J. Biol. Chem, 1998. 273: p. 9886-9893.

Woods, J.W., et al., 5-lipoxygenase and 5-lipoxygenase-activating protein are
localized in the nuclear envelope of activated human leukocytes. J. Exp. Med, 1993.
178(6): p. 1935-46.

Woods, J .W., et al., 5-Lipoxygenase is located in the euchromatin of the nucleus in
resting human alveolar macrophages and translocates to the nuclear envelope upon
cell activation. J. Clin. Invest., 1995. 95(5): p. 2035-46.

Lin, Y. , et al. , Docking phospholipase A2 on membranes using electrostatic potential-
modulated spin relaxation magnetic resonance. Science, 1998. 279(5 358): p. 1925-9.

Kudo, I., et al., Mammalian non-pancreatic phospholipases A2. Biochim. Biophys.
Acta, 1993. 1170(3): p. 217-31.

Murakami, M., I. Kudo, and K Inoue, Secretory phospholipases A2. J. Lipid Mediat.
Cell Signal, 1995. 12(2-3): p. 119-30.

Balsinde J, B.M., Dennis EA, Functional coupling between secretory phospholipase
A2 and cyclooxygenase-2 and its regulation by cytosolic group IV phospholipase A2.
Proc. Natl. Acad. Sci. U S A, 1998. 95(14): p. 7951-7956.

Reddy, ST. and HR. Herschman, Ligand-induced prostaglandin synthesis requires
expression of the TIS 1 O/PGS-2 prostaglandin synthase gene in murine fibroblasts and

76.

77.

78.

79.

80.

81.

82.

83.

84.

85.

86.

102

macrophages. J. Biol. Chem, 1994. 269(22): p. 15473-80.

Kirtikara, K, et al., Compensatory prostaglandin E2 biosynthesis in cyclooxygenase
l or 2 null cells. J. Exp. Med, 1998. 187(4): p. 517-23.

Reddy, ST. and HR. Herschman, Transcellular prostaglandin production following
mast cell activation is mediated by proximal secretory phospholipase A2 and distal
prostaglandin synthase 1. J. Biol. Chem, 1996. 271(1): p. 186-91.

Reddy, ST. and HR. Herschman, Prostaglandin synthase-1 and prostaglandin
synthase-2 are coupled to distinct phospholipases for the generation of prostaglandin
D2 in activated mast cells. J. Biol. Chem, 1997. 272(6): p. 3231-7.

Bingham, C.O., 3rd, et al., A heparin-sensitive phospholipase A2 and prostaglandin
endoperoxide synthase-2 are functionally linked in the delayed phase of
prostaglandin D2 generation in mouse bone marrow-derived mast cells. J. Biol.
Chem, 1996. 271(42): p. 25936-44.

Reddy, S.T., et al., Analysis of the secretory phospholipase A2 that mediates
prostaglandin production inmast cells. J. Biol. Chem, 1997. 272(21): p. 13591-6.

Balsinde, J., et al. , Inhibition of calcium-independent phospholipase A2 prevents
arachidonic acid incorporation and phospholipid remodeling in P388D1
macrophages. Proc. Natl. Acad. Sci. U S A, 1995. 92(18): p. 8527-31.

Balboa, M.A., et al., Novel group V phospholipase A2 involved in arachidonic acid
mobilization in murine P388D1 macrophages. J. Biol. Chem, 1997. 271(50): p.
32381-4.

Kuwata, H., Yoshihito Nakatani, Makoto Murakami, and Ichiro Kudo, Cytosolic
Phospholipase A2 Is Required for Cytokine-induced Expression of Type IIA
Secretory Phospholipase A2 That Mediates Optimal Cyclooxygenase-Z-dependent
Delayed Prostaglandin E2 Generation in Rat 3Y1 Fibroblasts. J. Biol. Chem, 1998.
273: p. 1733-1740.

Murakami, M., Satoko Shimbara, Terumi Kambe, Hiroshi Kuwata, Michelle V.
Winstead, Jay A. Tischfield, and Ichiro Kudo, The Functions of Five Distinct
Mammalian Phospholipase A23 in Regulating Arachidonic Acid Release. J. Biol.
Chem, 1998. 273: p. 14411-14423.

Dixon, R.A., et al., Requirement of a 5-lipoxygenase-activating protein for
leukotriene synthesis. Nature, 1990. 343(6255): p. 282-4.

Reid, G.K., et al., Correlation between expression of 5-lipoxygenase-activating
protein, 5-lipoxygenase, and cellular leukotriene synthesis. J. Biol. Chem, 1990.

87.

88.

89.

90.

91.

92.

93.

94.

95.

96.

97.

98.

103

265(32): p. 19818-23.

Kanai, N., et al., Identification and characterization of a prostaglandin transporter.
Science, 1995. 268(5212): p. 866-9.

Lu, R., et al., Cloning, in vitro expression, and tissue distribution of a human
prostaglandin transporter cDNA(hPGT). J. Clin. Invest., 1996. 98(5): p. 1142-9.

Ushikubi, F ., M. Hirata, and S. Narumiya, Molecular biology of prostanoid receptors;
an overview. J. Lipid Mediat. Cell. Signal, 1995. 12(2-3): p. 343-59.

Mais, D.E., T.A. True, and M.J. Martinelli, Characterization by photoaffinity
labelling of the human platelet thromboxane Az/prostaglandin H2 receptor: evidence
for N-linked glycosylation. Eur. J. Pharmacol., 1992. 227(3): p. 267-74.

Murray, R., E. Shipp, and GA. FitzGerald, Prostaglandin endoperoxide/thromboxane
A2 receptor desensitization. Cross-talk with adenylate cyclase in human platelets. J.
Biol. Chem, 1990. 265(35): p. 21670-5.

Murray, R. and GA. FitzGerald, Regulation of thromboxane receptor activation in
human platelets. Proc. Natl. Acad. Sci. U S A, 1989. 86(1): p. 124-8.

Habib, A., et al., Rapid, agonist-dependent phosphorylation in vivo of human
thromboxane receptor isoforms. Minimal involvement of protein kinase C. J. Biol.
Chem, 1997. 272(11): p. 7191-200.

Nishigaki, N., M. Negishi, and A. Ichikawa, Two Gs-coupled prostaglandin E
receptor subtypes, EP2 and EP4, differ in desensitization and sensitivity to the
metabolic inactivation of the agonist. Mol. Pharmacol., 1996. 50(4): p. 1031-7.

Ichikawa, A., Y. Sugimoto, and M. Negishi, Molecular aspects of the structures and
functions of the prostaglandin E receptors. J. Lipid Mediat. Cell Signal, 1996. 14(1-
3): p. 83-7.

Watkins, W. D. ,Peterson, M. B., Fletcher, J. R. ,Prostaglgm in Clinical Practice
1989, New York: Raven Press. 263.

Watabe, A., Y Sugimoto, A Honda, A Irie, T Namba, M Negishi, S Ito, S Narumiya,
and A Ichikawa, Cloning and expression of cDNA for a mouse EPl subtype of
prostaglandin E receptor. J. Biol. Chem, 1993. 268: p. 20175-20178.

Wahli, W. and E. Martinez, Superfamily of steroid nuclear receptors: positive and
negative regulators of gene expression. FASEB J., 1991. 5(9): p. 2243-9.

99.

100.

101.

102.

103.

104.

105.

106.

107.

108.

109.

110.

111.

104

Tontonoz, P., et al., mPPAR gamma 2: tissue-specific regulator of an adipocyte
enhancer. Genes Dev., 1994. 8(10): p. 1224-34.

Lehmann, J .M., et al., An antidiabetic thiazolidinedione is a high affinity ligand for
peroxisome proliferator-activated receptor gamma (PPAR gamma). J. Biol. Chem,
1995. 270(22): p. 12953-6.

Spiegelman, B.M., et al., PPAR gamma and the control of adipogenesis. Biochimie,
1997. 79(2-3): p. 111-2.

Kliewer, S.A., et al., A prostaglandin J2 metabolite binds peroxisome proliferator-
activated receptor gamma and promotes adipocyte differentiation. Cell, 1995 . 83(5):
p. 813-9.

Yu, K, et al., Differential activation of peroxisome proliferator-activated receptors
by eicosanoids. J. Biol. Chem, 1995. 270(41): p. 23975-83.

Forman, B.M., et al., 15-Deoxy-delta 12, 14-prostaglandin J2 is a ligand for the
adipocyte determination factor PPAR gamma Cell, 1995. 83(5): p. 803-12.

Ricote, M., et al., The peroxisome proliferator-activated receptor-gamma is a
negative regulator of macrophage activation Nature, 1998. 391(6662): p. 79-82.

Jiang, C., A.T. Ting, and B. Swd, PPAR-gamma agonists inhibit production of
monocyte inflammatory cytokines. Nature, 1998. 391(6662): p. 82-6.

Kim, J.B., Wright, H.M., Wright, M, and Spiegelman, B.M., ADDl/SREBPI
activates PPAR, through the production of endogenous ligand Proc. Natl. Acad. Sci.
USA, 1998. 95: p. 4333-4337.

Eling, TE. and AI. Ally, Pulmonary biosynthesis and metabolism of prostaglandins
and related substances. Environ. Health Perspect., 1984. 55: p. 159-68.

Matsuura, S., et al., Immunoelectron microscopy of the outer membrane of rat
hepatocyte nuclear envelopes in relation to the rough endoplasmic reticulum. Cell
Struct. Funct, 1983. 8(1): p. 1-9.

Martin, L., C. Crimaudo, and L. Gerace, cha cloning and characterization of
lamina-associated polypeptide 1c (Laplc), an integral protein of the inner nuclear
membrane. J. Biol. Chem, 1995. 270(15): p. 8822-8.

Furukawa, K, et al. , Cloning of a cDNA for lamina-associated polypeptide 2 (LAP2)
and identification of regions that specify targeting to the nuclear enve10pe. Embo. J. ,
1995. 14(8): p. 1626-36.

112.

113.

114.

115.

116.

117.

118.

119.

120.

121.

122.

123.

105

Bione, S., et al., Identification of a novel X-linked gene responsible for Emery-
Dreifuss muscular dystrophy. Nat. Genet, 1994. 8(4): p. 323-7.

Nagano, A., et al., Emerin deficiency at the nuclear membrane in patients with
Emery-Dreifirss muscular dystrophy. Nat. Genet, 1996. 12(3): p. 254-9.

Soullam, B. and H]. Worrnan, Signals and structural features involved in integral
membrane protein targeting to the inner nuclear membrane. J. Cell. Biol, 1995.
130(1): p. 15-27.

Smith, S. and G. Blobel, The first membrane spanning region of the lamin B receptor
is sufficient for sorting to the inner nuclear membrane. J. Cell. Biol, 1993. 120(3):
p. 631-7.

Ellenberg, J., et al., Nuclear membrane dynamics and reassembly in living cells:
targeting of an inner nuclear membrane protein in interphase and mitosis. J. Cell.
Biol, 1997. 138(6): p. 1193-206.

Shimokawa, T. and W.L. Smith, Essential histidines of prostaglandin endoperoxide
synthase. His-309 is involved in heme binding. J. Biol. Chem, 1991. 266(10): p.
6168-73.

Bhattacharyya, D.K., et al., Selective inhibition of prostaglandin endoperoxide
synthase-1 (cyclooxygenase-1) by valerylsalicylic acid. Arch. Biochem. Biophys,
1995. 317(1): p. 19-24.

Van Der Ouderaa, F.J., et al., Acetylation of prostaglandin endoperoxide synthetase
with acetylsalicylic acid. Eur. J. Biochem., 1980. 109(1): p. 1-8.

Lecomte, M., et al., Acetylation of human prostaglandin endoperoxide synthase-2
(cyclooxygenase-2) by aspirin J. Biol. Chem, 1994. 269(18): p. 13207-15.

Gierse, J .K, et al., A single amino acid difference between cyclooxygenase-1 (COX-
1) and -2 (COX-2) reverses the selectivity of COX-2 specific inhibitors. J. Biol.
Chem, 1996. 271(26): p. 15810-4.

McLean, I.W. and PK Nakane, Periodate-lysine-paraformaldehyde fixative. A new
fixation for immunoelectron microscopy. J. Histochem. Cytochem, 1974. 22(12): p.
1077-83.

Tokuyasu, KT., Use of poly(vinylpyrrolidone) and poly(vinyl alcohol) for
cryoultramicrotomy. Histochem. J., 1989. 21(3): p. 163-71.

124.

125.

126.

127.

128.

129.

130.

131.

132.

133.

134.

135.

136.

106

Kargman, S., et al., Characterization of Prostaglandin G/H Synthase 1 and 2 in rat,
dog, monkey, and human gastrointestinal tracts. Gastroenterologv, 1996. 111(2): p.
445-54.

Otto, J .C. and W.L. Smith, The orientation of prostaglandin endoperoxide synthase-1
in the endoplasmic reticulum of transiently transfected cos-1 cells. Adv.
Prostaglandin T hromboxane Leukot. Res, 1995. 23: p. 29-34.

Smith, W.L. and T.G. Bell, Immunohistochemical localization of the prostaglandin-
forming cyclooxygenase in renal cortex Am. J. Physiol, 1978. 235(5): p. F451-7.

Reiger, M.K , et al. , Subcellular localization of prostaglandin endoperoxide synthase-
2 in murine 3T3 cells. Arch. Biochem. Biophys, 1993. 301(2): p. 439-44.

Shimokawa, T., et al., Tyrosine 385 of prostaglandin endoperoxide synthase is
required for cyclooxygenase catalysis. J. Biol. Chem, 1990. 265(33): p. 20073-6.

Schilling, L.J. and P]. Farnham, Inappropriate transcription from the 5' end of the
murine dihydro folate reductase gene masks transcriptional regulation Nucleic. Acids

Res, 1994. 22(15): p. 3061-8.

Innis, J.W. and WA. Scott, Chromatin structure of simian virus 40—pBR322
recombinant plasmids in COS-1 cells. Mol Cell Biol, 1983. 3(12): p. 2203-10.

Hempel SL, M.M., Hunninghake GW, Lipopolysaccharide induces prostaglandin H
synthase-2 protein and mRN A in human alveolar macrophages and blood monocytes.
J. Clin. Invest., 1994. 93(1): p. 391-396.

Claria, J. and ON. Serhan, Aspirin triggers previously undescribed bioactive
eicosanoids by human endothelial cell-leukocyte interactions. Proc. Natl. Acad.
Sci.U.S.A., 1995. 92(21): p. 9475-9.

Evett, G.E., et al. , Prostaglandin G/H synthase isoenzyme 2 expression in fibroblasts:
regulation by dexamethasone, mitogens, and oncogenes. Arch. Biochem. Biophys,
1993. 306(1): p. 169-77.

Luong, C., et al., Flexibility of the NSAID binding site in the structure of human
cyclooxygenase-2. Nat. Struct. Biol, 1996. 3(11): p. 927-33.

Otto, J .C . and W.L. Smith, Photolabeling of prostaglandin endoperoxide H synthase-
1 with 3-trifluoro-3-(m-[‘251]iodophenyl)diazirine as a probe of membrane association
and the cyclooxygenase active site. J. Biol. Chem, 1996. 271(17): p. 9906-10.

Picot, D. and RM. Garavito, Prostaglandin H synthase: implications for membrane

137.

138.

139.

140.

141.

142.

143.

144.

145.

146.

147.

148.

107

structure. FEBS Lett, 1994. 346(1): p. 21-5.

T ontonoz, P., E. Hu, and BM. Spiegelman, Stimulation of adipogenesis in
fibroblasts by PPAR gamma 2, a lipid-activated transcription factor [published
erratum appears in Cell 1995 Mar 24;80(6):following 957]. Cell, 1994. 79(7): p.
1147-56.

Murakami, M., et al., c-kit ligand mediates increased expression of cytosolic
phospholipase A2, prostaglandin endoperoxide synthase-1, and hematopoietic
pro staglandin D2 synthase and increased I gE-dependent prostaglandin D2 generation
in immature mouse mast cells. J. Biol. Chem, 1995. 270(7): p. 3239-46.

Balsinde, J., et al., Arachidonic acid mobilization in P388D1 macrophages is
controlled by two distinct Ca(2+)-dependent phospholipase A2 enzymes. Proc. Natl.
Acad. Sci. USA, 1994. 91(23): p. 11060-4.

Murakami, M., Y. Nakatani, and 1. Kudo, Type II secretory phospholipase A2
associated with cell surfaces via C-terminal heparin-binding lysine residues augments
stimulus-initiated delayed prostaglandin generation J. Biol. Chem, 1996. 271(47):
p. 30041-51.

Kuhnacz, R.J., R.B. Pendleton, and WE. Lands, Interaction between peroxidase and
cyclooxygenase activities in prostaglandin-endoperoxide synthase. Interpretation of
reaction kinetics. J. Biol. Chem, 1994. 269(8): p. 5527-36.

Capdevila, J .H., et al. , The highly stereoselective oxidation of polyunsaturated fatty
acids by cytochrome P450BM-3. J. Biol. Chem, 1996. 271(37): p. 22663-71.

Swinney, D.C., et al., Differential allosteric regulation of prostaglandin H synthase
1 and 2 by arachidonic acid. J. Biol. Chem, 1997. 272(19): p. 12393-8.

Byrum, R.S., et al. , Role of the 5-lipoxygenase-activating protein (FLAP) in murine
acute inflammatory responses. J. Exp. Med, 1997. 185(6): p. 1065-75.

Abramovitz, M., et al., 5-lipoxygenase-activating protein stimulates the utilization
of arachidonic acid by S-hpoxygenase. Eur. J. Biochem., 1993. 215(1): p. 105-11.

Kargman, 3., Pl Vickers, and JP. Evans, A23187-induced translocation of 5-
lipoxygenase in osteosarcoma cells. J. Cell Biol, 1992. 119(6): p. 1701-9.

Vickers, P.J., S-Lipoxygenase-activating protein (FLAP). J. LipidMediat Cell
Signal, 1995. 12(2-3): p. 185-94.

Soullam, B. and H]. Worman, The amino-terminal domain of the lamin B receptor

149.

150.

151.

152.

153.

154.

155.

156.

157.

158.

159.

108

is a nuclear envelope targeting signal J. Cell Biol, 1993. 120(5): p. 1093-100.

Song, I. and W.L. Smith, C-terminal Ser/Pro-Thr-Glu-Leu tetrapeptides of
prostaglandin endoperoxide H synthases-1 and -2 target the enzymes to the
endoplasmic reticulum. Arch. Biochem. Biophys, 1996. 334(1): p. 67-72.

Wang, K, et al., Purification of an interleukin-1 beta converting enzyme-related
cysteine protease that cleaves sterol regulatory element-binding proteins between the
leucine zipper and transmembrane domains. J. Biol. Chem, 1995 . 270(30): p. 18044-
50.

Clark, J.D., et al., A novel arachidonic acid-selective cytosolic PLA2 contains a
Ca(2+)-dependent translocation domain with homology to PKC and GAP. Cell,
1991. 65(6): p. 1043-51 issn: 0092-8674.

Channon, J.Y. and CC. Leslie, A calcium-dependent mechanism for associating a
soluble arachidonoyl-hydrolyzing phospho lipase A2 with membrane in the
macrophage cell line RAW 264.7. J. Biol. Chem, 1990. 265(10): p. 5409-13.

Smith, W.L., Prostaglandin biosynthesis and its compartmentation in vascular
smooth muscle and endothelial cells. Annu. Rev. Physiol, 1986. 48: p. 251-62.

Ruan, KH., et al., Amino-terminal topology of thromboxane synthase in the
endoplasmic reticulum. J. Biol. Chem, 1993. 268(26): p. 19483-90.

Blobel, G., Intracellular protein topogenesis. Proc. Natl. Acad. Sci USA, 1980.
77(3): p. 1496-500.

Brunner, J. and G. Semenza, Selective labeling of the hydrophobic core of
membranes with 3-(trifluoromethyl)-3-(m-[1251]iodophenyl)diazirine, a carbene-
generating reagent. Biochemistry, 1981. 20(25): p. 7174-82.

Blanton, MP. and J .B. Cohen, Identifying the lipid-protein interface of the Torpedo
nicotinic acetylcholine receptor: secondary structure implications. Biochemistry,
1994. 33(10): p. 2859-72.

DeWitt, D.L., et al., Orientation of the active site and antigenic determinants of
prostaglandin endoperoxide synthase in the endoplasmic reticulum J. Biol. Chem,
1981. 256(20): p. 10375-82.

Movileanu, L., et al. , Selective association of phosphohpids as a clue for the passive
flip-flop diffusion through bilayer lipid membranes. Biosystems, 1997. 40(3): p. 263-
75.

 

160.

161.

162.

163.

164.

165.

166.

167.

168.

109

Paris, 8., et al., Role of protein-phospholipid interactions in the activation of ARFl
by the guanine nucleotide exchange factor Arno. J. Biol. Chem, 1997. 272(35): p.
22221-6.

Morrissey, J .H., et al., Factor VIIa-tissue factor: ftmctional importance of protein-
membrane interactions. T hromb. Haemost., 1997. 78(1): p. 112-6.

Haverstick, D.M. and M. Glaser, Visualization of Ca2+-induced phospholipid
domains. Proc. Natl. Acad. Sci. USA, 1987. 84(13): p. 4475-9.

Hempel SL, Monick MM, He B, Yano T, Hunninghake GW, Synthesis of
prostaglandin H synthase-2 by human alveolar macrophages in response to
lipopolysaccharide is inhibited by decreased cell oxidant tone. J. Biol. Chem. 1994,
269:52, p. 32979-32984.

Hempel SL, Monick MM, Hunninghake GW, Lipopolysaccharide induces
prostaglandin H synthase-2 protein and mRNA in human alveolar macrophages and
blood monocytes. J. Clin. Invest., 1994, 93:1, p. 391-396.

Hoff, T., DeWitt, D., Kaever, V., Resch, K,Goppelt-Struebe, M. Differentiation-
associated expression of prostaglandin G/H synthase in monocytic cells. 1993, FEBS
Lett, 320:], p. 38-42.

Funk, C.D., Furci, L., FitzGerald, G.A., et al., Point mutatioin in the seventh
hydrophobic domain of the human thromboxane A2 receptor allows discrimination
between agonist and antagonist binding sites. 1993, Mol Pharmacol. 4, 934-939.

Sugimoto, Y., Yamasaki, A., et al. Failure of Parturition in Mice Lacking the
Prostaglandin F2“ Receptor. 1997, Science 277: 681-683.

Tanaka, T., et al. Effects of growth temperature on the fatty acid composition of the
flee-living nematode Caenorhabditis elegans. 1996, Lipids 31:11, 1173-1178.

nrcuran STATE UNIV. LIBRRRIES
1|11111111llllllllmmllllWlllllllllllllllllll1111|
31293017662481