.,~:.;...:nr~ ‘q.

CIA—.1 at.“

é3w¢m¢4

#931

... ,1“ -

 

u
: 1‘
“13"!"

g

l
. 3} u’,
j 33’
1'3

J I” "f . u

- in :-!’J’ nil-7 '
t ,2“ ‘ (r 4.31.}. an a"
:23}; ‘33,: 3133, $5523},
7 ‘l’ '-

. .- i f;

 

 

This is to certify that the
dissertation entitled

STRUCTURAL AND CATALYTIC DETERMINANTS OF
INTRACELLULAR CYCLOOXYGENASE PROTEIN
DEGRADATION

presented by

URI R. MBONYE

has been accepted towards fulfillment
of the requirements for the

Doctoral degree in Biochemistry

 

 

/%/@g_ z . (524%

 

Major Professor’s Signature

rayon/a7

Date

MSU is an affinnative—action, equal-opportunity employer

 

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

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/07 p:/ClRC/DateDue.indd-p.1

STRUCTURAL AND CATALYTIC DETERNIINANT S OF INTRACELLULAR
CYCLOOXYGENASE PROTEIN DEGRADATION

By

Uri R. Mbonye

A DISSERTATION
Submitted to
Michigan State Universtiy

in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Biochemistry and Molecular Biology

2007

ABSTRACT

STRUCTURAL AND CATALYTIC DETERMINANTS OF INTRACELLULAR
CYCLOOXYGEN ASE PROTEIN DEGRADATION

By

Uri R. Mbonye

Cyclooxygenases (COX-1 and COX-2) are ER-resident, membrane-bound
glycoproteins that catalyze the committed step in the synthesis of biologically active lipid
hormones called prostanoids. COX-1 is constitutively expressed in many mammalian
cells, whereas COX-2 is usually expressed inducibly and transiently. In murine NIH/3T3
fibroblasts, COX-2 protein is degraded by a proteasome-dependent process with a half-
life (ha) of about 2 h whereas COX-1 is reasonably stable (t1/2> 12 h). The mature forms
of the COX isoforms are very similar in structure except that COX-2 has a unique 19-
amino acid (19-aa) segment of unknown function located just near the C-terminus. Here,
we provide evidence that the major role of the l9-aa cassette is to mediate entry of COX-
2 into the ER-associated degradation (ERAD) system that transports ER proteins to the
cytoplasm for degradation by the 26S proteasome. COX-l expressed heterologously in
HEK293 cells is quite stable (t1,2> 24 h) while COX-2 expressed heterologously is
degraded with a tug of ~5 h and its degradation is blocked by proteasome inhibitors. A
deletion mutant of COX-2 lacking 18 residues of the 19-aa cassette retains native COX-2
enzymatic activity and subcellular localization but, unlike native COX-2, is stable in
HEK293 cells (t1,2> 24 h). Conversely, inserting the COX-2 cassette near the C-terminus
of COX-1 yields a mutant ins594-612 COX-1 that is unstable (ti/2 ~ 3 h). A mannosidase

inhibitor that can block entry of ER proteins into the ER-associated degradation system,

retards the degradation of COX-2 and ins594-612 COX-l. Scanning mutagenesis of the
l9-aa cassette has revealed that Asn-594, an N-glycosylation site at the beginning of the
l9-aa cassette, is required, albeit insufficient to enable ERAD entry of COX-2 and
ins594-612 COX-1. The remaining 16 amino acids of the l9-aa cassette are not only
essential for COX-2 degradation but are also required for the enzyme to become
glycosylated at Asn-594. Structural and biochemical analysis of the region immediately
upstream of Asn-594 has revealed an 8-residue a—helical region that appears to impede
Asn-594 glycosylation. We propose that a structural conformational change of the or-
helical structure mediated by the C-terminal l6-residue sequence of the l9-aa cassette is
needed for Asn—594 glycosylation. We show that the helical region upstream of Asn-594
and the l9-aa cassette actually constitute a C-terminal 27 instability element (27 IE) that
dictates the intracellular stability of COX-2 by regulating glycosylation at Asn-594. We
have also identified a second distinct mechanism of COX-2 protein turnover that is
induced by arachidonic acid (AA), a major cylooxygenase (COX) fatty acid substrate.
AA-dependent degradation of COX-2 is blocked by COX inhibitors, is not affected by
inhibitors of proteasomal or lysosomal degradation, and occurs independently of the C-
terminal l9-aa cassette of the enzyme. A COX-inactive point mutant of COX-2 (G533A
COX-2) is resistant to AA-induced degradation even though it undergoes substrate-
independent degradation at the same rate as native COX-2. Loss of COX-2 specific
activity and prostanoid product formation is also observed in COX—2 expressing cells
treated with AA. We propose that substrate-dependent inactivation of COX-2 causes

structural damage to the enzyme that subsequently leads to its degradation.

To Baba, my Mom, and the loving memory of my grandfather Reuben Rwabusizoni

iv

 

 

 

ACKNOWLEDGMENTS

I would like to express my utmost gratitude to my advisor Dr. Bill Smith for his
wonderful mentorship. He has done an excellent job of guiding me through a very
challenging and thought-provoking thesis project. I have relied on his remarkable
guidance and intellectual and technical expertise to be able to successfully complete my
graduate career. I also sincerely acknowledge my advisory committee for providing me
with very helpful advice and constantly ensuring that my thesis research is on the right
track.

I extend my appreciation to members of our lab, both present and past, for all the
collaboration, advice, technical help, and friendship that I have received throughout my
graduate career. Much thanks goes to Dr. Rana Sidhu for teaching me how to use the
PyMOL software for protein structure analysis. He has also been a very cool and
approachable guy that I have constantly turned to for suggestions and help on wide-
ranging issues. Much thanks also goes to: Dr. Chong Yuan for teaching me, among other
things, the very useful technique of overlap-extension PCR; Dr. Masayuki Wada and Jill
Rieke for collaborative research related to examining the biological function of the COX-
2 C-terminal l9-amino acid cassette; Drs. Cindy DeLong and Inseok Song for helpful
discussions and teaching me how to perform TLC assays; and Drs. Jiayan Liu, Yeon-Joo
Kang, and Christine Harman for intellectually stimulating discussions. Thank you all for
providing me with a conducive and friendly atmosphere to pursue my thesis research.

Finally, I take this opportunity to also thank my family for their constant love and

support.

TABLE OF CONTENTS

LIST OF TABLES ............................................................................... viii
LIST OF FIGURES ................................................................................. ix
ABBREVIATIONS ................................................................................ xii
CHAPTER I: LITERATURE REVIEW
Introduction ................................................................. l
Mobilization of Endogenous AA for Prostanoid Synthesis .......... 6
Coupling between Cyclooxygenases and PLAzs ...................... 9
Functional Coupling between Cyclooxygenases and
Terminal Prostanoid Synthases ........................................ 10
Regulation of Cyclooxygenase Gene Expression .................... 13

CHAPTER II:

Cyclooxygenase Protein Structure and
Subcellular Compartmentation ......................................... 20

Cyclooxygenase Catalysis, Substrate Specificity,
and Suicide Inactivation ................................................. 28

Proteolytic Degradation of Intracellular Proteins .................... 33

Proteolytic Degradation of Endoplasmic Reticulum-
associated Proteins ....................................................... 35

THE 19 AMINO CASSETTE OF CYCLOOXYGENASE-2
MEDIATES THE ENTRY OF THE PROTEIN INTO THE ER-
ASSOCIATED DEGRADATION SYSTEM

Summary .................................................................. 46
Introduction ............................................................... 47
Experimental Procedures ................................................ 49
Results ..................................................................... 53

vi

Discussion ................................................................. 67

Acknowledgments ....................................................... 70
CHAPTER III: ASN-594 GLYCOSYLATION AND PROTEIN DEGRADATION
OF COX-2 IS REGULATED BY ITS C-TERMINAL 27 AMINO
ACID INSTABILITY ELEMENT
Summary .................................................................. 71
Introduction ............................................................... 72
Experimental procedures ................................................ 75
Results ..................................................................... 79
Discussion ............................................................... 108
Acknowledgments ...................................................... l 16
CHAPTER IV: SUBSTRATE DEPENDENT COX-2 PROTEIN
DEGRADATION
Summary ................................................................. 1 17
Introduction .............................................................. 1 18
Experimental procedures .............................................. 121
Results .................................................................... 126
Discussion ............................................................... 153
CONCLUSION .................................................................................... 160
APPENDIX ......................................................................................... 162
BIBLIOGRAPHY ................................................................................. 170

vii

Table 1.

Table 2.

Table 3.

Table 4.

LIST OF TABLES

Participants of glycoprotein ERAD- L in yeast and
mammalian cells .................................................................. 45

Relationship between the extent of Asn-594 glycosylation
and overall protein degradation ................................................ 107

C-terminal mutants designed by QuickChangeTM
site-directed mutagenesis ....................................................... 162

C-terminal mutants designed by overlap-extension PCR
mutagenesis ...................................................................... 165

viii

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Figure 5a.

Figure 5b.

Figure 6.

Figure 7a.

Figure 7b.

Figure 8.

Figure 9a.

Figure 9b.

Figure 10.

Figure 11.

LIST OF FIGURES

Synthesis and incorporation of AA into membrane
glycerophospholipids ............................................................... 2

Stimulus-induced synthesis of bioactive prostanoids .......................... 5

Domain alignment of COX-1 and COX-2 highlighting
major differences in their primary structures .................................... 6

Ribbons diagrams of the three dimensional structures
of ovine COX-1 and murine COX-222

A view from the membrane plane showing the
opening into the cyclooxygenase active site of
ovine COX-l created by the membrane binding domain .................... 23

A view from the top showing the peroxidase active
site of ovine COX-1. Proximal His-388 coordinates
the heme prosthetic group (red) at the active site by
bonding to the ferric heme ....................................................... 24

Ribbon diagrams showing two views (A and B) of
the C-terminal structure of murine COX-2 ................. _ ................... 27

Mechanism for bis-oxygenation of AA and subsequent

reduction of PGGz to form PGH; ............................................... 29
Proposed catalytic mechanism of the COX isoforms ........................ 3O
Mechanism for the polyubiquitination of
proteins prior to their degradation by the 26S proteasome .................. 37
Role of the calnexin/calreticulin cycle in ER quality control ............... 43

Processing of N-linked oligosaccharides in the ER lumen
during proten quality control .................................................... 44

Cyclooxygenase protein degradation in serum—treated
NIH/3T3 cells ...................................................................... 55

Stability of heterologously expressed native
and C-terminal cyclooxygenase mutants in HEK293 cells .................. 58

ix

Figure 12.

Figure 13.

Figure 14.

Figure 15.

Figure 16.

Figure 17.

Figure 18.

Figure 19.

Figure 20.
Figure 21.

Figure 22.

Figure 23.

Figure 24.

Figure 25.

Figure 26.

Subcellular localization of native huCOX-2, delS95-612
huCOX-2, native ovCOX-l and ins594-612 ovCOX-l
in HEK293 cells ................................................................... 60

Kinetic properties of purified native muCOX- 2 and
de1595 -612 muCOX- 2 ............................................................ 61

Stability of cyclooxygenase mutants in HEK293 cells
having various modifications of the 19 amino acid cassette ................ 63

Glycosylation of ins594-612 ovCOX-l and native huCOX-2
at Asn-594 ......................................................................... 66

Deletion and scrambled mutations of the 19-aa cassette
disrupt COX-2 degradation ...................................................... 83

The C-terminal end of l9-aa is essential for COX-2
protein degradation ............................................................... 87

KIF treatment stabilizes the Asn-594 glycosylated form

of COX-2 ........................................................................... 90
COX-2 stabilizing mutations of the l9-aa prevent

glycosylation of Asn-59493
Structure of the C-terminal region of COX-1 and COX-2 .................. 98
Degradation of the UPSTRM8 huCOX-2 mutants .......................... 101

Asn-594 glycosylation of the UPSTRM8 huCOX-2
mutants and their degradation by the ERAD pathway ...................... 106

Schemes illustrating how the C-terminus of COX-2
may regulate Asn—594 glycosylation .......................................... 112

Two possibilities of how Asn-594 of native
huCOX-2 may become accessible to glycosylation ......................... 113

A model of how Asn-594 glycosylation mediates the
entry of COX-2 into the ERAD pathway ..................................... 114

Cyclooxygenase inhibitors prevent the rapid
degradation of COX-2 in serum-treated NIH/3T3,
but not HEK293, cells ........................................................... 130

Figure 27. COX-2 degradation in HEK293 cells is enhanced
by exogenous AA treatment .................................................... 136

Figure 28. COX-2 degradation is selectively induced by PUFA
COX substrates for the enzyme ................................................ 139

Figure 29. The G533A cyclooxygenase-inactive COX-2 mutant
is not susceptible to substrate-dependent degradation ...................... 140

Figure 30. The enhancement of COX-2 degradation in NIH/3T3
cells by exogenous and endogenous AA is inhibited
by NSAIDS ....................................................................... 145

Figure 31. COX-1 stability is not significantly affected by
exogenous AA .................................................................... 146

Figure 32. Inhibitors of proteasomal and lysosomal
degradation do not prevent substrate-dependent
COX-2 protein turnover ......................................................... 150

Figure 33. COX activity and prostaglandin product formation
of His ovCOX-l and His muCOX-2 cells pretreated
with exogenous AA .............................................................. 152

Figure 34. Model for the basal and COX substrate-initiated

degradation of COX-2 in HEK293 and serum-treated
NIH/3T3 cells ..................................................................... 156

xi

2-AG ether
3’-UTR
19-aa
AA
BFA
CHX
COX
CPLAza
DHLA
EDA
Endo H
EPA
Epox
ER
ERAD
ER Man I
FB

FBS

GI

GH

Glc
GlcNAc
hu
IFN-Y
KIF

LA
Leup
LPS
Man
mu
NSAID
OA
OST

ov

PBS
POX
SPLAZ
tet

TLC
UGGT
UPSTRM8

ABBREVIATIONS

2-arachidonyl glycerol ether
3’-untranslated region
C-terminal l9-amino acid cassette of COX-2
arachidonic acid '
brefeldin A

cycloheximide
cyclooxygenase

cytosolic phospholipase Aza
dihomo-y-linolenic acid
eicosadienoic acid
endoglycosidase H
eicosapentaenoic acid
epoxomicin

endoplasmic reticulum
ER-associated degradation
ER Ot-l,2 mannosidase I
flurbiprofen

fetal bovine serum
glucosidase I

glucosidase 11

glucose

N-acetyl glucosamine
human

interferon-y

kifunensine

linoleic acid

leupeptin

bacterial lipopolysaccharide
mannose

murine

non-steroidal anti-inflammatory drug
oleic acid
oligosaccharyltransferase
ovine

phosphate buffered saline
peroxidase

secretory phospholipase A2
tetracycline

thin layer chromatography
UDP-glucose glucosyltransferase

the 8 amino acid segment immediately upstream of Asn-594 in COX-2

xii

CHAPTER I

LITERATURE REVIEW

Introduction

Prostanoids, namely prostaglandins, prostacyclin and thromboxanes, are potent
lipid signaling molecules that are synthesized in most mammalian tissues. These lipid
hormones belong to the family of twenty carbon fatty acid derivatives that are
collectively known as eicosanoids. Prostanoids act locally in an autocrine or paracrine
manner through transmembrane G-protein coupled receptors to elicit a wide range of
physiological and pathological responses (1-7). The major fatty acid precursor for
prostanoids is widely believed to be arachidonic acid (AA), a twenty carbon fatty acid
that is a metabolite of the dietary essential fatty acid linoleic acid. Once synthesized de
novo AA is preferentially esterified to the sn-2 position of glycerophospholipids in lipid
bilayers of cell membranes during phospholipid remodeling (Fig. 1) (8). The release of
free AA from membrane phospholipids by phospholipase A2 is tightly regulated and
represents the initial and first rate-limiting step of prostanoid synthesis (8-10). A second
key point of control of prostanoid synthesis is the subsequent step involving the
conversion of free AA to prostaglandin H2 (PGHz) by prostaglandin endoperoxide H2
synthase (PGHS) also known as cyclooxygenase (COX). This step is also rate-limiting
for prostanoid formation and is considered to be the committed step in the pathway (1,3).
Cell specific prostanoid synthases then catalyze the isomerization or reduction of PGHz

to form various bioactive prostanoids (Fig. 2).

Dietary linoleic Acid (18:2A”n(1)6)
ATP + CoA

AMP + PPi
linoleoyl-CoA

A6 desaturase

y-linolenoyl CoA (18:2A6’9'12w6)

malonyI-CoA

dihomo-y-linolenoyl—CoA (20:3A8’11'Mw6)

0
R2__.| |__ 0— Co A As desaturase

Arachidonoyl-CoA (20:4A5’3’ll’1‘ufi)

  

l o
o
o—P—o—X R2 | | o l (I)
o—lT-o-X

0

Figure 1. Synthesis and incorporation of AA into membrane glycerophospholipids.
All mammalian cells except erythrocytes synthesize AA from the essential fatty acid
linoleic acid (LA). As shown in the above scheme, AA synthesis from LA involves

formation of two additional double bonds (A5 and A8) and two-carbon chain elongation.
Once synthesized, AA is specifically incorporated into the sn-2 position of
glycerophospholipids by acyltransferase during phospholipid remodeling.

There are two known isoforms of cyclooxygenase: COX-1 and COX-2. Both are
membrane-bound heme-containing glycoproteins that possess two sequential catalytic
activities: a cyclooxygenase activity that bis-oxygenates AA to form the hydroperoxide

prostaglandin G2 (PGGz) and a peroxidase activity that reduces the 15-hydroperoxyl

group of PGGz to form PGHZ (Fig. 2) (1,3,7). COX-1 and COX-2 are structural and
functional homodimers and are known to associate monotopically with the luminal faces
of the endoplasmic reticulum (ER) membrane and the nuclear envelope (11-17). Their
membrane association is such that the opening of the cyclooxygenase active site is in the
lipid bilayer to allow the uptake of mobilized free AA. Despite being monotopically
associated with the membrane these isoenzymes are considered to be integral membrane
proteins because they can only be removed from the membrane with detergent and not
with chaotropic salts (15). Thus, by associating with one leaflet of the lipid bilayer, the
COXs constitute a unique class of integral membrane binding proteins.

COX-1 and -2 are encoded by different genes; nonetheless, they share ~60%
identity in primary structure. Both isoforms have a cleavable N-temiinal signal peptide
sequence, an epidermal growth factor-like domain, a membrane binding domain, and a
globular catalytic domain. At the C-termini of COX-1 and -2 are KDEL-like motifs that
have been shown to be important for retention of the enzymes in the ER (18,19).
Adjacent to the ER retention signal of COX-2 is a l9-amino acid insertion that is lacking
in COX-l (Fig. 3). This C-terminal insertion also confers COX-2 with an additional
consensus site for N-glycosylation. The biological role of this COX-2 C-terminal
insertion will be explored in detail in Chapters 11 and 111.

Although COX-1 and COX-2 are structural and functional isoforms, they have
different expression patterns and biological functions. COX-1 is constitutively expressed
in resting cells of most tissues (1,20). In contrast, COX—2 expression is usually absent in
resting cells and can be induced in fibroblast, epithelial, endothelial, macrophage, and

smooth muscle cells in response to growth factors, cytokines, mitogenic, and pro-

inflammatory stimuli (1,20-24). Induced COX-2 expression is usually observed to be
transient (21-24). It is possible that the differential expression profiles of COX-1 and -2
may contribute to their contrasting biological roles. In general, prostanoids generated as a
result of COX-1 activity are believed to play housekeeping roles (25) while those
generated by COX-2 are not only important physiologically but have also been
implicated in various pathophysiological processes such as inflammation (26), pain (27),
fever (28), angiogenesis (29), and tumorigenesis (30). The COXs are the best known
cellular target for non-steroidal anti inflammatory drugs (NSAIDS) such as aspirin,

ibuprofen, indomethacin, and flurbiprofen (1).

O
mass: A 333.53? 5383:. «not... 368:... 38—953 II.— _I|~.
/ >> 4

IX. 7on
2.? _
o

 

 

 

 

IIIIIIIIIIII

3m. . .3 W o I o:
Qasewgmoaoamo
w my \ o:u

      

 

 

TO I I

t». I OI OI 30 $3998

(4 r
019 z»

k n 3.83323

IO 4.3.0360qu
OI $559.8 $5598
81?

3 ‘IE/ \.\
..o D I I0 \\

    
 

m
" .32! m:
I0 01

mam—:6 N. mange—smbuacnaa mus—98mm 3. 38596 enema—seam. .25 mag B8 magnum find em Roman—Boa
mew—demo: mm 9o mange—swinging Bog—masses 0». >> ~89 505335 uremusosuam 3. 2.8232630 >~
Got/NV. 38 >> mm 8:333 8 wmmw m: a 25-3% Boone: 90x08 OBS—Ema 3 En owaooxwmoammo 805
3a Rain—ewe QOVO coming cm OOXL Ba -N. womim m 983 ©8953 mew $503 ESSA—cam man So:
mosoammoz mm 3855- BE oo=éfio mwoommo. Zm>~Um :55: 9o 00X waning e». OOXL Ba OOX-~.

A5068 Mn“ 0

i” ii: “l’
7 _.J

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Signal EGF MBD Globular-Catalytic . COX-1
Asn67 Asn144 Asn410 Asn594
T W W) Ii) .\ ERSTELCCC"
Signal ~—~ EGF MBD Globular-Catalytic E :: COX-2
// L\
/ \
/ \

594 / \612
§5§SSRSGLDDINPTVLLK human
§§§ASHSRLDDINPTVLIK mouse
§§§SSHSGLDDINPTVLLK ovine
§§§ASHSRLDDINPTVLIK rat

Figure 3. Domain alignment of COX-1 and COX-2 highlighting major differences in
their primary structures. The cleavable ER signal sequence of COX-2 is shorter (18
residues) than that of COX-1 (25 residues). COX-1 is found to be glycosylated at three
sites whereas COX-2 has four functional N-glycosylation sites. The last glycosylation site
of COX-2 (Asn-594) is variably glycosylated and is part of a unique C-terminal l9-amino
acid insertion (19-aa). Numbering for COX-1 begins with the Met at the translation start
site. Numbering for COX-2 parallels the COX-1 numbering in which the start of the
mature, processed COX-2 protein has the same number as the start of the mature,
processed COX-1 (12). The Asn-594 glycosylation site is underlined, and 19-aa residues
that are conserved in humans, mouse, sheep, and rat have been colored green. Images in
this dissertation are presented in color.

Mobilization of Endogenous AA for Prostanoid Synthesis
Membrane glycerophospholipids can be deacylated at the sn-2 position by
phospholipase A2 (PLA2) enzymes leading to the production of free polyunsaturated fatty

acids, usually AA (31), and lysophospholipids. The mobilization of AA from the cell

membrane for prostanoid synthesis depends on stimuli that will activate catalysis and/or

induce the expression of the PLA2s involved. Therefore, AA liberation from the
membrane is rate limiting for prostanoid metabolism. There is overwhelming evidence to
indicate that of the nearly 22 mammalian PLA2 enzymes that have been cloned and
characterized, cytosolic PLA20t (cPLA20L) and certain members of the family of secretory
PLA2s (SPLA2) are largely responsible for the mobilization of AA for prostanoid
metabolism (32-34). Of these two categories of CaZ+-dependent PLA2s, a consensus is
emerging that cPLA201 is critical for stimulus-dependent liberation of AA from
membrane phospholipids.

cPLA20t is expressed constitutively in most mammalian cell types (35). The
expression of cPLA20t can be elevated by various stimuli such as growth factors,
cytokines and proinflammatory agonists (36-40). It is tightly regulated by multiple
signaling pathways that control intracellular Ca2+ concentration and the phosphorylation
state of the enzyme. Under the appropriate stimuli that will raise the intracellular
cytosolic Ca2+ concentration to rnicromolar levels (eg. bradykinin and thrombin) (35,41-
43), cPLA20t will bind Ca2+ at its N-terminal C2 domain and, consequently, translocate to
the ER membrane and nuclear envelope (44-46). It is thought that if the Ca2+ flux is low
and transient, phosphorylation of the C-terminal catalytic domain of cPLA201 by mitogen
activated protein kinases (MAPKs) may increase the affinity of the Ca2+-bound lipase for
membrane binding (10,47). Caz+-dependent membrane association enables cPLA20L to
preferentially cleave AA from the sn-2 position of phospholipid substrates. That CPLA20t
is the phospholipase that is essential for mobilizing free AA from cellular membranes for
eicosanoid metabolism is unequivocally demonstrated by studies using genetically

deficient cPLA201 mice. These mice are generally reported to be healthy and fertile (48-

50). However, they are resistant to pathologies that require the actions of prostanoids and
other eicosanoids generated from the agonist-induced mobilization of AA (48,49,51-57).
These models include collagen-induced rheumatoid arthritis, bleomycin-induced lung
fibrosis, brain ischemia reperfusion injury, acute respiratory distress syndrome,
autoimmune diabetes, multiple sclerosis, and intestinal polyposis (48,49,51-57).

sPLA2s, particularly those belonging to Groups IIA, V, and X are thought to
complement cPLA20t activity by augmenting the release of cellular AA under certain
pathophysiological settings such as those of an inflammatory response where the stimulus
is robust and prolonged (58-60). Since sPLA2s are usually expressed extracellularly it is
reasonable to expect that these lipases would have a significant role in transcellular
prostanoid biosynthesis. sPLA2s require micromolar to millimolar concentrations of Ca2+
for catalytic activity and will indiscriminately hydrolyze cell surface phospholipids at the
sn-2 position (10,61). Recent evidence has shown that prior to their secretion sPLA2s
could also act on intracellular membranes to mobilize endogenous AA (33,34). The
expression of sPLA2s, which is usually low in unstimulated immune effector cells (ie.
macrophages, neutrophils, and mast cells) and in tissues of normal healthy individuals, is
induced by proinflammatory stimuli such as IL-lB, TNFOL, interferon-y, phorbol ester,
and bacterial lipopolysaccharide (62-66). Abnormally high levels of sPLA2-IIA can be
detected in synovial fluid of patients with rheumatoid arthritis (67,68) or in plasma of
patients experiencing systemic inflammation (69-71). Not surprisingly, the marked
elevation of sPLA2-IIA during an inflammatory response is often accompanied by an
induction of COX-2 expression and elevation of prostaglandins (10,72). The possible

functional coupling between sPLA2-IIA and COX-2 to generate prostaglandins in

inflammatory cells such as macrophages and mast cells appears to absolutely require
cPLA201 activity (34,73,74). These observations have implied that the activation of
cPLA20t precedes the mobilization of AA by sPLA2-HA for COX-2-mediated eicosanoid
formation. However, the mechanistic basis for this crosstalk between cPLA20t and

sPLA2-HA is not understood.

Coupling between Cyclooxygenases and PLA2s

Free AA that is mobilized by cPLA201 and sPLA2 is acted upon by
cyclooxygenases to form the intermediate PGH2 that is a substrate for terminal prostanoid
synthases. It is still not clear whether there is preferential coupling between
cyclooxygenases and the PLA2s. Several research groups have reported that there exists
two distinct phases of prostaglandin synthesis that are to a significant extent dictated by
the stimulus involved in activating PLA2, the endogenous levels of free AA, and the
temporal expression of COX-2 (75-77). The immediate phase usually occurs within
minutes after stimulation with bradykinin or calcium ionophore (A23187) and is
characterized by a rapid and transient increase in cytoplasmic Ca2+ concentration (76).
Since COX-1 is constitutively expressed, it is reasonable to expect that during the
immediate phase of prostaglandin and thromboxane synthesis cPLA20t is functionally
coupled to COX-1. The delayed phase of prostaglandin synthesis which is induced by
proinflammatory stimuli such as IL-IB occurs over several hours of cell stimulation and
is largely mediated by inducible COX-2 activity (76). The delayed phase also appears to
be dependent on cPLA2a activity even though this phase is not concomitant with the

cytoplasmic elevation of Ca2+ (36,78,79).

It is well known that COX-1 and COX-2 are differentially sensitive to the levels
of endogenous AA. During the immediate phase the levels of liberated AA reach a level
that is sufficiently high enough to be utilized by COX-l (76,77). However, the burst of
AA release during the immediate phase is transient and as cells progress into the delayed
phase AA release becomes limiting for prostaglandin synthesis. That COX-2 has been
shown to be preferentially capable of utilizing AA at low concentrations (1,76,77) would
reasonably explain why this COX isoform is the predominant cyclooxygenase during the
delayed phase. It has been suggested that differences in the enzymatic properties of COX-
1 and COX—2 may account for their segregated utilization of AA in cells that coexpress

both enzymes (1).

Functional Coupling Between Cyclooxygenases and Terminal Prostanoid Synthases

Cyclooxygenases are thought to exhibit distinct functional coupling with the
downstream prostanoid synthases, which include prostaglandin E2, prostaglandin D2,
prostaglandin F2a, prostacyclin, and thromboxane synthases. It is well established that in
platelet cells, which do not express COX-2, thromboxane A2 (T XA2) is synthesized due
to functional coupling between COX-1 and thromboxane synthase (TxAS), both of which
are constitutively expressed (80). Platelet-synthesized TXA2 acts in an autocrine fashion
to induce platelet aggregation and in a paracrine fashion to contract vascular smooth
muscle cells (81). The beneficial effect of low dose aspirin in reducing the risk of
myocardial infarction and other cardiovascular diseases is due to the suppression of
platelet TXA2 synthesis. Stimulation of macrophage cells with Ca2+ ionophore leads to

formation of TXA2 during the immediate phase (82). In this case, constitutively

10

expressed COX-1 is functionally linked to TxAS. Eliciting the delayed phase in
macrophage cells by stimulation with bacterial lipopolysaccharide (LPS) can also result
in TXA2 synthesis due to coupling between inducible COX-2 and TxAS (82).

The anti-thrombotic and vasodilatory activities. of prostacylin (PGI2) (83)
counteract the effects of TXA2 to achieve a balance that is crucial for maintaining
vascular homeostasis. Vascular endothelial cells synthesize prostacyclin under normal
physiological conditions (84) and during stress (85,86). Very low basal levels of PGI2 can
be detected in resting endothelial cells attributable to COX-1/PGIsynthase coupling (87).
In these cells mechanical stimuli, such as shear stress, and inflammatory stimuli, such as
LPS or TNFOL, result in coordinate induction of COX-2 expression and a dramatic rise in
the levels of PGI2 (85-87). That COX-2 is functionally coupled to PGI synthase in
vascular endothelial cells provides an explanation for the negative side effects of anti-
inflammatory COX-2 selective inhibitors on the cardiovascular system. Recent clinical
trials have shown that these inhibitors are associated with an increased risk of myocardial
infarction and stroke (88,89). Vascular smooth muscle cells, which express high basal
levels of PGI synthase (90), are reported to generate PGI2 under stress and/or
inflammatory conditions (91). In LPS- or TNFOL- stimulated macrophage cells, COX-2-
derived PGI2 is thought to be important for the resolution of the inflammatory response
through suppression of the synthesis of proinflammatory cytokines (92).

PGD synthase (PGDS) catalyzes the isomerization of the endoperoxide PGH2 to
form PGD2. Notably, PGD2 is produced in brain where it is involved in sleep induction
and regulation of pain (93,94). PGD2 is also synthesized by mast cells in response to

allergens and is believed to mediate allergic and inflammatory responses (95-97). The

11

chemical dehydration of PGD2 yields 15-deoxy-A'2’l4-PGJ2 which is thought to be the
endogenous ligand for the nuclear receptor PPARy (98-100). Acting via PPARy this
cyclopentenone prostaglandin has been shown to have anti-inflammatory activity and is
thought to be formed during the resolution phase to counter the proinflammatory effects
of other prostaglandins, including PGD2 (100-103). Two distinct forms of PGDS exist:
lipocalin-type PGDS (L-PGDS), a secreted protein that is highly expressed in the central
nervous system (CNS), and hematopoietic PGDS (H-PGDS) expressed in mast cells,
megakaryocytes, and macrophage cells (104). In a model of systemic inflammation,
PGD2 synthesis in the CNS by L-PGDS is dependent on COX-2 activity (105). In mast
cells, the generation of PGD2 upon challenge with IgF/antigen or A23187 is COX-1
dependent (40,82). However, stimulating mast cells with cytokines in the presence of
IgF/antigen not only induces COX-2 expression but also results in a six-fold increase in
PGD2 synthesis (106).

Prostaglandin E2 (PGE2) is the major prostanoid that has been implicated in the
pathophysiology of inflammation, pain, and fever. Genetic inactivation of COX-2 or
selective inhibition of the enzyme suppresses inflammation, pain, and fever concomitant
with a marked decrease in PGE2 generation (107-109). This observation is consistent
with the idea that PGE2 synthesis under these pathophysiological conditions is mainly
COX-2-derived. Three PGE synthase enzymes that catalyze the conversion of PGH2 to
PGE2 have been identified. These are cytosolic PGE synthase (cPGES), and membrane-
bound PGE synthases-1 and -2 (mPGES-l and -2). cPGES is constitutively expressed in a
wide variety of cell types (110) and its expression is usually not enhanced by

proinflammatory stimuli. This enzyme has been shown to be capable of forming PGE2

12

from COX-1-, but not COX-2-, derived PGH2, particularly during the immediate Ca2+-
dependent phase of prostaglandin synthesis (110). Of the three PGE synthases, mPGES-
1 is believed to be crucial for PGE2 generation under various pathological conditions.
mPGES-l is coordinately induced with COX-2 in cells and tissues in which COX-2-
derived PGE2 is believed to exert physiological and/or pathological actions (109). Just
like COX-2, mPGES-l expression can be inhibited by anti-inflammatory glucocorticoids
(111). Macrophage cells isolated from mPGES-l null mice generate much smaller
amounts of PGE2 relative to wild type upon stimulation with LPS (112). Just like cPLA20L
and COX-2 deficient mice (51,113), mPGES-l nulls have been reported to be resistant to
collagen-induced arthritis (114), a model of chronic inflammation. Moreover, similar to
COX-2-deficient mice (108), these mPGES-l null mice have an impaired LPS—induced
febrile response that is accompanied by a significant reduction of PGE2 levels in the
central nervous system (115). These observations suggest that functional coupling
between cPLA201, COX-2, and mPGES-l is important for stimulus-induced PGE2

formation under various stress or pathological conditions.

Regulation of Cyclooxygenase Gene Expression
Transcriptional Regulation
Although COX-1 and COX-2 exhibit a high degree of structural homology they
are the products of two distinct genes that are regulated differently. COX-l is
constitutively expressed in most mammalian tissues (1,20). The COX-1 gene is
approximately 22 kb in length, is located on chromosome 9, and is transcribed as a 2.8 kb

mRNA (20). The COX-1 gene promoter is TATA-less, CAAT—less, GC-rich, and

13

contains multiple transcription start sites (116). These promoter features are usually
characteristic of housekeeping genes that are constitutively expressed under basal
conditions. The COX-l promoter also contains three potential SP1 binding sites at -610, -
111, and -89 relative to the ATG start codon. Reporter gene assays have demonstrated
that the SP1 sites at -610 and -111 are functionally important in maintaining basal
constitutive expression of COX-1 (117). Inducible COX-1 expression has been observed
in a few cell types. For instance, differentiation of megakaryoblasts into megakaryocytes
in vitro upon stimulation with phorbol ester (PMA) is usually accompanied by at least a
5-fold increase in COX-1 mRNA and protein (118,119). Megakaryocytes are the
precursors of platelets which lack nuclei and constitutively express high basal levels of
COX-1. Recently, it has been reported that the SP1 site at -l 11 works in conjunction with
an intronic cis-acting AP-l site to facilitate PMA-induced expression of COX-1 in
megakaryoblasts (119). Inducible COX-l is also observed in vascular endothelial cells
that have been subjected to shear stress (86). Upon induction, COX—1 mRNA expression
is usually prolonged in these cells. In contrast, induction of COX-2 mRNA by shear
loading of endothelial cells is short-lived (86,120).

COX—2 was initially identified as a rnitogen-responsive immediate early gene in
fibroblast cells (121-124). Since then inducible COX-2 expression has been reported in a
wide variety of cell types, including smooth muscle cells, macrophages, endothelial cells,
synoviocytes, chondrocytes, pancreatic islet cells, and epithelial cells lining the airways,
stomach and intestine. Irnportantly, induced COX-2 is reported to be short-lived in
endothelial, smooth muscle, fibroblast, and epithelial cells (21-24). However, there are a

few exceptions to the inducible and transient expression of COX-2; constitutive

l4

expression of COX-2 has been reported in kidney, neuronal cells, and lung epithelial cells
under physiological conditions (125-127). COX-2 expression is notably absent in
erythrocytes, lymphocytes, and platelets. Various stimuli will induce COX-2 depending
on the cell type. These stimuli include mitogens, inflammatory cytokines (IL-1 B, IL-IOL,
TNFOL, and interferon-y), pro-inflammatory factors (eg. gram negative bacterial LPS),
growth factors (FGF, PDGF, VEGF, and EGF), hormones (estrogen, luteinizing
hormone, follicle stimulating hormone), oncogenes (v-Src and v-Ras), and mechanical
stimuli such as fluid shear stress (1,20). Many of these COX-2-inducing stimuli will also
activate the mobilization of AA from the cell membrane. More recently, it has been
shown that some prostaglandins, particularly PGE2, which would be formed during the
immediate phase due to COX-l activity, may activate their respective G protein coupled
receptors to induce COX-2 in a cyclic AMP- and protein kinase A-dependent fashion
(128).

The 8.3 kb COX-2 gene is located on chromosome 1 and is mainly transcribed as
a 4.5 kb mRNA (20). The last exon of the COX-2 gene encodes the 3’-untranslated
region (3’-UTR), which contains multiple copies of the ‘AUUUA’ RNA instability
elements that have a crucial role in the post-transcriptional regulation of COX-2 (20,129).
Unlike COX-l, the COX-2 promoter is replete with putative cis-acting regulatory
elements suggesting tight and complex regulation of the gene by numerous signaling
pathways. Of these cis elements, those that have been identified to have a regulatory role
in COX-2 transcription include E-Box, CAMP response element (CRE), NFKB, AP-l,
CAAT enhancer binding protein (C/EBP), SP1, serum response element (SRE), and

peroxisome proliferator response element (PPRE) (1,20,130,131). Subtle differences exist

15

in the COX-2 promoter sequence between the human and mouse gene. For instance, the
human gene has one CRE, one NFKB, and two C/EBP elements while the mouse gene
has one C/EBP, two CRE, and two NFKB sites.

Depending on the cell type and the stimulus, distinct combinations of cis-
regulatory elements will be utilized to activate COX-2 transcription. This is most evident
during septic shock, a systemic inflammatory response of the host to infection by gram
negative bacteria. The initial recognition of the LPS component of the bacterial cell wall
by monocytes and macrophages that comprise the host innate immune system will lead to
the activation of COX-2 gene transcription (130,132). LPS is recognized by toll-like
receptor 4 (TLR4) resulting in signaling cascades that lead to the simultaneous activation
of the NFKB, and the three MAP kinase (ERK, p38, and JNK) pathways (131,133). LPS-
induced MAP kinase signaling activates C/EBP, CREB, and the AP-l transcription factor
complex (130,131,133). These activated trans-acting factors localize to the nucleus and
bind the COX-2 promoter upon which they act in concert to stabilize the binding of
coactivator complexes that are responsible for recruiting the basal transcriptional
machinery (ie. RNA polymerase II and its associated factors) needed to initiate
transcription.

COX-2 inducible gene expression can be strongly retarded by glucocorticoids.
There is evidence indicating that these anti-inflammatory steroids may exert part of their
inhibitory action by suppressing COX-2 transcription. However, it is not clear how
glucocorticoids negatively regulate COX-2 transcription; moreover, the COX-2 promoter
lacks a glucocorticoid response element suggesting that the regulation may be indirect.

Several research groups have independently shown that dexamethasone, a glucocorticoid

16

analog, prevents the activation of NFKB and its subsequent nuclear localization by
upregulating the expression of IkBoc (134,135), an NFKB-interacting protein that
sequesters the transcription factor in the cytosol. Therefore, it is highly likely that
glucocorticoids may inhibit COX-2 transcription by interfering with the activation of
NFKB.

The anti-inflammatory cyclopentenone prostaglandin 15-deoxy-A'2‘14-PGJ2 (15-d
PGJ2), the dehydration product of PGD2, is known to suppress COX-2 transcription
(136,137). 15-d PGJ2 binding to its nuclear receptor PPARy inhibits PMA-mediated
induction of COX-2 in human epithelial cells by preventing binding of the AP-l trans-
acting complex to the promoter and the subsequent recruitment of the CBP/p300
coactivator complex (138). Inhibition of AP-l transactivation of COX-2 can be rescued
by a PPAR response element decoy oligonucleotide suggesting that ligand-bound PPARy

may bind the PPRE site of the COX-2 promoter to repress transcription.

Post-transcriptional regulation

COX-2 mRNA has a very short half-life. This can be explained by the presence of
multiple copies of the AUUUA motif within the 3’-UTR of COX-2 mRNA that are
known to direct mRNA decay (20,129). Deletion of these cis-acting decay motifs from
the 3’-UTR of COX-2 stabilizes the transcript (20,22). Unlike COX-2, COX-1 mRNA is
very stable and its 3’-UTR lacks these AU-rich elements (ARES) (22). The mechanism by
which ARES promote mRNA degradation in mammalian cells is not clearly understood.
There is general agreement that ARES promote deadenylation of the polyA tail which

precedes 5’ to 3’ or 3’ to 5’ exonuclease cleavage of the message (139-142). Most

17

recently, Stoecklin et al. have reported that ARE-mediated deadenylation leads to
decapping of the 5’ 7-methyl guanosine cap causing the message to be degraded S’to 3’
by the exonuclease Xml (142).

Although numerous ARE binding proteins have been shown to interact with
COX-2 ARES, the ARE binding protein(s) responsible for initiating ARE-mediated decay
of COX-2 mRNA is yet to be identified. The activation of the p38 MAPK pathway by
pro-inflammatory stimuli has been implicated in stabilizing COX-2 mRNA (143).
Stabilization of the COX-2 transcript by p38 MAPK signaling is inhibited by the
glucocorticoid dexamethasone and is believed to require a l23-nucleotide region within
the 3’-UTR and immediately downstream of the termination codon that has six ARE
copies (144). It is not clear what ARE binding proteins become activated by p38 MAPK
signaling to effect ARE-dependent COX-2 stabilization. However, various ARE binding
proteins such as CUGBP2 and HuR have been shown to bind the COX-2 3’-UTR and
stabilize the transcript. HuR-mediated COX-2 transcript stabilization has been reported in
colon cancer cells where COX-2 is aberrantly over-expressed (145). In contrast,
CUGBP2-mediated COX-2 transcript stabilization is observed in epithelial cells
undergoing apoptosis due to radiation exposure (146,147). In these cells, CUGBP2 has
also been shown to have the paradoxical role of inhibiting translation of the COX-2
transcript. This is believed to be the mechanism by which radiation-induced expression of
CUGBP2 in epithelial cells inhibits COX-2-mediated formation of PGE2 that is known to
have anti-apoptotic and mitogenic activities.

The translational silencing of COX-1 has been reported to occur in

megakaryocytes. During the PMA-induced differentiation of the megakaryocytic cell line

18

MEG-01 COX-1 mRNA is upregulated within a day of PMA treatment while an increase
in COX-1 protein is not observed for several days (118). Duquette and Laneuville have
reported a correlation between the occupancy of a putative 20-nucleotide cis element
within the COX-1 3’-UTR by a protein complex and the inhibition of COX-1 protein
synthesis (118). This complex of COX-1 mRNA binding proteins is yet to be

characterized.

Cyclooxyge’nase protein turnover

It is well established that the distinct profiles of expression of COX-1 and COX-2
observed in some mammalian cells could at least be partly attributed to differences in
regulation of the isoforms at the transcriptional and post-transcriptional levels. Although
it is not clear whether these enzymes are also regulated differently at the post-
translational level by proteolysis, their distinct patterns of expression suggest that this
may be the case. Certainly, rapid turnover of COX-2 in tissues where the enzyme is
transiently expressed may serve a significant physiological role in regulating the levels of
prostanoids whose synthesis is attributed to this COX isoform. Consistent with this idea
Rockwell and co-workers have reported observing the accumulation of COX-2 in its
native form and as polyubiquitin conjugates in HT4 neuronal-like cells treated with
inhibitors or disruptors of proteasomal degradation (148,149). In contrast, COX-1 protein
levels were unaffected by this treatment. These experimental results suggest that COX-2
may be selectively regulated by the ubiquitin-proteasome pathway, which has been

implicated in the proteolysis of intracellular proteins with short half-lives. My thesis will

19

attempt to identify and effectively address the molecular basis for the selective regulation

of COX-2 by proteolytic degradation.

Cyclooxygenase Protein Structure and Subcellular Compartmentation

Cyclooxygenases are membrane-bound, heme-containing glycoproteins that are
resident in the ER lumen and the contiguous lumen of the nuclear envelope (1 1-17). They
are believed to exist and function as homodimers with each monomeric subunit
possessing cyclooxygenase and peroxidase active sites (11,17,150). X-ray crystal
structures of mature COX-1 and COX-2 from several mammalian species have already
been determined. Their three dimensional structures exhibit a high degree of homology
which would be expected since both enzymes are ~60% identical in primary structure.

The maturation of COX-1 and COX-2 in the ER lumen involves cleavage of the
N-terminal ER targeting signal sequence, N-glycosylation at multiple sites, disulfide bond
formation, heme incorporation, membrane insertion, and dimerization. These co-
translational and post-translational modifications yield COX-1 and COX-2 mature
glycoproteins with monomeric masses of ~70 and ~72-74 kDa, respectively (151). There
are three major folding domains in mature COX-1 and -2: an epidermal growth factor
(EGF) domain, a membrane binding domain (MBD), and a globular catalytic domain
(Figs. 3 and 4). The EGF domain forms a significant portion of the dimer interface;
however, it is not clear whether it is critical for homodimerization. It has been proposed
that the EGF domain may have an important role in enabling the monotopic insertion of
the maturing cyclooxygenase into the membrane (152). The MBD is the only membrane

anchor for cyclooxygenases that has been identified thus far. It is made up of four

20

consecutive amphipathic or-helices that surround the opening of the cyclooxygenase
active site (Figs. 4 and 5). Hydrophobic and aromatic side chains protrude downwards
from these amphipathic helices and become inserted into one leaflet of the lipid bilayer.
Hydropathy analysis of the MBD of both isoforms indicates that the COX-1 MBD is
significantly more hydrophobic than that of COX-2 (16). Nonetheless, in both cases the
MBD-mediated monotopic insertion creates an unusually tight membrane association that
cannot even be disrupted with a high salt wash (15,16). Therefore, the cyclooxygenases
constitute a unique class of monotopic integral membrane binding proteins.

The cyclooxygenase catalytic domain is globular and is mostly comprised of 01-
helical secondary structure. The cyclooxygenase (COX) active site is an ~25 A long
hydrophobic channel that begins at the MBD and extends upwards into the protein
hydrophobic core. In order to be utilized as substrate, unesterified AA has to enter the
COX active site through the mouth formed by the MBD. The productive conformation of

AA at the COX active site is such that its (1) end extends into the hydrophobic core while

the A carboxyl end is stabilized at the mouth of the active site through electrostatic
interactions with the guanidium group of Arg-120 (153,154). The peroxidase (POX)
active site is solvent-exposed and contains a ferric-protoporphyrin IX heme group that is
bonded to the proximal histidine, His-388 (Fig. 5b). Both COX isoforms can be purified
as apo-enzymes indicating that the ferric heme is reversibly bound to the POX site.
Although the POX and COX active sites are spatially distinct they are functionally
interdependent. The COX oxygenation reaction requires activation by POX activity
(155,156). In turn, PGG2 formed at the COX active site diffuses to the POX active site

where it is reduced to form PGH2.

21

 

Figure 4. Ribbons diagrams of the three dimensional structures of ovine COX-1
(top) and murine COX-2 (bottom). The COX-1 and COX-2 structures are virtually
superimposable. One monomer of each enzyme is colored orange; the other monomer is
colored to show the N-terminal epidermal growth factor-like domain (green) at the dimer
interface, the four amphipathic helices that make up the membrane binding domain
(gray), and the globular catalytic domain (cyan). N-linked oligosaccharide groups (blue)
are also shown.

22

 

Figure 5a. A view from the membrane plane showing the opening into the
cyclooxygenase active site of ovine COX-1 created by the membrane binding
domain (gray). The carboxylate group of bound AA (yellow) forms a salt bridge with
the guadinium group of Arg-120 thereby stabilizing the substrate at the active site. Shown
in the background is the catalytic residue Tyr—385 which is suitably positioned to abstract
the l3pr0S hydrogen of AA.

Subtle differences exist between the primary structures of COX-1 and COX-2. The N-
terminal signal sequence of COX-1 is longer (25 residues) and more hydrophobic than
that of COX-2 (18 residues) (Fig. 3). After cleavage of the signal sequence the N-
terrninal most residue in both enzymes is usually an alanine. For convenience in
structural comparisons between the isoforms the sequence of mature COX-2 is numbered
such that it parallels the COX-1 numbering in which the start of the mature, processed
COX-2 protein has the same number as the start of the mature, processed COX-1 (12).
COX-2 numbering parallels that of COX-1 up to the C-terminus, where a 19-amino acid

insertion occurs in COX-2 six residues in from the C-terminal end. This C-terminal

23

 

Figure 5b. A view from the top showing the peroxidase active site of ovine COX-1.
Proximal His-388 coordinates the heme prosthetic group (red) at the active site by
bonding to the ferric heme. Mutagenesis studies have shown that both His-388 and distal
His-207 are important for POX catalysis. Although the role of distal His-207 in POX
catalysis has not been lucidly defined, it’s proposed to be important for deprotonation of
the hydroperoxide substrate and reprotonation of the alkyloxide ion intermediate to form
an alcohol (151). Tyr-385 and the membrane binding domain (gray) are shown to indicate
the position of the COX active site relative to the POX active site.

insertion imparts COX-2 with a consensus N—glycosylation site at Asn-594 (Fig. 3) that
has no counterpart in COX-l. As a result COX-1 is found to be N-glycosylated at three
sites, while COX-2 is glycosylated at its first three N-glycosylation sites and variably
glycosylated at Asn—594 (151).

It has been hypothesized that the C-terminal COX-2 insertion may serve as a
signal to regulate protein turnover or the subcellular compartmentation of COX-2

(1,152). In view of the short-lived expression of COX-2 protein in cellular contexts

24

where COX-1 is constitutive (21-24), it is possible that the insertion could mediate the
selective and rapid degradation of COX-2. The C-terminus of COX-2 has not been
successfully resolved in its entirety by X-ray crystallography analysis. The last resolved
residue in the highest resolution murine COX-2 crystal structure attained thus far (the
human COX-2 structure is not available in the PDB even though it has already been
solved) is Ser-596 (157), the final amino acid in the Asn-594 N-glycosylation consensus
sequence at the beginning of the insertion. This N—glycosylation site is part of a two-
tum helix (TKTATIlm) that is linked to an upstream helix by a long 15-residue loop
that includes 11 resolved residues (KGCPFI‘SFNVQ) and an unresolved 4-residue
sequence (DPQP) (Fig. 6). Asn-594 is not glycosylated in the murine structure
presumably because the amide group of its side chain is pointed upwards towards the
upstream helix that is at a distance of ~4.5 A. This space is barely sufficient to
accommodate an N-glycan group. Moreover, a disulfide bridge between Cys-569 of the
upstream helix and Cys-575 of the loop may further hinder Asn-594 glycosylation.
Therefore, in order for Asn-594 to be glycosylated a local conformational change would
have to occur, probably involving breakage of the disulfide bond and movement of the
long and inherently flexible intervening loop.

The inability to resolve the remaining 16 amino acids of the C-terminal l9-amino
acid insert in COX-2 crystal structures may be an indication that this portion is largely
disordered and lacks any degree of secondary structure. The hydroxyl side chain of Ser-
596 is solvent-exposed and is situated close to the surface of the membrane (157). The
KDEL-like C-terminal ER retention signal of COX-2 also has to be solvent-exposed and

near the membrane in order to be bound by membrane proteins involved in ER retention

25

(18,19). Therefore, the unresolved portion of the COX-2 insert may occur on the surface
of the protein close to the membrane surface. My thesis project has tested the hypothesis
that the biological role of the C-terminal 19-amino acid insertion is to selectively regulate
COX-2 protein turnover. I have also attempted to decipher the molecular mechanism by
which the COX-2 insertion may initiate and promote protein degradation. My findings

are presented and discussed in detail in Chapters 11 and III.

Ser-596

 

26

 

Figure 6. Ribbon diagrams showing two views (A and B) of the C-terminal structure
of murine COX-2. The last resolved residue in this structure is Ser-596 which is the final
amino acid of the Asn-594 consensus glycosylation site. Not surprisingly, Asa-594 is not
glycosylated most likely because its amide side chain is pointed away from the surface
and in the direction of a nearby helix. Both helices are separated by a long 15-residue
loop; the last four residues of this loop are not resolved in the structure. A disulfide bond
is formed between the thiols of Cys-569 and Cys-575 which may prevent glycosylation
of Asn-594. Futhermore, the distance between Asn-594 and Cys-569 is too small to
accommodate an N-glycan group. For Asn-594 to become glycosylated the disulfide
bond may have to be reduced to increase the flexibility of the loop. The catalytic domain
is colored green and the MBD is colored gray.

Cyclooxygenase Catalysis, Substrate Specificity, and Suicide Inactivation

The committed step in the prostanoid synthesis pathway, catalyzed by COX-1 and
COX-2, is actually a two-reaction step that involves: a) the incorporation of two
molecules of O2 to AA (5,8,11,14-eicosatetraenoic acid) to form the hydroperoxy
endoperoxide PGG2, and b) the two electron reduction of PGG2 to form the hydroxy
endoperoxide PGH2(1,3). The endoperoxide bridge of PGH2 is then acted upon by
distinct terminal prostanoid synthases to form different prostanoids. The proposed
catalytic mechanism of the cyclooxygenases is shown in Fig. 7 (156,158). Both the
peroxidase (POX) and cyclooxygenase (COX) reactions are dependent on the ferric-
protoporphyrin IX (Fe3+-IX) heme prosthetic group which is coordinated at the active site
by proximal His-388. The POX activity is required to initiate the COX activity.
Cyclooxygenase catalysis commences at the POX active site where an initiator
hydroperoxide that is yet to be identified oxidizes the heme group by removal of two
electrons yielding an oxyferryl heme radical intermediate called compound I.
Neutralization of this radical intermediate requires an electron provided by a neighboring
tyrosine residue (T yr-385) in an intramolecular reaction that leads to the formation of
intermediate 11 comprising an oxyferryl heme and a tyrosyl radical. Thereafter, the
reactive tyrosyl radical initiates COX catalysis by abstracting the 13-proS hydrogen from
AA at the COX active site yielding a radical at the 13-carbon (C-13). The AA radical
undergoes an isomerization that eventually transfers the radical to C-ll. Subsequently, an
endoperoxide bridge is formed between C-9 and C-11 after reaction with one molecule of
02. As this happens, a second molecule of O2 is added to the C-15 position forming the

hydroperoxy] PGG2.

28

 

CYCLOOXYGENASE

PEROXIDASE

 

Figure 7a. Mechanism for bis-oxygenation of AA and subsequent reduction of PGG2

NIT—W COOH
\=/<-_—/\/V

ARACHIDONIC ACID

to form PGH2. Reproduced with permission from Ref. 156.

29

POX suicide
0“ R01]

R00“ 0 0" Inactivation
Ai-h /

‘ Compound I
Resting enzyme Tyr-385 . Tyr-385

POX

 

 

Intermediate ll

Tyr-385 Tyr-385
0 0H % AA
COX
2 02
I COX suicide
AA

inactivation

Tyr-385

Figure 7b. Proposed catalytic mechanism of the COX isoforms. Cyclooxygenase
catalysis is initiated by hydroperoxide which oxidizes the ferric heme to form compound
I. This enzyme intermediate undergoes an intramolecular rearrangement of electrons
resulting in a shift of the radical from the heme to Tyr-3 85. Both radical intermediates are
very reactive and may cause modifications in the protein that lead to suicide inactivation.
The reactive tyrosyl radical at Tyr-385 will abstract the l3pr0S hydrogen of AA as
shown in Fig. 7a leaving a radical at C-13. This initiates bis-oxygenation of AA to form
PGG2 which is subsequently reduced at the POX site to form PGH2. The reduction of
PGG2 is coupled with formation of compound I and the initiation of a second catalytic
cycle.

There is no direct channel connecting the COX and POX active sites through which
PGG2 can diffuse. It has been proposed that PGG2 must exit the COX site by first
diffusing through the opening in the MBD and then traveling around the protein surface
in order to reach the POX site (156). The two electron reduction of PGG2 at the POX site

has the dual role of forming PGH2 and regenerating compound I for procession through

30

another catalytic cycle. The kinetic constants for the COX activity of COX-1 and -2 are
very similar. When AA is utilized as substrate, both enzymes have similar COX specific
activities (~25-30 umol/min of substrate per mg of enzyme) and the K1m of the purified or
microsomal (membrane-bound) forms of these enzymes is 7~ 5 11M (1).

The catalytic activities of the cyclooxygenase isoforms are distinctly regulated.
This became initially apparent when Reddy and Herschman demonstrated that stimulus-
induced prostaglandin synthesis in mitogen-stimulated 3T3 fibroblasts could almost
entirely be attributed to inducible COX-2 activity, even though COX-1 is constitutively
expressed at detectable levels in these cells (77). A year later, it was reported that the
COX activities of COX-1 and COX-2 are initiated by different levels of hydroperoxide
(155). Using varying amounts of the hydroperoxide scavenger, glutathione peroxidase, to
regulate the levels of hydroperoxide in the reaction mixture needed to initiate COX
activity, Kumalcz and Wang were able to show that both COX-1 and COX-2 COX
activities could be stimulated by nanomolar concentrations of hydroperoxide, with the
latter activity requiring 10-fold less hydroperoxide. Interestingly, they also noted that
addition of COX-2 to a COX reaction mixture containing COX-1 did not significantly
improve COX-1 activity at low hydroperoxide concentrations, suggesting that the amount
of PGG2 generated by COX-2 was not sufficient to initiate COX-1 COX activity.
Therefore, it is conceivable that in cells coexpressing both isoforms COX-2 could be the
functionally active isoform with COX-1 remaining in its latent or resting form at low
intracellular hydroperoxide levels. That COX-1 is mostly active during the immediate
phase of prostanoid synthesis would suggest the possibility of a peroxide flux that would

be sufficiently high to activate COX-1 during this phase but insufficient to activate COX-

31

1 during the delayed phase. Alternatively, the inability of COX-1 to utilize AA at low
concentrations as is observed during the delayed phase could also contribute to the
temporal segregation of COX-1 and COX-2 activities.

Although the hydroperoxide responsible for initiating COX activity in vivo is not
known, COX-1 and -2 can utilize a broad variety of hydroperoxides as substrates. In
addition to PGG2, COX-1 and -2 can reduce hydrogen peroxide, t-butyl hydroperoxide,
ethyl hydroperoxide, lS-hydroperoxy eicosatetraenoic acid (lS-HPETE), and cumene
hydroperoxide (156). Despite their broad substrate specificity, the COXs tend to prefer
secondary alkyl hydroperoxides such as 15-HPETE and PGG2 (156,159,160) and (Liu
and Smith, unpublished results). With respect to COX activity, AA, the physiological
substrate of both COX isoforms, is the most preferred substrate as determined by kinetic
measurements of oxygenation efficiency (156,161). Unlike COX-1, COX-2 can also
oxygenate a variety other substrates in vitro albeit not as efficiently as AA (156). These
substrates include, 2-arachidonyl glycerol (2-AG), an ester derivative of AA,
anandamide, an amide derivative of AA, eicosapentaenoic acid (EPA), and dihomo-y-
linolenic acid (DHLA). COX-2 is able to selectively utilize 2-AG and anandamide as
substrates because it has an active site side pocket extension that creates a larger active
site compared to that of COX-1 that can accommodate these arachidonyl derivatives (3).
This side pocket extension has been exploited by the pharmaceutical industry to design
selective COX-2 NSAIDS.

The POX and COX activities of the COXs are well known to undergo self
inactivation after an unknown number of rounds of catalytic turnover. POX and COX self

inactivation is irreversible and is also mechanism-based since it is induced by POX or

32

COX substrate, and proceeds from heme and tyrosyl radical intermediates formed during
the POX and COX reactions, respectively (3,156). However, the specific structural
changes in the holoenzyme that lead to suicide inactivation remain unknown. Mevkh et
al. noted that COX self inactivation is accompanied by dramatic changes in protein
structure as is manifested by the increased susceptibility of the inactive enzyme to trypsin
cleavage and the increased number of exposed histidine residues subject to chemical
covalent modification (162,163). POX and COX suicide inactivation have only been
observed in vitra and it is still not clear whether the COXs can undergo self-inactivation
in viva. Suicide inactivation could indeed serve as an additional control mechanism for
regulating prostanoid synthesis in viva. In this regard it would be interesting to
investigate whether the inactive forms of COX-1 and -2 are more susceptible to
degradation compared to the native forms. It is well established that misfolded or
structurally damaged proteins present in the ER can be targeted for degradation by a
process known as ER-associated degradation (164-166). It is possible that COX-2 might
selectively undergo suicide inactivation at endogenous hydroperoxide or AA levels that
would be insufficient to inactivate COX-1. If indeed the inactive protein gets degraded,
this could explain the short-lived nature of COX-2. Chapter IV of my thesis explores
inactivation-induced COX-2 protein turnover as a potential mechanism for regulating

COX-2 protein levels in viva.
Proteolytic Degradation of Intracellular Proteins

Lysosomal hydrolysis and ubiquitin-proteasome degradation are the two major

pathways for the degradation of intracellular proteins. Lysosomal protein degradation is a

33

non-selective process that is catalyzed by cysteine hydrolases called cathepsins which
function optimally at the low acidic pH of the lysosome (167). In addition to digesting
exogenous proteins that have been internalized through phagocytosis or receptor-
mediated endocytosis, lysosomes are also known to engulf and indiscriminately degrade
cytosolic and organelle proteins in bulk (168,169). This process, known as autophagy, is
usually activated for the clearance of cytosolic protein aggregates and damaged
organelles (168). Autophagy is also induced by cell starvation and can be inhibited by the
mTOR (mammalian target of rapamycin)/phosphatidylinositol-3 kinase signaling
pathway (170).

The selective degradation of intracellular proteins is largely non-lysosomal and
appears to be mediated mainly by the ubiquitin-proteasome pathway. The 268
proteasome is a ~2 MDa cytosolic and nuclear multi-protein complex that catalyzes the
selective degradation of three classes of protein substrates: a) normal short-lived proteins;
b) structurally defective proteins; and c) misfolded proteins (171). The 26S proteasome
consists of a barrel shaped, channel-like 2OS proteolytic core that is capped on each end
with a 198 ATPase subunit (172). The proteolytic core comprises three known catalytic
activities, namely trypsin-like, chymotrypsin-like, and peptidyl-glutarnyl activities (173).

For most proteasome substrates the prerequisite for degradation is
polyubiquitination, a process that involves the sequential attachment of molecules of the
highly conserved 76-amino acid ubiquitin to the e-amino group of one or more internal
lysine residues of the protein to form a long polymer of ubiquitin (174). Conjugation of
the protein substrate with a polyubiquitin tag is catalyzed by three different enzymes (Fig.

8) (174,175). Firstly, the carboxyl end of ubiquitin is conjugated to a thiol group at the

34

active site of a ubiquiting activating enzyme (E1) in an ATP-dependent manner.
Formation of the thioester linkage will activate ubiquitin for transfer to an active site
cysteine residue of ubiquitin conjugating enzyme (E2). Finally, ubiquitin protein ligase
(E3) will specifically recognize the protein substrate and aid in the transfer of the
ubiquitin moiety to the amide side chain of an exposed lysine residue. The successive
addition of ubiquitin molecules to one another will occur in a similar three-step reaction
where the carboxyl end of one ubiquitin molecule is conjugated to a specific lysine (Lys-
48) of the previously conjugated ubiquitin. The 19S subunit of the 26S proteasome will
specifically recognize the polyubiquitin tag on the substrate (174,175). Thereafter, the
polyubiquitin tag will be removed by deubiquitinating enzymes associated with the
proteasome (176). Meanwhile, the 19S subunit will employ its ATP-dependent
chaperone-like activity to unfold the protein and deliver it through the narrow pore of the
208 core for proteolytic cleavage (177). Some normally short-lived proteins like omithine
decarboxylase and cyclin D1 are known to undergo ubiquitin-independent proteasome
degradation (178,179). These proteins become targeted for proteasome destruction upon
conjugation with the molecule antizyme which has been shown to facilitate recognition of

the substrate by the 19S subunit (180,181).

Proteolytic Degradation of Endoplasmic Reticulum-associated Proteins
It has long been known that most substrates for the 26S proteasome complex are present
in the nucleus and cytosol. However, evidence has been accumulating within the last
decade to show that the 26S proteasome has a pivotal role in ER quality control involving

the elimination of misfolded, structurally damaged, and unassembled proteins that are

35

generated in the ER (l64-166,182-186). During the synthesis of membrane and secretory
proteins in the ER, these proteins undergo various co-translational modifications, such as
N-glycosylation and disulfide bond formation, that facilitate proper folding. The ER
possesses a surveillance or quality control system that ensures that newly synthesized
proteins are properly folded, N-glycosylated, and correctly assembled prior to exiting into
the secretory pathway (186). Since the native state of a protein is considered to possess its
lowest free energy (187), the surveillance system may selectively identify proteins with
non-native structure based on their lower thermodynamic stabilities. Non-native proteins
will be retained in the ER until they have achieved their native state. Otherwise, those
proteins that are irreversibly defective will be unfolded and exported to the cytosol where
they are degraded by the 26S proteasome. The pathway for the proteasomal degradation
of ER-associated proteins has been termed ER-associated degradation (ERAD) (l64-
166,182-186).

It has been recently proposed that the topology of the ERAD substrate relative to
the ER membrane and the subcellular location of its misfolded or defective lesion will
determine the manner by which the protein is selected and targeted for cytosolic
proteasomal degradation (188-190). ERAD glycoprotein substrates that are wholly
luminal or that are membrane-bound with substantially large luminal domains will be
selected for degradation in the ER lumen (189,190). The key mediators of substrate
selection in the ER lumen include the ER-resident molecular chaperone Hsp70 heavy
chain binding protein (BiP), the ER-resident lectin binding chaperone calnexin and its

homologue calreticulin, and ER-resident oxidoreductases such as ERp57 and protein

36

 

Figure 8. Mechanism for the polyubiquitination of proteins prior to their
degradation by the 26S proteasome. Ubiquitin is activated for conjugation by E1 in an
ATP-dependent manner. Activated ubiquitin is then transferred to the ubiquitin
conjugating enyzme E2. After selecting the protein substrate, E3 ligase in conjunction
with E2 will catalyze ubiquitin transfer to an exposed lysine residue on the protein.
Polyubiquitination results due to the successive addition of ubiquitin molecules to one
another where the carboxyl end of one ubiquitin molecule is conjugated to a specific
lysine (Lys-48) of the previously conjugated ubiquitin.

disulfide isomerase (PDI) (164-166,l83,186,190). Mueller et al. have recently
demonstrated that SELlL, a mammalian homologue of the yeast Hrdlp/Hrd3p E3 ligase
complex component Hrd3p is essential of the cytomegalovirus—mediated removal of
MHC class I heavy chains from the ER lumen (191). Their observations suggest that

SELlL may be a component of the ER luminal surveillance system and would therefore

37

have an important role in the recognition of lurrrinal ERAD glycoprotein substrates. The
putative role of SELlL in glycoprotein ERAD is consistent with its topology and that of
its yeast homologue Hrd3p; both are ER membrane proteins with relatively large luminal
domains (191). ER-resident chaperones will retain improperly folded glycoproteins in the
ER and assist them to fold properly. If the glycoprotein cannot be folded into its native
conformation, or if its native structure has been disrupted, it becomes selected for ERAD
(Fig. 9). The process is thought to begin with the recognition of exposed hydrophobic
sequences on the glycoprotein by BiP and other molecular chaperones. These
hydrophobic sequences are more likely to be exposed if the protein is unfolded,
misfolded, or if its native structure has been compromised. By binding to these
hydrophobic regions BiP prevents the aggregation of the misfolded protein in the ER
(192). A specific N-linked glycan on the surface of the misfolded protein is processed to
an oligosaccharide containing two molecules of N-acetyl glucosamine, nine mannoses,
and one glucose molecule (GlclMangGlcNAC2) by glucosidase I and glucosidase II. This
monoglucosylated oligosaccharide is specifically recognized by the membrane-bound
lectin calnexin and its soluble ER luminal homologue calreticulin. Upon binding, these
lectins in conjuction with BiP and the oxidoreductase ERp57 will assist in the productive
folding of the glycoprotein (193-195). If the protein becomes properly folded, cleavage of
the terminal glucose of the N-linked oligosaccharide by glucosidase U will release the
bound calnexin or calreticulin. If the protein is not properly folded, MangGlcNAC2 will be
reglucosylated by UDP-glucose glycosyltransferase (UGGT) so that the improperly
folded protein is reentered into the calnexin/calreticulin cycle. UGGT is a folding sensor

that will preferentially glucosylate ER proteins that have solvent-exposed hydrophobic

38

sequences (196,197). Repeated alterations of deglucosylation and reglucosylation will
retain the glycoprotein in the calnexin/calreticulin cycle. If the glycoprotein spends
significant time in this folding cycle, it will be considered to be terminally defective. In
this case, the 0t1,2-linked terminal mannose residue of the N-linked oligosaccharide is
irreversibly cleaved by ER 0L1,2 mannosidase I to form MangGlcNAC2 (198-200). A
Mans-binding lectin such as ER-degradation mannosidase I-like protein (EDEM) or
Yos9p will bind MangGlcNAC2 and facilitate the delivery of the terminally defective
glycoprotein to the export machinery on the ER membrane (201-205). At this juncture,
the ERAD substrate will have been unfolded by the combined action of PDI, which
breaks disulfide bonds, and other chaperones in order to be exported as an extended
polypeptide chain (206-209).

The molecular mechanism for the vectorial export or retrotranslocation of the
ERAD glycoprotein substrate across the ER membrane is yet to be clarified. The export
process requires energy and there is some evidence to indicate that the Sec6l translocon,
which is involved in ER protein import, may also be involved in protein export (206,210-
213). The recent X-ray crystal structure of the archeal SecY complex (214), a homolog of
the eukaryotic Sec6l translocon, shows that the protein-conducting channel can only
allow the passage of a molecule with a maximum diameter of 10-12 A (215). This pore
size of the translocon is barely large enough to permit the retrotranslocation of ER
polypeptides with their bulky N—glycan groups intact. However, there is evidence to
suggest that the Sec6l protein-conducting channel may be very flexible (216). Indeed,
while mediating cotranslational protein translocation the translocon pore size has been

measured to be about 40-60 A in diameter (217). Recently, a mammalian ER multi-

39

spanning membrane protein called Derlin-l was identified that shows weak resemblance
to the Derlp, a yeast transmembrane protein required for the degradation of yeast ERAD
luminal substrates. Derlin-l is proposed to be the putative retrotranslocation channel in
mammalian cells based on several studies showing its requirement for ERAD and its
ability to simultaneously associate with lurrrinal ERAD substrates and accessory cytosolic
components such as the p97(Cdc48)/del/Npl4 complex that have been shown to
facilitate export or removal of the ERAD substrates from the ER (164,218-221). After the
identification of Derlin-l two additional mammalian orthologues of Derlp, Derlin-2 and
Derlin-3, were later identified during a microarray screen for ER-stress inducible genes
(222). Derlin-2 and Derlin-3 exhibit ~75% homology and are ~30% identical to Derlin-l
(222). In the same study overexpression of Derlins-2 or -3 facilitated the degradation of a
misfolded mutant of the glycoprotein al-antitrypsin (222).

While the mechanism of ERAD is yet to be clearly defined in mammalian cells, it
has been extensively studied in yeast S. cerevisiae. Screening yeast genetic mutants that
are defective in ERAD, has enabled the functional components involved to be easily
characterized. By conducting homology database searches it was then possible to identify
homologs or functional equivalents of these yeast ERAD components in mammalian
cells. In both yeast and mammals the vectorial transport of the luminal ERAD substrate
across the putative export channel is ATP-dependent and is believed to be facilitated by a
cytosolic ATPase complex p97(Cdc48 in yeast)/de1/Npl4 that specifically recognizes
and binds polyubiquitin chains (164,220,223). As the substrate appears in the cytosol it is
polyubiquitinated by a series of membrane-bound ubiquitin conjugating enzymes such as

ch6p (UBC6) and ch7p (UBC7), and membrane-bound E3 ubiquitin ligases, that

40

include DoalO, Hrdlp (HRD), and gp78 (l64,188,l89,221,224,225). The mammalian
putative export channel Derlin-l can form membrane complexes with Hrdlp (226).
p97/del/Npl4 is recruited to the membrane by interacting with Derlin-l through an
accessory protein VIMP (p97-interacting membrane protein) (219). Recently, this
ATPase complex has also been shown to interact with the E3 ligases gp78 (227) and
Hrdlp (228) which may also contribute to its recruitment to the membrane. Membrane-
associated p97/del/Npl4 will bind the nascent polyubiquitin chain on the protein and
employ its ATPase activity to pull the polypeptide out of the ER (229). As the
glycoprotein substrate appears on the cytosolic side of the membrane it is deglycosylated
by peptidezN-glycanase (230,231). The deglycosylated substrate is then delivered to the
26S proteasome for degradation.

ER membrane proteins with defective lesions on their cytoplasmic or membrane
binding domains will be degraded by an ERAD pathway that is distinct from the one
described above (188,190). Degradation of these substrates does not appear to involve
recognition by the ER lunrinal surveillance system and is independent of their N-
glycosylation status. Thus, the targeting of ER luminal proteins for proteasomal
degradation has been designated ERAD-L to distinguish this ERAD pathway from
ERAD-C and ERAD-M which degrade ER membrane proteins with defective cytosolic
or transmembrane portions, respectively (188).

A few mammalian ER membrane-associated proteins are known to be regulated by
ERAD under physiological conditions. These include, HMG-CoA reductase, hepatic
microsomal cytochrome p450 CYP3A4, and inositol 1,4,5-triphosphate (IP3) receptor

(232-234). Even though all three membrane proteins have luminal portions, they are

41

likely to be degraded by an ERAD pathway other than ERAD-L since their turnover does
not appear to require the ER lurrrinal quality control system. 1P3 receptors, which form
channels that control Ca2+ release from the ER, are degraded in response to steady state
elevated intracellular levels of IP3 (234). ERAD of 1P3 receptors has been shown to be
mediated by ch7 and p97/de1/Npl4 (234,235). HMG-CoA reductase is a tightly
regulated enzyme that catalyzes the rate-limiting step in cholesterol biosynthesis
involving the conversion of 3-hydroxy-3-methylglutaryl CoA to mevalonate. It is a multi-
spanning ER membrane protein with a large cytosolic catalytic domain. Mevalonate or
sterols induce proteasomal degradation of HMG-CoA reductase by stimulating the
binding of the membrane protein Insig-l or Insig-2 to the membrane domain of the
reductase (236,237). Formation of the reductase-insig complex probably disrupts the
native structure of the enzyme because it is immediately polyubiquitinated by gp78 and
degraded by the proteasome (237-239). CYP3A4 undergoes substrate-induced suicide
inactivation, a process that irreversibly damages the active site heme prosthetic group and
makes the protein a substrate for ERAD (233).

Just like the cytochrome p450 CYP3A4, it is possible that COX-2 may undergo
irreversible substrate-induced suicide inactivation in viva leading to its clearance from the
cell by ERAD. However, since COX-2 is an ER-resident luminal glycoprotein, its

degradation would likely proceed through ERAD-L.

42

(MangGlcNAc2)
Unfolding

and refolding

I
K m
_ Calreticulin (GlclMangGlcNAc2)
GII ERp57

Properly folding

glycovrotcin (MangGIcNAc2)
8.3.; Misfolded
: [fig]. glycoprotein
EDEP ( ° “‘3 “m” GI and an

(MangGlcNAc2) i? ER Man I

    
 
 
 
   

  

Terminally misfolded
" glycoprotein

Proteasomal degradation

Figure 9a. Role of the calnexin/calreticulin cycle in ER quality control. N—glycosylation
usually takes place co-translationally as the elongating polypeptide is translocated into the ER
through the Sec6l translocon. Oligosaccharyltransferase (OST) transfers an intact
Glc3Man9GlcNAc2 from membrane-associated dolichyl pyrophosphate to an Asn-X-Ser/Thr site
on the polypeptide (where X is any amino acid except proline). As ER-resident molecular
chaperones such as BiP assist the Asn (N)-linked glycoprotein to properly fold, glucosidases I and
II (GI and GH) cleave off the two terminal glucose residues. If the glycoprotein is not properly
folded it will be brought into the calnexin/calreticulin cycle via the binding of
calnexin/calreticulin to the monoglucosylated N-glycan group of the protein. The heterodimer
complex of calnexin/calreticulin and ERp57 will unfold and refold the protein. The properly
folding glycoprotein will be released from the calnexin/calreticulin cycle by GII cleavage of the
remaining glucose residue. If the glycoprotein is not properly folded it will be maintained in the
cycle by the UGGT which reglucosylates the MangGlcNAc2 moiety. Misfolded glycoproteins that
remain in the cycle for too long are eventually released through the action of (11,2 ER
mannosidase I (ER Man I) which cleaves off the al.2-linked mannose to form MangGlcNAc2.
This initiates ERAD of the terminally misfolded glycoprotein. The protein is unfolded with the
help of BiP and PDI and exported out of the ER into the cytoplasm for proteasomal degradation.

43

0.00?.3300_02>0~

 

 

 

 

 

 

 

 

 

 

”. 0.0.2»: 0.02.8.
m .0 .> 052050.52?” 20:302.»: 20:032.»:
0/». e. be 00 00
H mm. ”nu.“ ”a. on
C O O
.u :03 u u
I i
_ COP—010.. 1:20.452 _|O|_IV |.|.IV .qude III..I1V— $20452 — IV — “50.3.2
Om... 0: - 2.0:. .
53.8302
D 0.0
_ a.» a."
z.
0 m: / 0......
I 0.022" \ 3.0

 

 

 

Enid e... 500000.:m 0.. 233.83 0..m00300..3...30m .3 :8 EN .530: 33...:m .308: 333.3. 00380..
.....0 0.50030030130 0:30.88 8.. 2.3.83 088.: «€003.38: .m 3000.30.03 0:8 30.83.. 34003000338
0. :8 3033830 0m .30 m” 00 0.0073330Q.0Z>0~. >m8~ 23:08.. 0.. 3.0 0..m00000..m...30 0:8 .30 388.: 3.
0..m00000.8Q.8:m..0800 AOmd. .. .0 9.000803 8 : .30:0m.:000<.083 was: AQ.0_7\.0300.0Z>0~. 3. 3.0
0033300 0.. E88888 . 3:3 .. Am. 0:: DE. 33... 0:00.00 .30 «750.088.: 8 .00 03883 .38 .30
8.300.380.8305? 8.0.0. .8308. 0.. 3.0 833.30. £30000 80.9.0 3. 0.. 8.03000 3.0 £80388? #0.:
3.0 03.39.380.885; 8.0.0. W0-m.:0omva.330: 3. CUw-m.:0000 «80035808800 Adam... 33.38.:0
3.0 388.: .3 3.0 8.0.0. .m .30 m.<00.3.08.: .0 833.30.? 3.08.30... mw 333833000 . 933: 5 0.0300 3.0
8.43.30. 0.. .N :38: 833.33. 3333000 80.330. .230 08: 23.0.. .m 3.3.0.83 3. 58:03.30 A53 .0 3.03m...
8 30 .30 00:33.83 08.: :8. .33080 mid 0.. .30 3080.78 m.v.00.3.08.:.

44

Table 1. Participants of glycoprotein ERAD-L in yeast and mammalian cells

Yeast
Lamina!
Kar2p
Cuelp
UGGT

(11,2 Man I

PDI

Htmlp

Yos9p

Membrane
Hrd3p

Sec61p

Derlp

ch6p and
ch7p

Hrdlp

Doa10

Cuelp

99
ee

Cytosolic
Cdc48

del

N pl4

Mammals
BiP
Calnexin/
Calreticulin/ERp57
UGGT

(11,2 Man I

PDI

EDEMS
(EDEM 1,2, and 3)
08-9?

SELlL

Sec61p

Derlins
(Derlin l, 2, and 3)

UBC6 and UBC7

gp78 and HRD

99
I.

??

VIMP

p97

de1
Npl4

peptidezN-glycanase

Role

prevents aggregation of ERAD substrates

glycoprotein folding

detects misfolded proteins and maintains
them in the calnexin/calreticulin cycle

initiates ERAD by irreversibly releasing
terminally misfolded glycoproteins from
the calnexin/calretican cycle

unfolds the terminally misfolded
glycoprotein in preparation for export
to cytosol

escorts the ERAD substrate to the
export/retrotranslocation channel
similar function as Htmlp/EDEM

role not clear; both have a large

luminal domain that may be involved

in selective recognition of ERAD substrates
retrotranslocation

retrotranslocation

E2 ubuquitin conjugating enzymes

E3 ubiquitin ligase

E3 ubiquitin ligase

anchors ch7p to the membrane

recniits p97 to the membrane

facilitates retrotranlocation

Cdc48/p97 cofactor

Cdc48lp97 cofactor

deglycosylates ERAD substrates
prior to their proteolytic cleavage

45

References

206,207

193-195

196,197

198-200

209,210

203-205
201,202

188,191

206,
210-213

218,219,
222

164,189

164,188,
189,224,
225

164,188,
189

189
219
164,188,
218-221
223
223

230,231

CHAPTER II

THE 19 AMINO ACID CASSETTE OF CYCLOOXYGENASE-2 MEDIATES ENTRY

OF THE PROTEIN INTO THE ER-ASSOCIATED DEGRADATION SYSTEM

Summary

Cyclooxygenase (COX) isoforms catalyze the committed step in prostaglandin
biosynthesis. The primary structures of COX-1 and COX-2 are very similar except that
COX-2 has a 19-amino acid (19-aa) segment of unknown function located just inside its
C-terminus. Here we provide evidence that the major role of the 19-aa cassette is to
mediate entry of COX—2 into the ER-associated degradation system that transports ER
proteins to the cytoplasm. COX-1 is constitutively expressed in many cells, whereas
COX-2 is expressed inducibly and transiently. In murine NIH/3T3 fibroblasts COX-2
protein is degraded by a proteasome-dependent process with a half-life (ti/2) of about 2 h
whereas COX—1 is reasonably stable (t1/2> 12 h). Similarly, COX-l expressed
heterologously in HEK293 cells is quite stable (II/2 >24 h) while COX-2 expressed
heterologously is degraded with a t1/2 of ~5 h and its degradation is blocked by
proteasome inhibitors. A deletion mutant of COX-2 was prepared lacking 18 residues of
the l9-aa cassette. This mutant retains native COX-2 enzymatic activity but, unlike native
COX-2, is stable in HEK293 cells (mp 24 h). Conversely, inserting the COX-2 cassette
near the C-terminus of COX-1 yields a mutant ins594-612 COX-1 that is unstable (tm ~3
h). Mutation of Asn-594, an N-glycosylation site at the beginning of the l9-aa cassette,
stabilizes both COX-2 and ins594-612 COX-1; nonetheless, COX mutants that have the

Asn-594 consensus glycosylation site but lack the remainder of the 19-aa cassette (ie.

46

de1597-612 COX-2 and ins594-596 COX-l) are stable. Thus, although the Asn-594
glycosylation site is necessary for COX-2 degradation, at least part of the remainder of
the 19-aa insert is also required. Finally, kifunensine, a mannosidase inhibitor that can
block entry of ER proteins into the ER-associated degradation system, retards COX-2

degradation.

Introduction

Prostanoids are an important class of lipid mediators that are synthesized in
almost all mammalian tissues. Prostanoids act in an autocrine and paracrine fashion
through G protein-coupled receptors to elicit a variety of physiological and pathological
responses (1). The committed step in the prostanoid synthesis pathway is catalyzed by the
cyclooxygenase isozymes, COX-1 and COX-2 (1,3,7,17). Within a species, COX-1 and
COX-2 exhibit ~60% amino acid sequence identity. Both isoforms are membrane-bound,
ER-resident, heme-containing glycoproteins that function as homodimers (l,3,7,17,150).
Structural and biochemical studies have shown that COX-1 and COX-2 are unusual
integral membrane proteins that associate monotopically with the luminal face of the ER
membrane and the contiguous inner membrane of the nuclear envelope
(11,12,15,l6,151).

Despite their close similarities in structure, catalytic function, and subcellular
localization, COX-1 and COX-2 differ remarkably in their profiles of expression. COX-l
is constitutively expressed in resting cells of many tissues in the absence of COX-2
( 1,20). Inducible expression of COX-2, which is usually short-lived, occurs in fibroblast,

epithelial, endothelial, macrophage, and smooth muscle cells in response to growth

47

factors, cytokines, and proinflammatory stimuli and expression is usually transient ( 1,20—
24). Differences in the patterns of expression of COX-1 and COX-2 genes have been
quite clearly delineated in serum-stimulated NIH/3T3 fibroblasts (22,124,240). COX-1
mRNA and protein levels remain unchanged before and after serum stimulation of
quiescent NIH/3T3 cells. In contrast, COX-2 mRNA and protein are barely detectable
prior to stimulation, become significantly up-regulated shortly after stimulation and then
rapidly decline to basal levels. It is well established that the different profiles of COX-1
and COX-2 expression as observed in serum-stimulated NIH/3T3 fibroblasts are partly
attributable to differences in regulation of the isoforms at the transcriptional and post-
transcriptional levels (l,20,240,241). It is still not clear the degree to which COX-1 and
COX-2 may also be regulated differently at the post translational level. However, rapid
degradation of COX-2 protein may serve a significant physiological role in regulating the
levels of prostanoids whose syntheses occur via this isoform. Previous studies have
indicated that COX-2 can be ubiquitinated and degraded by the 26S proteasome in the
cytoplasm (148,149,242), suggesting that COX-2 degradation can involve exit from the
ER via ER-associated degradation (ERAD) system(s) followed by proteolysis by the
proteasome (165,183,206,243,244).

In the present study, we establish that COX-2 protein is preferentially and rapidly
degraded in NIH/3T3 and HEK293 cells under conditions in which COX-1 is very stable.
We also confirm that even though COX-2 is an ER-resident enzyme its degradation is
proteasome-dependent. In investigating the molecular basis for the different rates of
COX-1 and COX-2 degradation, we found that the 19-amino acid segment unique to

COX-2 (Asn594-Lys612) and located six residues in from the carboxyl terminal end

48

targets the protein for rapid degradation through the ER-associated degradation (ERAD)

pathway (5) (164-166).

Experimental Procedures .

Materials. Dulbecco’s modified Eagle medium (DMEM), fetal bovine serum
(FBS), ponasterone A, tetracycline, and goat serum were from Invitrogen/Gibco. Bovine
calf serum was from Hyclone. Cycloheximide, puromycin, and bacterial
lipopolysaccharide (LPS) were obtained from Sigma-Aldrich. Kifunensine, M0132,
epoxomicin, E64, and leupeptin were purchased from Calbiochem. Endoglycosidase H
(Endo H) was purchased from Roche. Unlabeled arachidonic acid was purchased from
Cayman Chemicals, while radioactive l4C-arachidonic acid and l4C-eicosapentaenoic
acid were from American Radiolabeled Chemicals.

Construction of Pla_smids for Transfection. Recombinant ovine (ov) COX-1
cDNA and human (hu) COX-2 cDNA were subcloned into pIND (Invitrogen) and
pcDNAS/FRT/TO (Invitrogen), respectively. pIND is ecdysone-inducible while
pcDNAS/FRT/TO is tetracycline- inducible. After subcloning, the Quick-ChangeTM site-
directed mutagenesis kit (Stratagene) was used to create the following C-terminal
mutants: N594A huCOX-Z, delS95-612 huCOX-2, de1597-612 huCOX-2, and insS94-
596 ovCOX-l. The COX-1 insertion mutants, insS94-6l2 ovCOX-l, insS94-612(N 594A)
ovCOX-l, and insS94-612 ovCOX-l were created from the cDNA template for native
ovCOX-l by overlap extension PCR then subcloned into pIND using HindIII restriction
sites. Correct cDNA orientation and mutations were confirmed by sequencing. The
primers and the PCR conditions used to design the above mutants are shown in the

‘Appendix’ section.

49

Cell Culture and Transfection. NIH/3T3 fibroblasts at early passage (<6 passages)
were cultured in DMEM supplemented with 10% bovine calf serum and 100 u/ml
penicillin/streptomycin. To induce COX-2 expression, the cells were first made quiescent
by serum starvation for 48 h in DMEM containing 0.2% bovine calf serum and thereafter
treated with DMEM supplemented with 20% FBS for 4 h.

RAW 264.7 macrophage-like cells were cultured in DMEM supplemented with
10% FBS and 100 u/ml penicillin/streptomycin. To stimulate COX-2 expression, the cells
were challenged with 200 ng/ml LPS.

HEK293-derived cell lines stably expressing native or mutant COX constructs
were generated using either the tetracycline-inducible and ecdysone-inducible
mammalian expression systems (Invitrogen) according to the manufacturer’s protocol.
Constructs that were expressed under the control of a tetracycline-inducible promoter
were native huCOX-2, delS95-612 huCOX-2, N594A huCOX-2, and delS97-612
huCOX-2; those expressed under the control of the ecdysone-inducible promoter were
native ovCOX-l, ins594-612 ovCOX-l, insS94-612(N594A) ovCOX-l, and insS94-596
ovCOX-l. Stably transfected HEK293 cells were cultured in DMEM supplemented with
10% FBS, 100 u/ml penicillin/streptomycin, and the appropriate pharmacological
selective reagents. Inducible expression was achieved by a 24 h serum starvation in
serum-free medium followed by treatment with 10 ug/ml tetracycline or 10 M
ponasterone A (ecdysone analog) for the appropriate times in normal culture medium.

Protein Degradation and Drug Treatments. A cycloheximide (CI-IX) decay
experiment was performed to measure the protein stabilities of COX-1 and COX-2 in

serum-treated NIH/3T3 cells. Briefly, quiescent cells were serum-stimulated for 4 h then

50

incubated for different times with 50 M CHX in DMEM supplemented with 10% FBS.
The protein degradation experiment was performed in the absence or presence of MG132
(20 11M and 50 11M), epoxomicin (3 M and 5 uM), 25 MM leupeptin, and 25 M E64.

HEK293 cells inducibly expressing wild-type or mutant cyclooxygenase
constructs were grown to 80% confluency, serum-starved for 24 h, and then treated with
10 rig/ml tetracycline or 10 M ponasterone A in complete culture medium to induce
expression. Afterwards, the cells were incubated for various times with 50 MM
puromycin to block translation in the absence or presence of 20 MM MG132, 25 M KIF,
or 50 11M KIF.

Envzmatic Deglvcosvlation. For complete deglycosylation, HEK293 whole cell
lysates were denatured by boiling in NuPAGE SDS sample loading buffer (Invitrogen)
and then treated for at least 12 h with endoglycosidase H (Endo H) at a concentration of
0.4 mU/ul.

Western Transfer Blottig. After the appropriate treatments, NIH/3T3, RAW
264.7, and HEK293 cells were scraped into ice-cold PBS (phosphate-buffered saline), pH
7.4 containing 5 mM EDTA and a cocktail of protease inhibitors (Roche) and lysed by
sonication. RIPA lysis buffer (10 mM Tris, pH 7.2, 150 mM NaCl, SmM EDTA, 0.1%
SDS, 1.0% Triton X-100, 1% deoxycholate) containing a protease inhibitor cocktail
(Roche Applied Science) was also used for cell lysis. Protein concentrations were
determined using the BCA protein assay kit (Pierce). The NuPAGE system (Invitrogen)
was used to resolve the proteins in the whole cell lysates on a 7% tris-acetate
polyacrylamide gel. After transfer to a nitrocellulose membrane, immunoblotting was

performed with the appropriate primary antibody. Horseradish peroxidase-conjugated

51

anti-rabbit or anti-mouse IgG antibodies (Bio-Rad) were used as secondary antibodies.
Immunodetection was performed using the Western Lighting Chemilurninescent kit
(Amersham Biosciences) followed by exposure to X-ray film. Densitometry analysis was
performed using ImageQuant TL software (Amersham Biosciences).

Antibodies Specific for COX-1 or COX-2. A previously generated (13), peptide-
specific, polyclonal primary antibody for murine (mu) COX—2 against the epitope
Ser598-Ly3612 was used in the current study to detect muCOX-2 by immunoblotting. A
polyclonal antibody raised against whole recombinant ovCOX-l was used to detect
muCOX-l. Peptide-specific, polyclonal primary antibodies for ovCOX-l and huCOX-2
were synthesized by Covance Research Products against the following epitopes: Leu272-
Gln283 of ovCOX-l and ProS83-Asn594 of huCOX—2.

Immunocflofluorescence. Stably transfected HEK293 cells were grown on poly
L-lysine-coated cover slips and induced with either 10 ug/ml tetracycline or 10 M
ponasterone A for the appropriate times. After fixation in 3.7% formaldehyde and
perrneabilization with 0.2% Triton X-100, the cover slips were blocked in 0.1% Triton-
X-100 solution supplemented with 10% goat serum. Thereafter, the cover slips were
incubated with the appropriate primary antibody against either ovCOX—l or huCOX-2,
extensively washed with PBS, and incubated with Alexa Fluor 488 fluorophore-
conjugated secondary antibody (Invitrogen). After three additional extensive washes with
PBS the cover slips were dried and mounted onto glass slides using Prolong antifade kit
(Invitrogen). Microscopy was conducted on a Zeiss Atto Arc® 2 HBO 100W

fluorescence microscope at a magnification of 200X.

52

Results

Rapid, Protea_some—dependent Dggra_d§tion of cox-g. COX-1 is expressed
constitutively in murine NIH/3T3 fibroblasts, whereas COX-2 is expressed inducibly and
transiently(22,124,240). This raises the possibility that the protein stabilities of the COX
isoforms are different with COX-2 being more susceptible to degradation. A
cycloheximide (CHX) decay experiment in NIH/3T3 cells was performed to compare the
stabilities of COX-1 and COX-2 (Figs. 10a and b). COX-2 was degraded with a short
half-life (11/2) of ~2 h while COX-l was more stable and did not appear to undergo
degradation during the experiment (111/2 >12 h).

Although COX-2 expression is typically transient, a few tissues and cell types are
reported to express COX-2 in a prolonged and/or constitutive manner
(125,127,242,245,246). It is possible that COX-2 protein is stable under these conditions.
To address this, we examined COX-2 protein degradation in LPS-stimulated murine
RAW264.7 macrophage cells where COX-2 protein expression continues to increase up
to 24 h after LPS treatment. COX-2 was also degraded rapidly (t.,2~2 h) in LPS-
challenged RAW 264.7 cells (data not shown) indicating that increased COX-2 stability
is not responsible for prolonged expression of the enzyme at least in this cell type.

We performed experiments with several classes of protease inhibitors to
determine the pathway responsible for the selective degradation of COX-2 in NIH/3T3
cells. The selective 26S proteasome inhibitors, epoxomicin and M6132 (Figs. 10c and d),

significantly slowed COX-2 protein degradation at 4 and 8 h after CHX treatment

53

10A. 0 1 2 4 6 12 Timewith CHX(h)

cox-2 5-3 in... .-.... .. e ,.

cox-1 F‘Hflfifi

Actin w- “A an; L was wad

 

§§§

f—O—W“ A" fic'oxa 5
j_.fl 1cox-2j

% Protein
Remaining
s s 8

 

 

\
E‘F-

Time after CHX (h)

 

No protease
inhibitor 5 uM Epox 25 pM Leup
_0 2 4 8 2 4 8 2 4 8 TimewithCHX(h)

54

10C' N o protease

(cont’d) inhibitor 50 “M MG132 3 pM Epox
. 0_ 2 .11.. 8 2 4 >8 2_ 74 8 TimewithCHX(h)
I-Ii and un- « ~ ‘c-fl u-d' -~‘ a—a u—a «it

 

Actin W m m

D.

 

_
N
O

E

    

—0 - CHX only
+cux + M6132

\L\
“ i-——§

% Protein Remaining
s

 

 

 

0 2 4 6 8 10
Time with CHX (b)

Figure 10. Cyclooxygenase protein degradation in serum-treated NIH/3T3 cells. A,
Quiescent 3T3 cells were stimulated with 20% FBS for 4 h to induce COX-2 expression.
Cycloheximide (CHX; 50 [2M final concentration) was added to the medium and the cells
were incubated for the indicated times. Cell lysates were then prepared and analyzed by
Western blotting for COX-1, COX-2 and actin as described in the Methods section. B,
Densitometry was performed to quantify the levels of COX-1 and COX—2 which were
normalized to those of actin. Error bars denote +/- SE of the mean for three independent
experiments. C, COX—2 expression was induced in murine NIH/3T3 cells and 50 11M
CHX added to the medium as described in A. The cells were then treated for the indicated
times with or without 3 11M or 5 uM epoxomicin (Epox), 25 “M leupeptin (Leup) or 50
uM MG132. COX-2 protein levels were examined by Western blotting and densitometry.
D, Quantitative analysis of the effect of 20 11M MG132 on COX-2 protein stability in
3T3 cells. Densitometry values for each time point were from six independent
experiments performed as described above and represent the mean +/- SE. Asterisks
denote that a p value < 0.05 based on a Student’s paired t-Test.

55

while two different cysteine protease inhibitors of lysosomal degradation, leupeptin (Fig.
100) and E64 (data not shown), did not. At the 2-h time point in six independent
experiments with 20 uM M0132, there was a trend but not a statistically significant
inhibition (Fig. 10d). Accordingly, it is not clear whether a protease other than the 26S
proteasome can also be involved in COX-2 degradation.

It should be noted that both 26S proteasome inhibitors prevented the degradation
of the two alternatively glycosylated (ie. 72 and 74 kDa) forms of COX-2 (Figs. 10c and
d). Overall, our findings and those of others (148,149,242) indicate that COX-2 can be
degraded in the cytosol by the 26S proteasome. Because COX-2 is located on the luminal
surface of the ER (13,15,16), our results imply that its degradation must involve transport
across the ER membrane to the cytosol.

The Unique C-terminal l9—amino Acid Segment is Involved in COX-2
Degradation. Despite the high degree of structural similarity between the COX isoforms,
COX-2 is distinct from COX-1 in possessing a unique l9-amino acid '(19-aa) cassette that
is located six residues in from the carboxyl-terminal end (Fig. 11a). The N-terminal most
residue of the 19-aa insert is Asn-594, which is one of four functional N-glycosylation
sites of COX-2. COX-l has the other three N—glycosylation sites in common with COX-2
We used site-directed mutagenesis to test the hypothesis that the unique C-terminal 19-aa
of COX-2 affects its degradation. HEK293 cells, which do not have detectable levels of
COX-1 or COX-2, were used to stably and inducibly express native or C-terminal mutant
versions of the cyclooxygenase enzymes (Fig. 11b). Native human (hu) COX-2 inducibly

expressed in HEK293 cells was degraded with a t”; of ~5 h while native ovine (ov)

56

 

11A.

Asn68 Asn l 44

1% l L
| Signal | EGF ] MBD lGlobular-Catalytic l

Asn67 Asn l 44

Asn410

Asn4l0

RPPTEL C00“

COX-l

Am594

I ‘ 1 ERS'rELC00H
I Signal j EGF I MBD lGlobular-Catalytic - COX-2

B.

5‘14 -(~ 1 3

 

53°FSVPDPELIK'I'VTI“hisSSIRSGLDDI.‘li'lP'.l.'VI-Iaml7{S'I‘EL618
Sa"FSVPDPELIKTVTIN

 

 

EIRS‘I‘EL600
S“IE'SVIE’DPELIKIV‘IT ‘ .. iEL“8
58"FSVPDPELIK‘I V I INKS ---------------- ERSTEL‘SOZ
58°FHVPDPRQEDR“CV/E RPPI‘EL600

 

Sa”FI'IVPDPRQEDRPGVMAS8SRSGIJJDIN'P'I'VIIZJLKERPP’I‘EL519
580IE‘HVPDPRQEIJR. v . "“1VLLKERPPTEL‘19
58°FHVPDPRQEDRPGVENAS ---------------- RPP‘TEL603

native huCOX—Z
de1595-612 huCOX—2
N594A huCOX—Z
delS97—612 huCOX—2

native ovCOX—l

insS94—612 ovCOX-l
in5594—612(N594A) ovCOX—l
insS94-596 ovCOX—l

 

 

C.

native ovCOX—l
0 4 s 12 24 (h)

AAA—ya

Actin

‘3'! Illlqlla Clint—II

11,, > 24 h

delS95-612 huCOX-2
0 4 3 12 24 (h)

lira-nuflhiliuu-y

Actin

mac-vac.

t],2 > 24 h

57

 

native huCOX-2
0 2 4 s 24 (h)
r“ - d“. W 1

F

    

ins594-612 ovCOX-l

0 2 4 8 24 (h)
~a‘
fine-bud

1 ID' native huCOX—2

 

No protease inhibitor 20 1.1M MG132
0 4 812 24 4 8 12 24 (h)

‘ manure: . u. - . .‘r,"r --.-'.

 

 

. .’ 3", r; ' “_ ”,1: ',‘~_‘~."-rfl" :1! at
”IV 'm¢ .9 r a. - ,

l

M-‘ia‘JcflQ-l. ht. _ A. I‘“m‘dwl‘0.l'H-J‘MV §\l~- I. .' .'.I." «hHMI-N'fdf‘gfl-‘m'i‘lfl. 5-5. .a Vanna

insS94-612 ovCOX-l

No protease inhibitor 20 “M MG132
2 4 8 2 4 8 (h)

"‘F'"_r‘-‘-T'TQT" ‘ " ”.Tm’ "'"— ' ' ~‘ " I "" ' ‘ '_“‘-"-‘

 

 

 

 

* r We“ 2”“ ”*5 E

Fig. 11. Stability of heterologously expressed native and C-terminal cyclooxygenase
mutants in HEK293 cells. A, Alignment of major folding domains of COX-1 and COX-
2 showing consensus N-glycosylation sites and the relative position of the unique 19-
amino acid (l9-aa) insert located near the C-terminus of COX-2. Numbering for COX-l
begins with the Met at the translation start site. Numbering for COX-2 parallels the COX-
1 numbering in which the start of the mature, processed COX-2 protein has the same
number as the start of the mature, processed COX-1(12). B, Amino acid sequences of the
C-termini of cyclooxygenase constructs that were stably transfected into HEK293 cells.
The underlined sequence is the consensus N-glycosylation site at the start of the unique
COX-2 l9-aa cassette. Proteins that were expressed under the control of a tetracycline-
inducible promoter were native huCOX-Z, delS95-612 huCOX-2, N594A huCOX-2, and
delS97-612 huCOX-2; those expressed under the control of the ecdysone-inducible
promoter were native ovCOX-l, ins594-612 ovCOX-l, ins594-612(N594A) ovCOX-l,
and insS94-596 ovCOX-l. C, HEK293 cell lines stably and inducibly expressing the
constructs shown in B were grown to ~80% confluency, subjected to serum starvation for
24 h, and then treated with the appropriate inducing agent (10 ug/ml tetracycline or 10
“M ponasterone A) for 24 h (native huCOX-2, delS95-612 huCOX-Z, and insS94-612
ovCOX-l) or 12 h (native ovCOX-l). Puromycin (50 uM) was then added to block
translation and protein levels were analyzed at different times by Western blotting.
Densitometry was performed as described in the legend to Fig. l. The Western blotting
results are representative of at least three separate experiments from which protein half-
life measurements were made. D, Following serum-starvation and COX protein
induction of HEK293 cells expressing huCOX-2 and ins594-612 ovCOX-l as described
in C, 50 11M puromycin was added to inhibit translation in the presence or absence of 20
uM MG132. At the indicated times, cells were lysed and subjected to Western blotting.
The results shown for each protein are representative of two independent experiments.

58

COX-1 was not degraded at a detectable rate (tug >24 h) (Fig. 11c). A deletion mutant
delS95-612 huCOX-2 was prepared that lacks 18 amino acids of the 19-aa; additionally,
Asn-594 cannot be glycosylated in this mutant as the consensus N-glycosylation is no
longer present. This deletion mutant was stable with a t1}; comparable to that of native
ovCOX-l (>24 h) (Fig. 11c). Inserting l9-aa of huCOX-2 near the C-terminus of
ovCOX-l yielded a mutant ins594-6l2 ovCOX—l that had a relatively short “/2 of ~3 h
(Fig. 11c). Degradation of both native COX-2 and ins594—612 ovCOX-l was retarded by
treatment with the proteasome inhibitor M6132 (Fig. 11d). These observations suggest
that the l9-aa segment of COX-2 is involved in the rapid degradation of this COX
isoform, perhaps by targeting it for proteolysis by the 26S proteasome.

The native and mutant proteins used in the studies depicted in Fig. 11 and in
subsequent figures are inducibly expressed in HEK293 cells with different time courses.
To the extent possible, we performed protein degradation experiments with HEK293 cells
expressing comparable amounts of COX protein to achieve a more direct comparison of
half-lives.

The subcellular distribution of the native and mutant proteins heterologously
expressed in HEK293 cells was determined by immunofluorescent staining. Both mutant
enzymes exhibited a perinuclear and diffuse pattern characteristic of COX-1 and COX-2
(Fig. 12) (13,14,19); we also determined by Western blotting that delS95-6l2 huCOX-2
and insS94-612 ovCOX-l are present in a high speed, microsomal membrane fraction
prepared from HEK293 cells (data not shown). Dr. Masayuki Wada measured and
compared the Km and Vmx values of native murine (mu) COX-2 and delS95-612

muCOX-2 for oxygenation of arachidonic acid (Fig 13a). He also analyzed the products

59

formed from [1-14C] arachidonic acid and [1-14C] eicosapentanoic acid by both enzymes
(Fig. 13b). The kinetic constants and product distributions were the same for the native
and mutant enzymes. The specific cyclooxygenase activity of purified ins594-612
ovCOX-l was also comparable to that of the purified native ovCOX-l (38.2 and 40

pmol/min/mg, respectively; Dr. Rana Sidhu, unpublished results).

A- (I) native huCOX-Z (11) delS95-612 huCOX—2

   

(I) native ovCOX-l (II) native ovCOX-l

   

Fig. 12. Subcellular localization of native huCOX-Z, delS95-612 huCOX-Z, native
ovCOX—l and ins594-612 ovCOX-l in HEK293 cells. HEK293 cells stably expressing
the various COX variants were grown on poly L-lysine-coated cover slips and induced
with either 10 ug/ml tetracycline or 10 uM ponasterone A as described in the legend to
Fig. 11. The cells were then fixed with formaldehyde, permeabilized with Triton X-100
and incubated with an appropriate primary antibody. A, For huCOX-2 (I) and delS95-612
huCOX-2 (II) a rabbit anti-peptide antibody to residues 583—594 of huCOX-2 was used.
B, For native ovCOX-l (I and II) and insS94-612 ovCOX-l (III) a rabbit anti-peptide
antibody to residues 272-283 of ovCOX-l was used for (I) while an antibody against
whole ovCOX-l was used for (11) and (111). After washing with 1X PBS the cells were
incubated with fluorophore-conjugated anti-rabbit IgG secondary antibody, washed with
1X PBS, mounted using an anti-fade reagent, and examined by fluorescence microscopy
at a magnification of 200X.

60

 

 

 

E‘
A. €53 35,
.2 9 30‘
H 4
< a,25+
.2 E 20~
’5 E ‘ Natlve muCOX-2 - _. .
0 E 15* Km=7.6
9"“ ‘ VM=32unltslmg
(D o 10‘
X _ ‘ del595-612 muCOX-2 o ------- o
O 0 5: Km=9.9iiM
U E Vm=28 unltslmg
1 0
v ' l l I v r ,
0 so 100 150 200
[Arachidonic Acid] (pM)
B. E "i' fi'i' ” 0,! {:63
e >< 3* E at 3x
2 is es 7‘3 25" Se
, a ,

   
 

'«EPA

MS we. "i“

. “C-AA(20 iiM) _ ' "C-EPA (20 pM) .

Figure 13. Kinetic properties of purified native muCOX-2 and delS95-612 muCOX-
2. A, Hexahistidine-tagged versions of muCOX-2 and delS95-612 muCOX-2 were
expressed in Sf21 insect cells and purified. Oxygen electrode assays were performed to
compare the specific cyclooxygenase activities of the wild type and mutant enzyme. B,
Similar amounts of purified hexahistidine—tagged muCOX-2 or delS95—612 muCOX-2
(~23 mU) were incubated with 20 W [1-I4C]arachidonic acid (AA) or [1-
l4C]eicosapentaenoic acid (EPA) for 30 sec. The products were extracted, separated by
thin-layer chromatography, and visualized by autoradiography. The radioactive band
between EPA and PGH3 has chromatographic properties of the PGH3 degradation
product 17-hydroxy-5,12,15—heptadecatrienoic acid. The experiments in A and B were
carried out by Dr. Masayuki Wada.

61

Collectively, these results suggest that effects on substrate turnover or subcellular
localization do not account for the large difference in the rates of degradation between the
native and mutant enzymes.

Asn-594 Glycosylation Site is Necessary but not Sufficient to Effect COX-2
Degradation. A point mutation of the Asn-594 N—glycosylation site at the beginning of
the l9-aa segment was sufficient to extend the [”2 of the mutated COX-2 (N594A
huCOX-2) to >24 h (Fig. 14a). Consistent with this, ins594-6l2 ovCOX-l was also
stabilized by mutating the Asn-594 glycosylation site (Fig. 14a). However, a deletion
mutant delS97-612 huCOX-2 carrying the Asn-594 glycosylation site but lacking the rest
of the l9-AA insert was as stable as delS95-612 huCOX-2 (Fig. 14b). Similarly, insertion
of a consensus N-glycosylation site near the C-terminus of ovCOX-l (ins594—596
ovCOX-l) did not destabilize this COX isoform (Fig. 14b), although this mutant is at
least partly glycosylated at Asn-594 (Fig. 15a, panel I).

In evaluating the glycosylation status of Asn-594 in various forms of COX-1 and
COX-2, we compared the molecular mass of ins594-612 ovCOX-l with that of ins594-
612(N594A) ovCOX-l and the mass of native huCOX-2 to N594A huCOX-2. The
approximate mass of an N-linked oligosaccharide moiety is 2 kDa (151). Both native
huCOX-2 and insS94-612 ovCOX-l would be expected to have a total of four N-
glycosylation sites (151,247). As anticipated, the electrophoretic mobility of insS94-612
ovCOX-l was less than that of insS94-612(N594A) ovCOX-l by a difference
corresponding to about 2 kDa (Fig. 15a, panel I); moreover, deglycosylation of ins594-

612 ovCOX-l and ins594-612(N594A) ovCOX—l by Endo H increased the mobilities of

62

 

 

 

 

 

N594A huCOX-2 ins594-612(N594A) ovCOX-l
o241224(h) 0241224(h)
. 4t .‘C'f' ’T".f first“ ' fig
W““""”"i mun-inn!"
Actin ”II-5 in» we at 'v ’ all. iii—vuln- A: "I .-
t,,2 > 24 h t1,2 > 24 h
B.
delS97-612 huCOX-2 ins594-596 ovCOX-l
024824(h) 024122401)
m _ 1";
A w in- if“ r’ Q Q; i a v-
Actin “your“ “kw—e.
“Mi“ and: .....~._. .31!
t1/2>24h tm>24h

Figure 14. Stability of cyclooxygenase mutants in HEK293 cells having various
modifications of the 19 amino acid cassette. A and B, Amino acid sequences of the C-
terrnini of cyclooxygenase constructs that were stably transfected in HEK293 cells are
shown in Fig. 118. Proteins that were expressed under the control of the tetracycline-
inducible promoter are N594A huCOX-2 and delS97-612 huCOX—2; those expressed
under the control of the ecdysone-inducible promoter were ins594-612(N594A) ovCOX-
l and ins594-596 ovCOX-l. Expression was induced for 12 h (N594A huCOX-2 and
delS97-612 huCOX-2) or 24 h (ins594-612(N594A) ovCOX-l and ins594-596 ovCOX-
1). Thereafter, the time course of degradation of the proteins was determined as described
in the legend to Fig. 11. The results shown in A are representative of at least three
independent experiments from which protein half-life measurements were made. The
results shown in B are representative of two independent experiments.

63

both mutants such that they had apparently identical molecular masses (Fig. 15a, panel
11). These results suggest that the difference in molecular mass between the two COX—l
insertion mutants results from glycosylation of Asn-594.

In contrast to the results obtained with the COX-1 mutants, we found no
difference between the electrophoretic mobilities of native huCOX-2 and N594A
huCOX-2 expressed in HEK293 cells (Fig. 15b, panel I). This is somewhat similar to
what is seen with murine COX-2 expressed in cos-1 cells (151) where the major bands
for native muCOX-2 and N594Q muCOX—2 have the same electrophoretic mobilities;
however in the case of muCOX-2 expressed in cos-1 cells, a more slowly moving band
was also observed. A second, less mobile band was observed with huCOX-2, but not
N594A huCOX-2, when HEK293 cells were treated with KIF, an al,2 ER mannosidase I
inhibitor (Fig. 15b, panel H). Moreover, Endo H treatment led to deglycosylated forms of
native COX-2 and N594A COX-2 that had the same molecular mass (Fig. 15b, panel 11).
In addition to causing the appearance of a second, alternatively glycosylated form of
huCOX-2, KIF treatment of HEK293 cells expressing native huCOX-2 inhibits COX-2
degradation (Fig. 15b, panel II). Degradation of ins594-612 ovCOX-l is also inhibited by
KIF (Fig. 15c). This is a diagnostic feature of N-glycosylated proteins whose degradation
involves transport to the cytoplasm by the ERAD system (199,200,248—251). These
results are consistent with those obtained with 26S proteasome inhibitors in indicating
that COX—2 degradation involves entry into the ERAD system prior to degradation by the

26S proteasome.

64

15A. (0

l 2 3 4
l. native ovCOX—l

2. insS94-596 ovCOX-l

3. ins594-612 ovCOX-l

4. ins594-612 (N 594A) ovCOX-l

(lo
4.

I “12 :3"
fl .

w

insS94-612 ovCOX—l
insS94-612(N 594A) ovCOX-l
Endo H cleavage of l
Endo H cleavage of 2

95”.“?

1. native huCOX—2
2. N594A huCOX-Z

 

 

(ll)
native huCOX-z N594A huCOX-Z
l 2 3 4 1 2 3 4
3 an H c- .-

1. tet-induced

2. tet-induced + 12 h KIF
3. Endo H cleavage of l
4. Endo H cleavage of 2

65

15C.

 

25 uM KIF
0 2 4 8 24 2 4 8 24 Time with puromycin (h)
"'7‘?“ .42... a A: A A
Actin ‘l‘?’ " 3" a “a “‘3 Wig-taut 1‘

Figure 15. Glycosylation of ins594-612 ovCOX-l and native huCOX—Z at Asn-594. A
(I) Native ovCOX-l (1), ins594-596 ovCOX-l (2), insS94-612 ovCOX-l (3), or ins594-
612(N594A) ovCOX-l (4) expression was induced with 10 M ponasterone A in
HEK293 cells for 24 h. (II) After protein induction of insS94-612 ovCOX-l and in3594-
612(N594A) ovCOX-l cell lysates were prepared and boiled in SDS sample loading
buffer for 10 min and then treated with or without Endo H (0.4 mU/ul) for at least 12 h.
Proteins were resolved by SDS-PAGE on a 7% tris-acetate polyacrylamide gel and
subjected to Western blotting. B (I), native huCOX-2 (1) or N594A huCOX-2 (2)
expression was induced with 10 ug/ml tetracycline in HEK293 cells for 24 h and 3 h,
respectively. ( II ), after protein induction as in (I) cells were rinsed in PBS and incubated
for an additional 12 h in normal culture medium with or without 50 M KIF. Thereafter,
cell lysates were prepared and boiled in SDS sample loading buffer for 10 min and then
treated with or without Endo H (0.4 mU/ul) for at least 12 h. Proteins were resolved and
subjected to Western blotting as in A. Arrows indicate differences in molecular mass due
to N-glycosylation. C. insS94-612 ovCOX-l was induced in HEK293 cells as described
in the legend to Fig. 2. The cells were then incubated with puromycin (50 M) in the
presence or absence of KIF (25 M) for the indicated times. Cell lysates were prepared
and subjected to SDS-PAGE and Western blotting as described in the legend to Fig. 11.
The results shown are representative of two independent experiments.

66

Discussion

Substitution of the COX-1 coding region for COX-2 in transgenic mice partially
rescues the typical phenotype of COX-2 null mice (252). Thus, a key difference between
COX-1 and COX-2 is in their patterns of expression and not simply in their enzymatic
properties. The difference in the profiles of COX-1 and COX-2 protein expression are at
least partly attributable to differences in the regulation of transcription of the two genes
(reviewed in (1,20,241)). In general, COX-1 gene expression is constitutive, and COX-2
expression is inducible. In addition, COX-2 mRNA with its repetitive AUUUA
degradation sequences is subject to post-transcriptional control (145,253,254). In this
report, we confirm and extend previous findings indicating that COX-l is a stable protein,
whereas COX-2 is degraded relatively rapidly, and that COX-2 can be ubiquitinated and
degraded by the 26S proteasome (148,149,242). COX-2 is localized to the luminal
surface of the ER (13-16), and so to be degraded, it must first be transported to the
cytosol from the ER.

In the studies reported here, we have focused on the molecular basis for the
relatively rapid rate of COX-2 degradation when compared with that of COX-1.
Specifically, we evaluated the 19-aa cassette (Asn-594—Lys-612) that is unique to COX-2
for its potential role in degradation. This 19-aa insert is located 6 residues in from the
COX-2 C terminus. It has only one recognizable consensus sequence, an N-glycosylation
sequence involving Asn-594—Ser—596. We have demonstrated that the 19-aa cassette of
COX-2, with an intact Asn-594 N-glycosylation site, targets COX-2 for proteasomal
degradation. This is based on the following observations. a) A deletion mutant of

huCOX-2 lacking 18 amino acids of the l9-aa insert (delS95—612 huCOX-2) and a

67

functional N—glycosylation site at Asn-594 has native COX activity and the same
subcellular location as native enzyme but is refractory to protein degradation. b) Point
mutation of the Asn-594 N-glycosylation site (N594A huCOX-2) also stabilizes COX-2.
c) Inserting the COX-2 l9-aa cassette at the C terminus of ovCOX-l (insS94—612
ovCOX-1)destabilizes ovCOX-l. d) Point mutation of Asn-594 in insS94—612 ovCOX-l
stabilizes this insertion mutant. e) The degradation of both native huCOX-2 and insS94—
612 ovCOX-l can be retarded by M6132, a 268 proteasome inhibitor. Although the Asn-
594 N-glycosylation site is required for degradation, we found evidence that all or part of
the remainder of the l9-aa insert is also necessary.

Our results suggest that glycosylation of Asn-594 of COX—2 is critical for
initiating entry of the enzyme into the ERAD system for transport to the cytoplasm. The
ERAD system is thought to involve unfolding of N-glycosylated proteins in the ER lumen
and ATP-dependent retrograde transport across the ER membrane, and frequently, this is
followed by ubiquitination of the transported protein by ubiquitin-prOtein isopeptide (E3)
ligases associated with the cytosolic face of the ER membrane and subsequent
degradation by the 26 S proteasome (165,183,206,243,244). The ERAD system is usually
regarded as a quality control pathway for the clearance of misfolded, unassembled, or
toxic proteins from the ER lumen (165,206,243,244,255). It has also been implicated in
the degradation of native ER-associated proteins such as hydroxymethylglutaryl-CoA
reductase, hepatic microsomal cytochrome P450 CYP3A4, and inositol 1,4,5-triphosphate
receptor (232-234), all of which are ER transmembrane proteins. An ER luminal
glyc0protein is selected for the ERAD pathway by a complex process that involves

enzymatic processing of specific N-linked oligosaccharide groups on the surface of

68

glycoproteins (166,204,255,256). This requires the action of 0L1,2 ER mannosidase I
whose inhibition by KIF leads to retention and stabilization of glycoproteins in the ER
(199,200,248-251). In the current study, we show that KIF retards the degradation of
COX-2 and a COX-l mutant containing the 19-aa cassette of COX-2.

Spear and Ng (256) have recently shown that in a multiply glycosylated
glycoprotein a single, specific N-linked oligosaccharide group serves as the determinant
for entry into the ERAD pathway. In the present study, we have provided evidence
suggesting that only one of the four N-glycosylation sites of COX-2 (Le. Asn-594) is
involved in COX-2 degradation; however, degradation also requires at least some amino
acids downstream of the consensus N-glycosylation sequence. We presume that a
carbohydrate moiety linked to Asn-594 along with adjacent amino acids of the 19-aa
insert interact with the complex of proteins that provide entry into the ERAD system.
Very little of the l9-aa cassette can be resolved in the COX-2 crystal structure (12), and
therefore, the l9-aa insert is thought to lack secondary structure. The last resolved residue
in the murine COX-2 crystal structure is Ser—596, the final amino acid in the N-
glycosylation consensus sequence. Ser-596 is solvent-exposed and appears to be oriented
in such a way that its C terminus would point toward the lipid bilayer. The COX-2
KDEL-like ER retention motif at the C terminus (19) also has to be on the protein surface
to be recognized (18). Thus, the 19-aa cassette of COX-2 is likely to be solvent-exposed
and situated close to the membrane surface.

In summary, we have identified a 19-aa cassette having a consensus N-
glycosylation site at its N terminus that can target a protein to the ERAD system.

Scanning mutagenesis should be helpful in identifying the critical residues in the l9-aa

69

cassette that work in conjunction with the Asn-594 N-glycosylation site to target COX-2
for degradation. It will also be important to characterize the oligosaccharide group linked

to Asn-594.

Acknowledgments
We thank Dr. Anne Vojtek and Jennifer Taylor for their kind help with
immunocytofluorescence. We are also grateful to Dr. Billy Tsai for his thoughtful
suggestions relevant to this work. This work was supported in part by NIH Grant

GM68848.

7O

CHAPTER III

ASN-594 GLYCOSYLATION AND PROTEIN DEGRADATION OF
COX-2 IS REGULATED BY ITS C-TERMINAL 27 AMINO ACID

INSTABILITY ELEMENT

Summary

The C-terminal 19 amino acid cassette (19-aa) of COX-2 is required for the entry
of the protein into the ER-associated degradation (ERAD) system. We have previously
shown that the Asn—594 glycosylation site at the start of the 19—aa cassette is necessary,
but not sufficient, to mediate COX—2 ERAD. Here, we demonstrate that the C-terminal 16
amino acid portion of l9-aa is also essential for COX-2 degradation. We also show a
positive correlation between the extent of glycosylation at Asn-594 and the overall
degradation of COX-2. Evidence is provided to demonstrate that the 19-aa cassette is part
of a larger 27 amino acid sequence that dictates the intracellular stability of COX-2 by
regulating glycosylation at Asn-594. Analysis of the region immediately upstream of
Asn—594 suggests that an 8-residue sequence with OL-helical structure impedes Asn-594
glycosylation, and in order for COX-2 to become glycosylated at Asn-594 there must be
a change in the structure or conformation of this a—helix. We propose that this is elicited
by the C-terminal l6-residue sequence of l9-aa. The upstream 8-residue sequence also
appears to play a role in the processing of the N-glycan group at Asn-594 that preceeds

translocation of COX-2 to the cytoplasm for proteasomal degradation. Therefore, the

71

interplay of the upstream 8-residue segment and the C-terminal 16 amino acid portion of
l9-aa to regulate Asn—594 glycosylation and its N-glycan processing may control the

timing and the extent of COX-2 protein degradation.

Introduction

Cyclooxygenases-l and -2 (COX-1 and COX-2) catalyze the key committed step
in prostanoid synthesis (l,3,7,17). These enzymes are multiply N-glycosylated ER
resident proteins that exist as homodimers and exhibit ~60% identity in primary structure
(l,3,7,17,150). They are also integral membrane proteins that monotopically insert into
the lipid bilayer so that they are largely compartmentalized in the ER lumen or the
contiguous lumen of the nuclear envelope (14-16). The mature forms of the COX
isoforms are very similar in structure except that COX-2 has a unique l9-residue
insertion (l9-aa) near its C-terminus. Previously, we have shown that the 19-aa is
essential for ER-associated degradation (ERAD) of COX-2 by the 26S proteasome (257).
We have also demonstrated that the Asn-594 glycosylation site at the N-terminus of 19-
aa is required for COX-2 ERAD (257). Immunoblotting of a variety of cell lines for
COX-2 usually detects 72 and 74 kDa variably glycosylated forms of the enzyme due at
least in part to alternative glycosylation at Asn-594 (148,151,258). Inhibition of co-
translational N-glycosylation of COX-2 with tunicamycin at its four consensus
glycosylation sites eliminates the cyclooxygenase (COX) and peroxidase (POX) activities
of the enzyme (151). However, point mutation of the Asn-594 glycosylation site actually

enhances COX—2 COX specific activity (151) most likely due to the increased stability of

72

the mutant enzyme (257). These results imply that while the Asn-594 glycosylation site is
dispensable for proper co—translational protein folding, it is critical for the degradation of
the enzyme.

In HEK293 cells stably expressing recombinant native human COX-2 the Asn-
594 glycosylated form of the protein can only be detected upon addition of kifunensine,
an ER (X-l,2 mannosidase I inhibitor (257). This is an indication that Asn-594
glycosylated COX-2 is rapidly degraded by ERAD. It is well established that the ER
possesses a quality control or surveillance system that selectively identifies glycoproteins
with non-native structure so that they are retained in the ER long enough to achieve their
native state (l64—l66,182—186). Glycoproteins that are irreversibly defective are degraded
by ERAD, a process that involves unfolding of the protein and its export to the cytosol
where it undergoes proteolysis by the 26S proteasome (164-166,182-186,188,189). COX—
2 is the only ER luminal integral membrane protein that is known to be degraded via
ERAD in its native or properly folded form. This raises the question as to why and how a
properly folded glycoprotein is targeted for ERAD by the ER protein quality control
system. We believe that a thorough structural and biochemical analysis of the C-terminal
19-aa of COX-2 may provide an answer to this conundrum.

Very little of the l9-aa can be resolved in the highest resolution mouse COX-2 X-
ray crystal structure attained thus far (157). The last resolved residue in this 2.4 A
structure is Ser-596, the final amino acid in the Asn-594 N-glycosylation consensus
sequence at the N-terminal end of 19-aa. The inability to resolve the remaining 16
residues of the 19-aa by X-ray crystallography may be an indication that this segment is

largely disordered and lacks any degree of secondary structure. The Asn-594

73

glycosylation site is at the end of a two-tum helix that is linked to an upstream helix by a
long lS-residue loop (Fig. 6). Asn-594 is not glycosylated in this structure presumably
because the amide of its side chain is pointed upwards toward the upstream helix that is a
distance of ~4.5 A. Since the space between the helices is barely sufficient to
accommodate an N-glycan group, a local conformational change, probably mediated by
movement of the long and inherently flexible intervening loop, would have to occur to
expose Asn-594 for N-glycosylation.

In our previous study, a HEK293 heterologous system was utilized to uncover the
role of 19-aa in directing COX-2 protein degradation. In a different study, Dr. Masayuki
Wada prepared fibroblast primary cultures from skin tissue of wild-type mice or a knock-
in mouse carrying the mutant delS95-612 murine (mu) COX-2 in place of the wild-type
gene (W ada and Smith, unpublished results). To induce COX-2 expression, these primary
cultures were first made quiescent by 24 h serum starvation followed by challenge with
fetal bovine serum. NIH/3T3 fibroblasts that have been growth—arrested by serum
starvation lack detectable COX-2 expression but constitutively express COX-1 (22). By
immunoblotting analysis, Dr. Masayuki Wada made the interesting finding that while
COX-l was the only isoform that could be detected in growth-arrested primary cultures
of wild-type mice, quiescent cultures prepared from delS95-612 COX-2 mutant (-/-) mice
constitutively expressed both COX isoforms. Fibroblasts from heterozygous (+/-) mice
also constitutively expressed COX-2; however, the COX-2 protein levels were lower
compared to the -l- cells indicating that protein expression was gene dose-dependent.
Serum-stimulation of these quiescent primary cells resulted in the time-dependent

accumulation of COX-2 protein in ~/- cells in deep contrast to the more transient

74

expression of COX-2 in serum-treated wild-type cells. This is a confirmation that the 19-
aa of COX-2 is essential for the degradation of the protein. More importantly, these
results strongly argue that COX-2 ERAD that is mediated by l9-aa could represent a
significant physiological mechanism for the regulation of COX-2 gene expression.

In the present study we set out to examine what specific features of the C-
terrninal l9-aa, in addition to Asn-594 glycosylation, are critical for targeting the enzyme
to the ERAD system. The role of the helical region upstream of the Asn-594
glycosylation site in regulating Asn-594 glycosylation and COX-2 protein degradation
was also investigated. Our findings have enabled us to identify a C-terminal 27 instability
element of COX-2 that regulates the glycosylation of the enzyme at Asn-594, and

subsequently, controls the timing and extent of protein degradation.

Experimental Procedures

Materials. Dulbecco’s modified Eagle’s medium (DMEM), fetal bovine serum
(FBS), ponasterone A, and tetracycline were from Invitrogen. Bovine calf serum was
from Hyclone. Cycloheximide, puromycin, and bacterial lipopolysaccharide (LPS) were
obtained from Sigma. Kifunensine and MG132 were purchased from Calbiochem.
Endoglycosidase H (Endo H) was purchased from Roche Applied Science.

gastruction of Flam for Trflfection. Recombinant human (hu) COX-2
cDNA was subcloned into the tetracycline-inducible vector pcDNAS/FRT/TO
(Invitrogen). After subcloning, the QuickChangeTM site-directed mutagenesis kit
(Strategene) was used to create the following C-terminal mutants: del602-612 huCOX-2,

del607-612 huCOX—Z, P607A T608A huCOX-2, V609A huCOX-Z, L610A huCOX-Z,

75

L611A huCOX-2, K612A huCOX-2, V591P T5920 huCOX-2, ovine (ov) COX-l
UPSTRM8 huCOX-2, and murine (mu) COX-2 UPSTRM8 huCOX-2. The COX-1
insertion mutant insS94-60l ovCOX-l was created by QuikChangeTM site-directed
mutagenesis using insS94-612 ovCOX-l as template. In3594-612 ovCOX-l mutant was
previously made by overlap extension PCR and subcloned into the ecdysone-inducible
pIND vector (Invitrogen). The COX-2 mutants, scrambled insert (Sins) 595-612 huCOX-
2 and Sins597-6l2 huCOX-2 were created from the cDNA template for native huCOX-2
by overlap extension PCR and then subcloned into pcDNAS/FRT/TO using BamHI and
XhoI sites. Correct cDN A orientation and mutations were confirmed by sequencing. The
primers and the PCR conditions used to design the above mutants are shown in the
‘Appendix’ section.

Cell Culture and Transfection. NIH/3T 3 fibroblasts at early passage (<6 passages)
were cultured in DMEM supplemented with 10% bovine calf serum and 100 u/ml
penicillin/streptomycin. To induce COX—2 expression, the cells were first made quiescent
by serum starvation for 48 h in DMEM containing 0.2% bovine calf serum and thereafter
treated with DMEM supplemented with 20% FBS for 4 h.

RAW 264.7 macrophage-like cells were cultured in DMEM supplemented with
10% FBS and 100 u/ml penicillin/streptomycin. To stimulate COX-2 expression, the cells
were challenged with 200 ng/ml LPS for 12 h.

HEK293-derived cell lines stably expressing native or mutant COX constructs
were generated using either the tetracycline-inducible and ecdysone-inducible
mammalian expression systems (Invitrogen) according to the manufacturer’s protocol.

Constructs that were expressed under the control of a tetracycline-inducible promoter

76

were native huCOX-2, del602-612 huCOX-2, del607-6l2 huCOX-2, P607A T608A
huCOX-2, V609A huCOX-Z, L610A huCOX-2, L611A huCOX-2, K612A huCOX-2,
V591P T5920 huCOX-2, ovCOX—l UPSTRM8 huCOX-2, muCOX-2 UPSTRM8
huCOX-2, SinsS95-612 huCOX-2, and Sins597-612 huCOX—2. The insS94-60l ovCOX-
1 construct was expressed under the control of the ecdysone-inducible promoter. Stably
transfected HEK293 cells were cultured in DMEM supplemented with 10% FBS, 100
u/ml penicillin/streptomycin, and the appropriate pharmacological selective reagents.
Inducible expression was achieved by by treatment with 10 jig/ml tetracycline or 10 MM
ponasterone A (ecdysone analog) for 24 h in normal culture medium.

Protein Degradation and Drug Treatments. Quiescent 3T3 cells were serum-
stimulated for 4 h then treated with 50 M cycloheximide (CHX) for different times in
the presence or absence of 25 M kifunensine (KIF). RAW264.7 macrophage cells were
challenged with 200 ng/ml LPS for 12 h then treated with 50 M CHX in the presence or
absence of treatment with 50 M KIF. HEK293 cells stably and inducibly expressing
wild-type or mutant cyclooxygenase constructs were grown to ~80% confluency, serum-
starved for 24 h, and then treated with 10 ug/ml tetracycline or 10 uM ponasterone A in
complete culture medium to induce expression. Afterwards, the cells were incubated for
various times with 50 MM puromycin to block translation in the absence or presence of 20
MM M0132 or 25 “M KIF. Alternatively, stable transfectants of HEK293 cells were
grown to ~80% confluency, treated with 10 ug/ml tetracycline for 24 h to induce
expression, and thereafter treated for 12 h with 25 M KIF.

Enyzmatic Deglycosylation. For complete deglycosylation, HEK293 whole cell

lysates were denatured by boiling in NuPAGE SDS sample loading buffer (Invitrogen)

77

and then treated for at least 12 h with endoglycosidase H (Endo H) at a concentration of
0.4 mU/ul.

Western Transfer Blotting. After the appropriate treatments, NIH/3T3, RAW
264.7, and HEK293 cells were scraped into ice-cold PBS (phosphate-buffered Saline), pH
7.4 containing 5 mM EDTA and a cocktail of protease inhibitors (Roche) and lysed by
sonication. RIPA lysis buffer (10 mM Tris, pH 7.2, 150 mM NaCl, SmM EDTA, 0.1%
SDS, 1.0% Triton X-100, 1% deoxycholate) containing a protease inhibitor cocktail
(Roche Applied Science) was also used for cell lysis. Protein concentrations were
determined using the BCA protein assay kit (Pierce). The NuPAGE system (Invitrogen)
was used to resolve the proteins in the whole cell lysates on a 7% his-acetate
polyacrylamide gel. After transfer to a nitrocellulose membrane, immunoblotting was
performed with the appropriate primary antibody. Horseradish peroxidase-conjugated
anti-rabbit or anti-mouse IgG antibodies (Bio-Rad) were used as secondary antibodies.
Immunodetection was performed using the Western Lighting Chemiluminescent kit
(Amersham Biosciences) followed by exposure to X-ray film. Densitometry analysis was
performed using Quantity One software (Bio-Rad).

Antibodies for Western Analysis. A previously generated (13), peptide-specific,
polyclonal primary antibody for murine (mu) COX-2 against the epitope Ser598—Lys612
was used in the current study to detect muCOX-2 expressed in NIH/3T3 and RAW264.7
cells. Peptide-specific, polyclonal primary antibodies for ovCOX-l and huCOX-2 were
synthesized by Covance Research Products against the following epitopes: Leu272-
01n283 of ovCOX-l and Pr0583-Asn594 of huCOX-2. A polyclonal antibody raised

against whole muCOX-2, which also specifically detects native huCOX-2, was used to

78

immunoblot for the mutants V591P T5920 huCOX-2, ovCOX-l UPSTRM8 huCOX-2,

and muCOX-2 UPSTRM8 huCOX-2.

Results

The C-terminal region of 19-& is important for COX-2 degradation. It has
previously been shown that the Asn-594 glycosylation site is necessary, but not
sufficient, for enabling the proteasomal degradation of COX-2 (257). Therefore, an
additional portion of l9-aa is required to mediate COX-2 degradation. We embarked on
determining which amino acids of the C-terminal l9-aa are needed for this process.
Initially, two C-terminal deletion mutants of human (hu) COX-2 were prepared, namely
del602-612 huCOX-2 and del607-612 huCOX-2 (Fig. 16a, panel I). We also prepared an
insertion mutant of ovine (ov) COX-l, ins594-601 ovCOX-l, that carried an identical
truncated region of l9-aa as del602-612 huCOX-2 (Fig. 16a, panel I). These mutants
were stably transfected into HEK293 cells and their degradation profiles examined. All
three mutants, including del607-612 huCOX-2, were stable and did not degrade within
the 24 h time frame of the experiment (Fig. 16a, panel 11). These results indicated that
removal of the last six residues of 19-aa was sufficient to eliminate COX-2 degradation.
Preventing COX-2 degradation by trimming the l9-aa raised the possibility that a correct
length for this insertion, rather than its primary structure, was needed to mediate protein
turnover. To eliminate this possibility, we prepared two mutants of COX-2 that carried
scrambled versions of the 19-aa (Fig. 16b, panel I). Sins597-612 huCOX-2 has an intact
Asn-594 glycosylation site but the rest of the l9-aa has been scrambled. SinsS95-612

huCOX-2 is not only mutated at Asn-594, but it also has a randomized version of the rest

79

of the l9-AA. After stably transfecting these scrambled mutants into HEK293 cells their
protein stabilities were analyzed. Both mutants were much more stable than native
huCOX-2 because they degraded very slowly with a half-life (tug) > 24 h (Fig. 16b,
panels 11 and 111). Overall, these results suggest that there is primary structure information
within the 19-aa cassette, in particular its last six residues, that is essential for dictating
COX-2 protein degradation.

To identify which of the last six amino acids of 19-aa were important for enabling
COX-2 degradation we performed alanine scanning mutagenesis of this region (Fig. 17a).
The mutant P607A T608A huCOX-2 had a tug of ~8 h which was not far removed from
that of the native huCOX-2 (ti/2 ~5 h) (Figs. 17b and c). Its degradation kinetics was also
similar to that of the native enzyme (Fig. 17c). The V609A huCOX-2 mutant also
degraded with a Ill/2 of ~8 h (Figs. 17b and c). However, careful examination of its overall
degradation profile revealed that V609A huCOX-2 was degraded less efficiently than
both the native and P607A T608A huCOX-2 (Fig 17c). There seemed to be a significant
loss of V609A huCOX-2 protein at the earlier time points after inhibiting protein
translation. Thereafter, at the 12 h and 24 h time points, this mutant appeared to be
relatively stable compared to native and P607A T608A huCOX-2 (Figs. 17b and c).
Although the mutations L610A huCOX-2, L611A huCOX-2, and K612A huCOX-2 did
not completely prevent COX-2 degradation, they significantly stabilized the enzyme; the
11/2 for the degradation of all three mutants was >24 h (Figs. 17b and c). Collectively,
these results point to a three amino acid sequence at the C-terminal end of 19-aa, namely

Leu-6lO, Len-611, and Lys-612, as being essential for enabling COX-2 degradation.

80

16A. (1)

 

 

 

 

 

 

 

 

5BOFSVPDPELIKTVTInnisssasC4Llanrnp-r'vnracERSTEL618 native huCOX—2
53°FSVPDPELIKTVTINA_ ERSTEL607 del602—612 hucox—2
58°FSVPDPELIKTVTINASSSR8GLDDIN —————— ERSTELGlz de1607-612 hucox-z
SBOFHVPDPRQEDRPGV ERPPTEL5°° native ovCOX—l
58°FHVPDPRQEDRPGVNASSSRSGLDDINPTV'LLKERPPTELG19 insS94—612 ovCOX-l
SBOFHVPDPRQEDRPGVNA ERPPTEL6°8 insS94—601 ovCOX—l
A. (II)
del602-612 huCOX-Z del607—612 huCOX-Z
0 2 4 12 24 (h) 0 2 4 12 24 (h)

‘ r 7i ' ’T” " , ’T! if"? ”7‘ h ” " " —_” ' 71‘

z: . a

in " ' I ‘N ~'.‘. . . .. . ,

Actin? fl-n-nq'v

I

Actinii v" ”A " * Q

rm > 24 h rm > 24 h

iu5594-601 ovCOX-l
o 2 4 12 24 (h)

 

Actin

 

t,,, > 24 h

 

16B. (1)

 

 

 

 

 

58°FSVPDPELIKTVTINASSSRSGIDDIN’PTVLLKERSTELSIB native huCOX-Z
SBOFSVPDPELIKTV'I‘INASIRGPLTKLSDNSLVSDERSTELGlB Sins597—612 hucox-2
SBOFSVPDPELIKTVTIAASIRGPLTKLSDNSLVSDERSTELEle SinsS95—612 hucox—z
B. (11)
Sins595-612 huCOX-Z Sins597-612 huCOX—Z
0 2 4 12 24 (h) 0 2 4 12 24 (h)
"'r‘r'xA—x “all-Arr.“

Acfinwflflad Actin “'.w

(M > 24 h t” > 24 h

82

 

16B. (111)

 

 

 

 

 

 

c»
C
.E —-
.5 —-___-—
E __~—— -
2 —_
.5
8
2
(L
32 ‘5
0 . . e .
0 5 10 15 20 25

Time after puromycin (h)

. O native nucox_2

 

‘- — — — i SinsS97-612 “cox-2

'- — — — '- SinsS95—612 huCOX-Z

Figure 16. Deletion and scrambled mutations of the l9-AA cassette disrupt COX-2
degradation. A (I) and B (1), Amino acid sequences of the C-termini of cyclooxygenase
constructs that were stably transfected into HEK293 cells. The underlined sequence is the
consensus N-glycosylation site at the start of the COX-2 l9-aa cassette. The C-terminal
sequences of native huCOX-2, native ovCOX-l, and ins594—612 ovCOX-l are included
for comparison. A (II) and B (11), HEK293 cells stably and inducibly expressing the
constructs shown in (I) were grown to ~80% confluency, subjected to serum-starvation
for 24 h, and then treated with the appropriate inducing agent (10 ug/ml tetracycline or
10 [1M ponasterone A) for 24 h. Puromycin (50 M) was then added to block translation
and COX-2 and actin protein levels were analyzed at different times by Western blotting.
The Western blotting results in A(II) and EU!) are representative of two and three
separate experiments, respectively, from which protein half-lives were determined. B
(III), Densitometry analysis of the degradation of native huCOX-Z, Sins595-6l2 huCOX-
2, and Sins597-612 huCOX-Z. The results are based on at least three different
experiments for each protein. The error bars denote +/- SE of the mean.

83

17A.

 

58017SVPDPELIKTVTIEAjSSRSGmDINPTVLMERSTELMB
5SOFSVPDPELIKTVTIESSRSGLDDINAAVLIKERSTELMB
58°FSVPDPELIKTVTIflhjSSRSGLDDINPTAI-LKERSTEL“8
58°FSVPDPELI KTVTI_NA_SSSRSGLDDINPTVALKERSTEL618
58°1='svapiz:LIKTV'I‘I_1~m_sssitsGm!):tmvrvniutfi:RsTEL618
53°FSVPDPELIKTVT1&8s8118mm)ImuvrvnnnERsTEL618

 

native huCOX—2
. P607A T608A huCOX—2
V609A huCOX—2

L610A huCOX-2

L61 1A huCOX-Z

K612A huCOX-2

 

 

 

B.
P607A T608A huCOX-Z V609A huCOX-2
0 2 4122401) 0 2 4122401)
n “ A fi H .- 2..
Actin rc r. m m are; Aetin v- v n w. ~

tr2"8h

L610A huCOX—Z

L61 1A huCOX-2

 

 

0 2 4 12 24(h)

‘III-If-|‘-"‘

 

Actin I'- F- i- O '-
Actin
tm > 24 b
K612A huCOX-Z
0 2 4 8 24 (h)
Actin Actin

“have m 6’3w «”4. 55¢

3? 1503 é‘ufl it? " .4-

g1>24h

84

'Ili- '

“WAA " ~

0 2 412 24 (h)

iqw ,. .. (“an .9 - ‘ w

finfii‘ifiq

t1,2 > 24 h

native huCOX-Z
0 2 4 3‘24 '(h)

 

 

{‘1'

are: new we. a“. if.»

in ~ 5 h

 

17 C.

 

 

 

 

 

 

   

 

 

 

P607A T608A huCOX—2
100

a: 80 -
.E
.E
2 60 ‘
2
.E
6 .
§ 40
a
.\’

20 ‘

O . . f .
0 5 10 15 20 25
Time (h)
V609A huCOX-2

100
g 80 .
.E
3 so -
2
5
3 4o ‘
2
:1.
,\° 20 +

0 r .
O 5 10 15 20 25
Time (h)

¢—0 native

.______.. mutant

85

17C. (cont’d)

 

 

 

 

 

 

 

 

 

 

L610A huCOX-Z
100 \ ‘ ’
t.
804 \ \
D \ \
.5 ‘ x
-----------
2
5 J
‘3 4o -
E
at
20 ~
0 . . r .
0 5 10 15 20 25
Time (h)
L611A huCOX-Z
100
a 80 "
C
E
(I
E 60 ‘
2
.E
3 40 ~
2
n.
=2 20 j
0 . . . .
0 5 1O 15 20 25
Time (h)

¢——9 native

.. ..... _. mutant

86

K612A huCOX—Z

 

100

  

00
O

\‘I---{ ______ _ —————— i

O)
O

  

4h
0

°/o Protein remaining
N
O

 

 

 

O

10 15 20 25
Time (h)

C
m .

‘—9 native

'- ..... .. mutant

Figure 17. The C-terminal end of l9-aa is essential for COX-2 protein degradation.
A, Alanine scanning mutants of the C-terminal portion of l9-aa that were stably
transfected into HEK293 cells. The underlined sequence is the Asn-594 glycosylation
site. B, HEK293 cells stably and inducibly expressing the constructs shown in (A) were
grown to ~80% confluency, subjected to serum-starvation for 24 h, and then treated with
10 ug/ml tetracycline for 24 h. Thereafter, the time course of degradation of the proteins
was determined as described in the legend to Fig. 16. Protein half-lives of all proteins
except P607A T608A huCOX-Z were determined from at least three separate
experiments. C, Densitometry analysis of the degradation of the mutants shown in A and
B. The results for native huCOX-2 and for all mutants except P607A T608A huCOX-Z
are based on at least three independent experiments. Densitometry analysis of P607A
T608A huCOX-2 is based on a single experiment. The degradation profile of huCOX-Z
has been included for comparison. The error bars denote +/— SE of the mean.

87

Glycosylation of Asn-594 correlates with COX-2 protein degwon. We
previously reported that kifunensine, which inhibits ERAD by preventing N—glycan
processing by ER 0L1,2 mannosidase I, stabilizes huCOX-Z stably expressed in HEK293
cells (257). At the same time kifunesine (KIF) leads to the appearance of a less mobile,
alternatively glycosylated form of the native enzyme in HEK293, NIH/3T3, and RAW
264.7 cells even after protein translation has been inhibited (Figs. 180, b, and c). This
higher glycosylated form is not observed after KIF treatment of HEK293 cells expressing
N594A huCOX-2 (257). Endo H treatment of native huCOX-2 expressing cells collapses
both alternatively glycosylated forms to the 66 kDa fully deglycosylated form of the
enzyme (Fig. 18a). Therefore, we believe that KIF stabilizes the 74 kDa Asn-594
glycosylated form of the native enzyme. The enhanced levels of the 74 kDa form upon
KIF treatment could be used to assess Asn-594 glycosylation of COX-2 in HEK293 cells.
Experiments were performed to determine if the deletion mutants delS97-612 huCOX-Z,
del602-612 huCOX-2 and del607-612 huCOX-Z, and the scrambled mutant Sins597-612
huCOX-2 are glycosylated at Asn-594. The deletion mutant delS97-612 huCOX-2 has the
same electrophoretic mobility as delS95-612 huCOX-Z, which lacks the Asn—594
glycosylation site (Fig. 19a, panel I). KIF treatment of cells expressing delS97-612
huCOX-2 did not affect the mobility of this mutant suggesting that it is not glycosylated
at Asn-594 (Fi g. 19a, panel II). Similarly, del602-612 huCOX-2 and del607-612 huCOX-
2 did not appear to be glycosylated at Asn-594 (Fig. 19a, panel 11). The scrambled mutant
SinsS97-612 huCOX—Z carrying an intact Asn-594 glycosylation site had the same

electrophoretic mobility as SinsS95-612 huCOX—2 with or without KIF treatment

88

 

18A.

 

 

 

74 kDa
72 kDa
66 kDa —*
1. tet-induced
2. tet-induced + 12 h KIF
3. EndoH cleavage of l
4. EndoH cleavage of 2
B.
25 EM KIF
0 2 4 8 24 2 4 8 24 Time with puromycin (h)
HEK293 pan-i.“ put-grit“:
120
.E’
.E
g ._. No inhibitor
:3 L". 25 uM KIF
.5
3 L ...... A 50 “M KIF
O.
32
0

 

 

 

Time with puromycin (h)

89

18C.

25 pM KIF 50 pM KIF
0 2 4 8 2 4 8 4 8 Time with CHX (h)
1::

 

NIH/3T3 II'_‘___‘____,;_;....:....
d------""" Actin
so KIF
o 2 4 2 4 Time with CHXOI)
RAW 264.7 {,figifl-Wtwgz

Figure 18. KIF treatment stabilizes the Asn-594 glycosylated form of COX-2. A,
Native huCOX-2 expression in HEK293 cells was induced with 10 ug/ml tetracycline for
24 h . After protein induction cells were incubated for an additional 12 h in normal
culture medium with or without 25 “M KIF. Thereafter, cell lysates were prepared and
boiled in SDS sample loading buffer and then treated with or without Endo H (0.4
mU/ul) for at least 12 h. Proteins were resolved by SDS-PAGE on a 7% tris-acetate
polyacrylamide gel and subjected to Western blotting. B, Native huCOX-2 expression
was induced as described in the legend to Fig. 16. Puromycin (50 M) was then added to
block translation in the absence or presence of 25 51M KIF. Thereafter, COX-2 protein
levels were analyzed at different times by Western blotting and densitometry. The
densitometry analysis is based on three different experiments that were performed to
examine the effect of 25 “M and 50 M KIF on COX-2 degradation. The error bars
denote +/- SE of the mean. C, Quiescent NIH/3T3 cells were stimulated with 20% FBS
for 4 h to induce COX-2 expression. Cycloheximide (CI-IX; 50 11M final concentration)
was added to the medium and the cells were incubated for the indicated times with or
without additional treatment with 25 11M or 50 11M KIF. RAW264.7 cells were treated
with 200 ng/ml LPS for 12 h to induce COX-2 expression. Thereafter, cells were treated
with 50 ”M CHX for the indicated times in the presence or absence of 50 M KIF. Cell
lysates were prepared and analyzed for COX-2 protein levels. Arrows in B and C show
the 72 and 74 alternatively glycosylated forms of COX-2 detected by immunoblotting.

9O

(Fig. 19b). Just like delS97-612 huCOX-2, treatment with KIF did not result in the
detection of a less mobile glycosylated form of SinsS97—612 huCOX-2 (Fig. 19b).
Collectively, these results indicate that mutations of the 19-aa that confer stability to
COX-2 also prevent glycosylation of Asn-594. Consistent with this observation, the
mutant V609A huCOX-2, which is degraded to a lesser extent compared to native
huCOX-2, is also glycosylated less efficiently at Asn-594 in KIF-treated cells (Fig. 19c).
Therefore, the C—terminal 16 amino acids of 19-aa appear to play an important role in
effecting glycosylation of Asn-594. This could be one mechanism by which the 19-aa
cassette mediates COX-2 protein degradation.

The helical region upstream of Asn-594 negatively regu_lates Nglvcosvlation.
Previously, we showed that inserting the COX-2 19-aa near the C-terminus of ovCOX-l
yielded an unstable mutant ins594-612 ovCOX—l that was degraded with a tm of ~3 h
(257). A comparison of the molecular masses of ins594-612 ovCOX—l with that of its
stable, Asn-594-mutated counterpart ins594-612 (N594A) ovCOX-l suggested that the
former was fully glycosylated at Asn-594 (257). Unlike native huCOX-Z, KIF treatment
of HEK293 cells expressing ins594-612 ovCOX-l failed to elicit the appearance of a
higher glycosylated form of this ovCOX-l mutant; instead, KIF significantly stabilized
insS94-612 ovCOX-l (257). The apparent disparity between native huCOX-2 and
insS94-612 ovCOX-l with respect to Asn-594 glycosylation suggested that the ease with
which the two proteins became glycosylated at this site was different. The fact that the “/2
of ins594-612 ovCOX-l is ~2 h shorter than that of native huCOX-2, also suggested that

the efficiency of Asn-594 glycosylation appears to positively correlate with the rate of

91

19A. (1) 1 2

5 ~ . 1. delS95-612 hucox-z
“ fl 2. delS97-612 huCOX-Z

A. (II)
(191597-512 huCOX—Z del602-612 huCOX-Z del607-612 huCOX-Z
1 2 3 4 1 2 3 4 l 2 3 4
fl ‘9‘} “It U W I
-~,’ wtfi w~ ‘V
l. tet-induced
2. tet-induced + 12 h KIF
3. EndoH cleavage of 1
4. EndoH cleavage of 2
B.

l 2 3 4

. SinsS95-612 huCOX-Z
Sins597-612 huCOX-Z
Sins595-612 huCOX-Z + 12 h KIF
SinsS97-612 huCOX-2 + 12 h KIF

AWN—-

92

19C' native V609A hucox-z

l 2 l 2 3 4

 

 

' “'1':-

1"?“

1. tet-induced

2. tet-induced + 12 h KIF
3. EndoH cleavage of l

4. EndoH cleavage of 2

Figure 19. COX-2 stabilizing mutations of the 19-aa prevent glycosylation of Asn-
594. A (I), The expression of delS95-6l2 huCOX-2 or delS97-612 huCOX-2 was induced
with 10 jig/ml tetracycline for 24 h. (11), After protein induction as in (1) cells were
incubated for an additional 12 h in normal culture medium with or without 25 M KIF
(delS97-612 huCOX—2) or 50 M KIF (del602-6l2 huCOX-2 and del607-612 huCOX-2).
Thereafter, cell lysates were prepared and boiled in SDS sample loading buffer for 10
min and then treated with or without Endo H (0.4 mU/iil) for at least 12 h. Proteins were
resolved by SDS-PAGE and subjected to Western blotting as in 18A. B, Expression of
SinsS95-612 huCOX-2 or Sins597-612 huCOX-2 was induced as in A(I). After protein
induction cells were incubated for an additional 12 h with or without 25 11M KIF.
Proteins were resolved by SDS-PAGE and subjected to Western blotting as in 18.4. C,
Expression of native huCOX-2 or V609A huCOX-2 was induced as in A( I). After protein
induction cells were incubated for an additional 12 h with or without 25 uM KIF. Cell
lysates were then prepared and boiled in SDS sample loading buffer for 10 min and then
treated with or without Endo H (0.4 mU/ttl) for at least 12 h. Proteins were resolved by
SDS-PAGE and subjected to Western blotting as in 18A. Arrows indicate differences in
molecular mass due to N—glycosylation.

93

COX degradation. Native huCOX-2 and ins594-612 ovCOX-l have very similar
sequences downstream of Asn-594; therefore, we decided to probe the region upstream of
Asn-594 in huCOX-2 for its ability to negatively regulate N-glycosylation. In the
muCOX-2 X-ray crystal structure, the Asn-594 glycosylation site is at the C-terminal end
of a 9-residue (it-helical region (Figs. 20a and b). This helix (helix A) is connected to a
nearby helix (helix B) by a long 15-residue loop; four of these residues are presumed to
be part of the loop because they cannot be resolved in the crystal structure (Figs. 20a and
b). The amide side chain of Asn-594 appears to be pointed upwards in the direction of
helix B which is at a distance of ~4.5 A (Fig. 20c). The length of two GlcNAc residues of
an N-glycan group is ~8 A. Therefore, helix B is situated such that it may impede the
transfer of an N-glycan group to the amine group of the amide chain of Asn-594.
Furthermore, a nearby disulfide bridge formed between Cys-569 of helix B and Cys-575
of the loop may also impede Asn-594 glycosylation because it is situated ~4.7 A away
(Fig. 20c). Moreover, the disulfide bridge may limit the flexibility of the loop, thereby,
constraining helix A into a conformation in which Asn-594 cannot become glycosylated.
The loop and the disulfide bridge can both be seen in the X-ray crystal structure of
ovCOX-l, indicating that they are conserved in both isoforms (Fig. 20d). The last
resolved residue in the ovCOX-l structure is Pro-583, which is part of the loop; therefore,
the region that would be immediately upstream of Asn-594 in ins594-612 ovCOX-l (ie.
Asp-584 to Val-593) cannot be resolved, suggesting a lack of secondary structure.
Furthermore the presence of proline and glycine at two consecutive positions, 591 and
592, respectively, makes it highly improbable that this region would possess any

secondary structure. It is therefore likely that ins594—612 ovCOX-l is fully glycosylated

94

at Asn-594 because this glycosylation site is situated in a loop that is long enough to be
very flexible even in the presence of the disulfide linkage (Fig. 206).

To test this hypothesis, we prepared and stably transfected huCOX-2 mutants
bearing various modifications of an 8-residue sequence (UPSTRM8) immediately
upstream of Asn-594. The mutation V591P T592G huCOX—2 introduces a proline and a
glycine into UPSTRM8 to disrupt the helix that is presumably formed by this region of
huCOX-Z. This mutation was also designed to mimic the corresponding region of
ovCOX-l (Fig. 21a). A second mutant was made in which the whole UPSTRM8 region
of huCOX—2 was replaced with the corresponding region of ovCOX-l (Fig. 21a). This
mutant is designated ovCOX-l UPSTRM8 huCOX-2. The initial rates of degradation of
V591P T592G huCOX-2 (ti/2 ~4-5 h) and ovCOX-l UPSTRM8 huCOX-2 (ti/2 ~3 h)
were similar to that of the native enzyme (Figs. 21b and c). However, overall, both
mutants appeared to be degraded to a greater extent than native huCOX-2 (Fig. 21c). For
instance, essentially all of V591P T592G huCOX-2 or ovCOX-l UPSTRM8 huCOX-2
was degraded by the 24 h time point, while ~15% of native huCOX-2 remained (Fig.
21c). Also, these mutations resulted in the detection of the 74 kDa glycosylated form
even without KIF treatment (Figs. 21b, 22a, and 22b). This interesting observation led us
to hypothesize that disrupting helix A facilitated Asn-594 glycosylation, and
consequently, enhanced the overall degradation of the protein. If this is true, then
replacing the UPSTRM8 region of huCOX-2 with that of muCOX-2 should not change
the glycosylation pattern at Asn-594. As shown in Fig. 21a substantial differences exist
between the UPSTRM8 regions of muCOX-2 and huCOX—2. Nonetheless, we expected

that helix A of muCOX-2 would be conserved in huCOX-2. The initial degradation rate

95

of the mutant muCOX-2 UPSTRM8 huCOX-2 (ti/2 ~4-5 h) was similar to that of native
huCOX-2 and the ovCOX—l UPSTRM8 mutants (Figs 21b and c). Moreover, the extent
of the overall degradation of this mutant appeared to be intermediate between that of the
native huCOX—2 and the ovCOX-l UPSTRM8 mutants (Fig. 21c). The Asn-594
glycosylation pattern of muCOX-2 UPSTRM8 huCOX-2 was similar to that of the native
huCOX-2 in that unlike the ovCOX-l UPSTRM8 mutants, very little of the Asn-594
glycosylated form of this mutant could be detected in the absence of KIF or M6132
treatment (Figs. 22a and b). The modest detection of Asn-594-glycosylated muCOX-2
UPSTRM8 huCOX-2 is consistent with the observation that the extent of its overall
degradation is slightly greater than that of native huCOX-2. Nonetheless, these results are
consistent with our expectation that the UPSTRM8 regions of muCOX-Z and huCOX—2
have conserved helical secondary structure which appears to negatively regulate Asn—594
glycosylation. It is important to note that the putative helix A formed by the UPSTRM8
of huCOX-2 may actually be 2 residues longer since, unlike muCOX-2 UPSTRM8, this
region lacks a proline residue at its N-terminus (Fig. 21a).

Just like native huCOX-2, the degradation of all three UPSTRM8 huCOX-2
mutants involved entry into the ERAD system because the 26S proteasome inhibitor
MG132 and the ER al,2-mannosidase I inhibitor KIF stabilized these mutants (Figs. 22a,
b, and c). These findings lead us to propose that the UPSTRM8 of huCOX-2 acts in
concert with the downstream l9-aa to control the timing of the entry of the protein into
the ERAD system and the extent of protein degradation by regulating glycosylation of

Asn-594.

96

20A.

 

 

helix loop * helix
5638‘ IQSLICMGCPFTSFNVBDPQPEKTATINABS96 mouse
SIQSLICNNVKGCPFTSFSVPDPELIKTVTINAS human

 

 

Helix B
n-570

-569

 

Ser 6
Helix A

 

97

 

 

Pro-583
helix
+ -------- >
56‘ LKKLVCLRTKTCPYVSFHVPDPRQEDRPGV ——————————————————— ERPPTELGUO ovcox-l
55‘ IQSLICNNVKGCPFTSFSVPDPELIKTVTINASSSRSGLDDINPTVLLKERSTEL“8 hucox—z

 

 

 

E.

 

helix loop -k helix
5“ IQSLICNNVKGCPFTSl-“SVPDPELIKTVTImisssrtsammIiu=t'1'v1.i:.RERS'I‘ELM native hucox—z
9“ LKKLVCLNTKTCPWSFHVPDPRQEDRPGVNAa "M... . . rent” insS94—612 ovCOX—l

V helix loop

 

 

 

 

Figure 20. Structure of the C-terminal region of COX-1 and COX-2. A, Amino acid
sequences of the C-terminal regions of muCOX-2 and huCOX-Z that are located
upstream of the Asn-594 glycosylation site (underlined). Residues that are colored blue
are conserved between the two species. Segments that form helices in the muCOX-2 X-
ray structure are highlighted. The region marked with an asterisk is not resolved in the
muCOX-2 X-ray structure and is very likely to be a continuation of the resolved loop.
The disulfide linkage between Cys-569 and Cys-575 is also shown. B, Ribbon diagrams
depicting the structure of the C-terminal region of muCOX-2 whose sequence is shown in
A. The last resolved residue in the muCOX-2 X-ray structure is Ser-596, the final amino
acid of the Asn-594 glycosylation site. Asn-594 is situated in a helix (Helix A) that is
positioned right under an adjacent helix (Helix B). The amide chain of Asn-594 appears
to be pointed upward in the direction of helix B. The position of the disulfide linkage
relative to Asn-594 is shown. C, A ribbon diagram showing the distances between Asn-
594 and Asn-570 of helix B (4.5 A), and Asn-594 and the disulfide linkage between the
loop and helix B (4.7 A). D, Top, Structure of the C-terminal region of ovCOX-l. The
last resolved residue in the X-ray structure of ovCOX-l is Pro-583. The conserved

98

disulfide linkage between Cys-569 and Cys-575 is shown. Bottom, Amino acid
sequences of the C-termini of ovCOX-l and huCOX-2. The sequence in ovCOX-l that
forms a helix is shown. The sequence highlighted with a dashed double arrow is the
region of ovCOX-l that is not resolved in the X-ray crystal structure whose
corresponding region is resolved in the X-ray structure of muCOX-2. Conserved residues
between ovCOX-l and huCOX-2 are highlighted in red. The conserved cysteines that
form the disulfide bridge in either isoform are colored blue. The last resolved residue,
Pro-583, is also colored blue. E, Comparison of the C-termini of native huCOX-2 and
insS94-612 ovCOX-l. Segments expected to form a helix are highlighted with black
double arrows. Segments expected to form a loop are highlighted with blue double
arrows. The area of native huCOX-2 marked with an asterisk is likely to be a
continuation of the helix. The conserved Cys-569 and Cys-5 75 are colored red.

21A.

 

53°FSVPDPELIKTVTIESSRSGLDDINPTVLLKERSTELe18 native hucox-2
SBOFSVPDPELIKTPGIESSRSGLDDIN‘PTVLLKERSTEL‘S18 V591P T592G hucox-z
58°FSVPDPRQEDRPGVEA—SSSRSGLDDINPTVLLKERSTEL518 ovCOX-l UPSTRM8 huCOX—2
530FSVPDPQPTKTATIESSRSGLDDINPTVLLKERSTEL518 mucox—z UPSTRM8 huCOX-Z

 

 

 

 

 

B.
V591? T592G huCOX—Z OVCOX-l UPSTRMS huCOX—Z
0 2 4 12 24 (ii) “-9. ”2., ....‘3_ -12. 24 0')
Actin can... Actin “C-..
tug” 3 h
tm ~ 4-5 h

muCOX-Z UPSTRM8 huCOX—Z
0 2 4 12 24 (h)

 

Actin ..“~

t1]; ~ 4-5 h

99

21C.

°/o Protein remaining

O————O

A- ----- A

100

“I. Protein remaining

80-

60‘

40-

20-

ovCOX-l UPSTRM8 huCOX-2

 

100

on
O

O)
O

    

 

 

 

40
20 \
i ' — ~ _ *
0 I I I I—¢
0 5 10 15 20 25 30
Time (h)
native
mutant

V591P T592G huCOX—Z

 

 

 

 

 

\
0 5 10 15 20 25
Time (h)

.—9 native

l- ----- -I mutant

lOO

21C. (cont’d)

muCOX-2 UPSTRM8 huCOX-2

100

 

on
O

60‘

40-

% Protein remaining

20-

 

 

 

 

Figure 21. Degradation of the UPSTRM8 huCOX-2 mutants. A, Amino acid
sequences of the C-termini of UPSTRM8 huCOX-2 mutants that were stably transfected
into HEK293 cells. The underlined sequence is the Asn-594 glycosylation site. B, After
stably transfecting the UPSTRM8 huCOX-2 constructs shown in A, protein expression
was induced as described in the figure legend to Fig. 16. Thereafter, cells were treated
with 50 M puromycin for the indicated times and remaining levels of COX-2 and actin
were analyzed by Western blotting. The Western blotting results are based representative
of three separate experiments from which protein half-life measurements were made.
Arrows indicate differences in molecular mass due to N—glycosylation. C, Densitometry
analysis of the degradation of the UPSTRM8 huCOX-2 mutants relative to native
huCOX-2. Densitometry is based on three separate experiments. Error bars denote +/— SE
of the mean. Red asterisk indicates that the difference in % protein levels between the
mutant and native protein at a given time point is statistically significant based on a
Student’s paired t-Test (p value < 0.01). Black asterisks indicate that the overall
degradation of the mutant relative to the native protein is statistically significant based on
a Student’s paired t-Test (p value < 0.05).

101

22A. (1)

Endo H cleavage

 

123456123456

“kDa—v

74 kDa :
72 kDa ”W . . ,
MV

1. native huCOX-Z

2. native huCOX-2 + KIF

3. V591P T5926 huCOX—Z

4. V 591P T592G huCOX—Z + KIF

5. muCOX-2 UPSTRM8 huCOX-Z

6. muCOX-Z UPSTRM8 huCOX-Z +KIF

Endo H cleavage

 

l. native huCOX-Z

2. native huCOX-Z + KIF

3. ovCOX-l UPSTRM8 huCOX-Z

4. ovCOX—l UPSTRM8 huCOX-2 + KIF

102

22B. (1) V591P r5920 huCOX-Z

 

 

20 “M MG132
0 2 4 12 24 2 4 12 24 (h)

Actin "wanton nun-d

 

 

B. (11) ovCOX-l UPSTRMS huCOX-Z
20 1.1M MGI32
o 4 12 24 4 12 24 (h)
:z I! n in «a.

-., .—

 

 

 

B. (III)
muCOX—Z UPSTRMs huCOX-Z
20 MI MG132
0241224 24122401)
:2: UI-iudir- i"§i‘g‘“‘
Actin . “mi! Wilt. W “8'

103

22B. (IV)

native huCOX-2

 

20 pM MGl32
0 2 4 8 24 2 4 12 24 (h)

-..__-.____ _. --.—a—.-— _.-__.._..-....,.._.._. ..._..._- ._:_.....
(. .. _ . “J. _ . _, " ~.. , . ‘3‘}?

 

 

,i,. r

“155': .'-1-_. :53 .‘fl: 3'; r151; 11' 'i- L,- 1l- - .- ‘ -.‘ 1' ,. --'-'-'a¥!-'. ’.- ' M9 “:I-JKC. - ‘af'? .. hi”- ' ' ‘
W" {‘2‘}; 3m 5 ‘t 3.} r ' J 'a‘ . ‘. 1’ 1‘1. ."L'. ‘_ ""rs'c ' _ “If ‘ is; "w ‘\.‘ T 7' 21".. 1
it .
. f
H i
. ‘r. u“ -
Actin ~ II ' ~ ~ |~

avw -u- v-r-I— In.“ ‘ 1:- a.-. rv v-uu --——' .. v-awr—n- - u-nu——-— r——-—-—fi - ”—— _————-. ., 9

C.
V591 P T5926 huCOX-2

 

% Protein remaining

 

Time (h)

Q. _____ .9 No protease inhibitor

I 1 20 W M6132

 

104

22C. (cont’d)

ovCOX1 UPSTRM8 huCOX2

% Protein remaining

 

Time (h)

muCOX—2 UPSTRM8 huCOX—2

1 20
1 00
80
60

4O

% Protein remaining

20

 

O 5 1 0 15 20 25
Time (h)
0. _____ 4 No protease inhibitor

 

I— 1 20 11M MG132

105

22C. (cont’d)

 

 

 

 

 

Native huCOX—2
E
E
3
5:
Time (h)
.1. ————— 4 No protease inhibitor

I—-—I 20 11M MG132

Figure 22. Asn-594 glycosylation of the UPSTRM8 huCOX-2 mutants and their
degradation by the ERAD pathway. A (I) and (II), The expression of native huCOX-2
or the UPSTRM8 huCOX-2 mutants was induced with 10 ug/ml tetracycline for 24 h.
After protein induction cells were incubated for an additional 12 h in normal culture
medium with or without 25 M KIF. Thereafter, cell lysates were prepared and boiled in
SDS sample loading buffer for 10 min and then treated with or without Endo H (0.4
mU/ul) for at least 12 h. Proteins were resolved by SDS—PAGE and subjected to Western
blotting as in 18A. Arrows indicate differences in molecular mass due to N—glycosylation.
B (I), (II), (III), and (IV), HEK293 cells stably expressing the UPSTRM8 huCOX-2
mutants were grown to ~80% confluency, subjected to serum-starvation for 24 h, and
then treated with 10 ug/ml tetracycline for 24 h. Puromycin (50 M) was then added to
block translation with or without additional treatment with 20 “M MG132. Thereafter,
COX-2 and actin protein levels were analyzed at different times by Western blotting.
Arrows indicate differences in molecular mass due to N—glycosylation. C, Quantitative
analysis of the effect of 20 M MG132 on the protein stability of native huCOX-2 and
the UPSTRM8 huCOX-2 mutants.

106

 

   

 

Protein Extent of Asn-594 Protein half-llfe Extent of protein
glycosylation clearance
Full glycosylation Full clearance
ins594-612 ovCOX-l ~ 3 h
ovCOX-l UPSTRM8 huCOX-Z ~ 3 h
V591? T5926 huCOX-z ~ 4—5 h
muCOX-z UPSTRM8 huCOX-Z ~ 4-5 h
Native huCOX-Z -— 5 h
V609A huCOX-Z ~ 8 h
31115597612 huCOX-2 No glycosylation >24 h
(MSW-612 huCOX-z No clearance
del602-612 huCOX-Z > 24 h
412507-612 huCOX-Z

 

 

Table 2. Relationship between the extent of Asn-594 glycosylation and overall
protein degradation. The extents of Asn-594 glycosylation and protein clearance are
arbitrary and are based on Western blotting and densitometry data. Protein clearance is a
relative measure of how much protein is remaining at the end of the 24 h experiment.

107

Discussion

Inhibiting the N—glycosylation of COX-2 during its synthesis and maturation in
the ER severely compromises its catalytic activity (151). Therefore, co-translational N-
glycosylation of COX-2 appears to be required for the proper folding of the enzyme. It is
well known that there are ER-resident chaperones, namely calnexin and calreticulin, that
associate with folding or misfolded glycoprotein substrates mainly by binding to their
exposed N—linked oligosaccharide groups ( 166). Asn-594, the last N-glycosylation site of
COX-2, is usually variably glycosylated (148,151,258), and is not required for the
enzyme to achieve a catalytically competent conformation (151). Previously, we have
shown that the Asn-594 glycosylation site is essential for the ER-associated degradation
(ERAD) of COX-2 (257). In the present study, we have addressed the question of
whether glycosylation of Asn-594 is essential for enabling COX-2 ERAD. We provide
evidence that Asn-594 glycosylation is regulated by interplay of an upstream segment,
approximately 8 residues long and a downstream l6-residue sequence that forms part of
the 19-aa cassette. Furthermore, we observed a positive correlation between the extent of
Asn-594 glycosylation and the overall degradation of the protein (Table 2). This is based
on the following findings: a) l9-aa deletion mutants of huCOX-2 that have an intact Asn-
594 glycosylation site but lack various portions of the l9-aa appear not to be glycosylated
at Asn-594 and are completely stable; b) scrambling the l6-amino acid sequence
downstream of the Asn-594 glycosylation site prevents glycosylation of Asn-594 and
yields a mutant that is degraded much slower than native COX-2; c) inserting the
huCOX-2 19-aa near the C-terminus of ovCOX-l results in the complete glycosylation of

Asn-594, and the COX-1 mutant has a shorter tl/z (~3 h) compared to native huCOX-2

108

(~5 h)(257); and d) replacing the 8-residue sequence immediately upstream of Asn-594 in
huCOX-2 with the corresponding region of ovCOX-l facilitates Asn-594 glycosylation
concomitant with a shorter 11/2 (~ 3 h) and more effective clearance of the protein.

Both the 72 and 74 kDa alternatively glycosylated forms of COX-2 can be readily
detected by immunoblotting in various cell types, including NIH/3T3 fibroblasts and
RAW264.7 macrophages. However, heterologous expression of recombinant huCOX-2
in HEK293 cells yields predominantly the 72 kDa form. The 74 kDa Asn-594-
glycosylated huCOX-2 form can be detected in these cells after treatment with
kifunensine (KIF), an inhibitor of ER a1,2 mannosidase I, which we have shown to
stabilize COX-2 (257). ER a1,2 mannosidase I catalyzes the committed step in the
ERAD pathway (259,260). Its activity releases irreversibly misfolded glycoproteins from
the calnexin/calreticulin folding cycle by cleaving a single mannose residue from an N-
linked MangGlcNAcz (198-200). A putative Mans-binding lectin Yos9p or the ER-
degradation mannosidase I-like protein (EDEM) will bind MangGlcNAcz and facilitate
the delivery of protein having a processed N—glycan to a membrane retrotranslocon for
export out of the ER and into the cytosol for proteasomal degradation (201-
204,206,207,248). The identity of this membrane retrotranslocon has not been
determined. KIF inhibition of ER a1,2 mannosidase I prevents the retrotranslocation and
proteasomal degradation of aberrant glycoproteins by causing them to be retained in the
ER (199,200,248-251). Since KIF inhibits COX-2 degradation and leads to accumulation
of the Asn-594 glycosylated form of the enzyme in HEK293, NIH/3T3, and RAW264.7
cells, we reason that COX-2 ERAD requires Asn-594 glycosylation and subsequent

processing of the attached N-glycan group by ER a1,2 mannosidase I.

109

Even though ERAD is considered to be an ER quality control mechanism for the
elimination of aberrantly folded glycoproteins, COX-2 appears to be degraded from a
properly folded conformation. Sifers and co-workers have proposed that glycoprotein
ERAD (GERAD) requires at least two signal determinants: a) an N—glycan component,
and b) non-native protein structure (259). Their proposal is based on the observation that
inhibiting the glycosylation of misfolded glycoproteins prevents their degradation (261).
However, it is still unclear how GERAD quality control is able to selectively eliminate
terminally defective glycoproteins while sparing folding glycoproteins that have yet to
achieve their native protein structure. It is possible that the intimate physical interaction
of folding glycoproteins with ER resident molecular chaperones protects them from being
degraded. Taking into consideration both the criteria for GERAD and our overall
findings, we provide a model for the initiation of COX-2 degradation by Asn-594
glycosylation (Fi gs 23, 24, and 25). In this model, COX-2 synthesis and maturation in the
ER involves co-translational glycosylation of its first three N-glycosylation sites, but not
Asn-594. Glycosylation of these sites as the COX-2 nascent protein grows from the
Sec61 translocon would serve the purpose of recruiting the chaperones calnexin and/or
calreticulin to assist in protein folding. COX-2 maturation also entails formation of
disulfide bonds, membrane insertion, heme incorporation, and assembly into a
homodimer. This yields a catalytically competent COX-2 having an Asn-594 site that
may not be easily accessible for glycosylation because it is situated in an ordered helix
(helix A) such that its amide group is protected by an adjacent helix (helix B) ~4.5 A
away. A disulfide bridge between Cys-569 of helix B and Cys-575 of a 15-residue loop

connects helices A and B. This covalent modification may further contribute to

110

constraining helix A into a conformation which prevents glycosylation of Asn-594.
However, this disulfide bridge is also present in COX-1 and we have found that the
COX-1 insertion mutant ins594-612 ovCOX-l appears to be fully glycosylated at Asn-
594. Therefore, we propose that the degradation of COX-2 is initiated by a local
conformational change in the helix A-loop-helix B region that enables the post-
translational glycosylation of the protein at Asn-594. This conformational change
probably involves the last 16 amino acids of the l9-aa cassette since this region appears
to be required for Asn-594 to become glycosylated. The mechanism by which the C-
terminal 16 amino acid segment of l9—aa promotes glycosylation of Asn-594 is yet to be
determined. It is possible that after this putative loop region may form hydrophobic and
electrostatic interactions with helix A, thereby, causing helix A to change conformation
so that it is no longer stacked against helix B. Alternatively, the C-terminal 16 amino
acids of 19-aa may serve as a docking site for a protein complex that will remodel the
helix A-loop-helix B region so that the side chain of Asn-594 is exposed for
glycosylation. Based on our experimental results, it is unlikely that cleavage of the
disulfide bridge between the intervening loop and helix B is required for Asn-594 to
become glycosylated. Otherwise, one of the proteins in this complex may be a protein
disulfide isomerase that would catalyze the breakage of the disulfide bond between the
loop and helix B. Oligosaccharyltransferase (0ST) could also be part of this complex.

Since OST is a membrane-bound complex that is associated with the protein translocon

111

23A.

Helix B

native huCOX-2

 

Helix A l
l
\v - ERSTELCOOH
B. i
:3” 3?:
‘ in3594-612 ovCOX-1
\ N5“
\ <
\ \
X ‘ _ _ ‘ , .ERPPTEL°°°”

Figure 23. Schemes illustrating how the C-terminus of COX-2 may regulate Asn-
594 glycosylation. A, In native huCOX-2, Asn-594 glycosylation may be hindered by the
ordered helical conformation of the region immediately upstream of Asn-594. The C-
terrninal 16 amino acid segment of l9-aa appears to promote glycosylation of Asn-594.
B, Since insS94-612 ovCOX-l lacks secondary structure in the vicinity of Asn-594, this
residue is easily accessible to glycosylation. Asn-594 glycosylation of ins594-612
ovCOX-l may not require reduction of the disulfide bond. The green continuous line
represents the loop that is resolved in the X-ray structures of muCOX—Z and ovCOX-l.
The green dashed line is the putative loop region upstream of Asn—594 that is not
resolved in either isoform. The black dashed line represents the C-terminal l6-amino acid
segment of 19-aa.

112

24.
Possibility A

Helix B 5.5
. 321
<1.
<1
.e

.‘e
l i.
Helix A “ Ma.
\\ z - ERSTELCOOH _
\

\ I ‘ ERSTELC°°H

 

 

Possibility B
Helix B
-'
S—S ‘
N’“ (:5-
- — ' l
Helix A 1 3’0...
| | .‘
\w - ERSTEL°°°“ ' :
M».

‘w - ERSTELCOOH
Figure 24. Two possibilities of how Asn-594 of native huCOX-Z may become
accessible to glycosylation. The loop region from Cys-575 to helix A could be long and
flexible enough to enable movement of helix A to permit Asn-594 glycosylation
(Possibility A). Alternatively, reduction of the disulfide bond may also be needed to
sufficiently expose Asn—594 for glycosylation (Possibility B). In either case, the C-

terminal 16 amino acid segment of 19-aa would be required to mediate Asn-594
glycosylation.

113

25.

HelixB
y GI Gil S
"i. ' m
'--' ERp57
.fe , 3:. Disassembly, unfolding
‘ 5“. '. e-‘ and refolding
PV (GlcaMangGlcNAC21‘5 (Glc1MangGlcNAc2)
Helix ‘i, ".l 611
I |
|‘.. - ERSTEL°°°" \.- ERSTEL°°°"

   

UGGT (ManeGlcNAcz)
f l '
( o ding sensor) I ‘0'."
I '..
' 0
w
'\
I
‘.
ER Man I ‘m ERSTEL°°°N

 

O
{3" (MansGlcNAcz)
/ %: KIF
We.

_—’
,—

w - ERSTEL°°°"

Figure 25. A model of how Asn-594 glycosylation mediates the entry of COX-2 into
the ERAD pathway. An N—glycan group consisting of GIC3Man9GlcNAcz is transferred
en bloc from ER membrane-bound dolichol pyrophosphate to Asn-594. This occurs after
the helix A-loop-helix B region has been remodeled to permit glycosylation. The
remodeling process could be mediated by a complex of proteins that dock onto the 16-
amino acid segment of 19-AA. Glucosidases I and 11 (G1 and GII) will cleave off the two
terminal glucose residues leaving a monoglucosylated N-glycan that is specifically
recognized by calnexin, a membrane-bound lectin binding chaperone. The heterodimer
complex of calnexin and ERp57 will participate in the disassembly, unfolding, and
refolding of the protein in an attempt to correct the structural modification to the native
enzyme. The protein will be maintained in the calnexin cycle by the actions of G11 and
the folding sensor and glucosyltransferase UGGT. Since COX-2 is irreversibly
glycosylated at Asn-594, the protein will eventually be released from the calnexin cycle
by ER (21,2 mannosidase 1 (ER Man I) which cleaves off the a1,2-1inked mannose to
form MangGlcNAcz. This will initiate ERAD of Asn-594 glycosylated COX-2. Processing of
the MangGlcNAcz oligosaccharide group by ER Man I is the committed step of ERAD and is
inhibited by KIF.

114

(262,263), the C terminal region of mature COX-2 has to be in the vicinity of the
membrane for Asn-594 to become post-translationally glycosylated. Introduction of the
local conformational change in the helix A-loop-helix B region and the subsequent
glycosylation of Asn-594 may constitute the signals needed to initiate COX-2 GERAD.
The above model for COX-2 degradation could be tested in several ways. We
have preliminary evidence that COX-2 undergoes post-translational glycosylation at Asn-
594. This is based on the observation that in addition to preventing COX-2 ERAD in
HEK293 cells, KIF treatment causes the accumulation of what appears to be the Asn-594
glycosylated form, even in the absence of protein synthesis; this COX-2 glycosylated
form is barely detectable in untreated HEK293 cells. Accumulation of the 74 kDa Asn-
594 glycosylated form of COX-2 in KIF-treated cells may inhibit further glycosylation of
Asn-594. This would explain why KIF also stabilizes the 72 kDa form of the native
enzyme. To the best of our knowledge, the only published report of post-translational N-
glycosylation is that of the secretory glycoprotein human coagulation factor VII (FVII).
FVH has two consensus glycosylation sites, and, like COX-2, exists as two alternatively
glycosylated forms of 54 and 56 kDa (264). Using 35S pulse chase analysis, Steenstrup
and co—workers elegantly demonstrated that the 54 kDa form is a precursor for the 56
kDa form, and by Endo H treatment were able to show that the difference between the
two forms was due to N-glycosylation. Similarly, it will be important to perform 358
pulse chase analysis to confirm that the 72 kDa form of COX-2 is indeed the precursor
for the 74 kDa form in KIF-treated cells. It would also be interesting to examine the
effect of mutagenesis of Cys-575 on the extent of Asn-594 glycosylation and the overall

degradation of COX-2.

115

In conclusion, we have identified a potential mechanism for the initiation of
COX-2 ERAD that involves Asn-594 glycosylation, its negative regulation by an 8-
residue segment that is immediately upstream, and its positive regulation by the last 16
amino acids of the l9-aa cassette. The upstream 8-residue sequence may also play a role
in the processing of the N—glycan group at Asn-594 that preceeds translocation of COX-2
to the cytoplasm for proteasomal degradation. We reason that both the 19-aa cassette and
the upstream 8-residue sequence form a 27 amino acid sequence that constitutes the

instability element responsible for controlling the intracellular stability of COX-2.

Acknowledgments
We thank Dr. Randal Kaufman and Dr. Billy Tsai for helpful suggestions relevant

to this study.

116

CHAPTER IV

SUBSTRATE DEPENDENT COX-2 PROTEIN DEGRADATION

Summary
Cyclooxygenases (COX-1 and COX-2) catalyze the committed step in prostanoid
synthesis. Previously, we showed that the unique C-terminal 19 amino acid cassette of
COX-2 enables degradation of this COX isoform in a proteasome-dependent manner.
Here, we demonstrate that COX-2 degradation in NIH/3T3 and HEK293 cells is
enhanced by treatment with exogenous arachidonic acid (AA) at concentrations as low as
5-20 11M. At similar concentrations AA did not cause significant COX-l degradation in
either cell type. Endogenous COX-2 degradation in serum-treated NIH/3T3 cells was
enhanced by treatment with 10 11M calcium ionophore (A23187) and was almost
completely blocked by cyclooxygenase (COX) inhibitors. AA—induced COX-2
degradation was retarded by COX inhibitors. The 26S proteasome inhibitor M0132,
which blocks basal COX-2 protein turnover in HEK293 cells, did not affect the AA-
dependent degradation of the protein. Moreover, AA-dependent degradation of COX-2
occurred independently of its C-terminal 19 amino acid cassette. Only fatty acid
substrates of COX-2 were able to enhance degradation. Also, 3 6533A COX-inactive
point mutant of COX-2 was resistant to AA-induced degradation even though it
underwent substrate—independent degradation at the same rate as native COX-2. Loss of
COX specific activity and prostaglandin product formation was also observed in COX-2-

expressing cells treated with AA. We propose that substrate-dependent inactivation of

117

COX-2 causes structural damage to the enzyme that subsequently leads to its

degradation.

Introduction

Cyclooxygenases-1 and -2 (COX—1 and COX-2) are membrane-bound, ER-
resident hemoproteins that catalyze the committed step in prostanoid synthesis
(l,3,7,17,150). Although encoded by different genes, COX-1 and COX-2 are structural
isoforms that share ~60% identity in primary structure. COX-1 is a stable protein that is
constitutively expressed in resting cells of many tissues, most notably in platelets and
vesicular gland (1,257,265,266). In contrast, COX-2 is a stimulus-inducible protein that is
quite unstable and whose expression is observed to be short-lived in epithelial,
endothelial, smooth muscle, and fibroblast cells (20,21,23,24,257). The short half-life of
COX-2 protein mimics that of the COX-2 mRNA. Both the COX-2 mRNA and protein
have been found to possess instability elements that target them for rapid degradation. At
the 3’ untranslated region (UTR) of COX-2 mRNA are multiple AUUUA elements that
are known to target the message for rapid exonuclease cleavage (20,129). COX-1 mRNA
lacks these 3’ UTR AU-rich elements which would explain the prolonged expression of
COX-1 mRNA in megakaryocytes and vascular endothelial cells (117,118,267). We have
recently reported that a unique C-terminal l9-amino acid insertion near the C-terminus of
COX-2 confers the enzyme with a short half-life by rapidly and selectively initiating its
proteasomal degradation (257). This cassette is part of a larger 27 amino acid sequence
that appears to be essential for directing ER-associated degradation of the enzyme. It

could therefore be postulated that the evolution of COX-2 must have entailed the

118

development of post-transcriptional and post-translational mechanisms that contribute to
the sophisticated tight regulation of the synthesis of COX-2-derived prostanoids.

The principal endogenous fatty acid substrate of COX-1 and —2 is arachidonic acid
(20:4A5'8'11’”) which is mobilized from the sn-2 position of membrane phospholipids upon
the activation of cytosolic phospholiase A201 (cPLAzoc) with bradykinin, thrombin,
growth factors, calcium ionophore, or cytokines (l,3,8,36-41,43). The cyclooxygenase
isoforms have two catalytic activities that occur at two spatially distinct active sites. At
the cyclooxygenase (COX) active site, arachidonic acid (AA) is oxygenated to form the
hydroperoxide and endoperoxide prostaglandin G; (PGGZ) (1,3,7). This prostaglandin
intermediate is then moved to the peroxidase (POX) site where its hydroperoxy group
undergoes a two—electron reduction to form the alcohol prostaglandin H2 (PGH2)
(l,3,7,156). PGH2 is a common substrate for downstream terminal prostanoid synthases
that act upon it differently to form various prostanoids.

The POX and COX activities of COX-1 and -2 undergo irreversible suicide
inactivation in vitro during catalysis. POX and COX self inactivation is mechanism-based
because it proceeds from heme and tyrosyl radical intermediates that are formed during
the POX and COX reactions, respectively (3,156). The specific structural changes in the
holoenzyme that lead to suicide inactivation remain unknown. Mevkh et al. noted that
COX-l self inactivation is accompanied by dramatic changes in protein structure as is
manifested by the increased susceptibility of the inactive enzyme to trypsin cleavage and
the increased number of exposed histidine residues subject to chemical covalent
modification (162). It has not been established whether the COXs can undergo suicide

inactivation in vivo. If so, this could serve as an additional regulatory mechanism for

119

prostanoid synthesis. In this regard it would be interesting to investigate whether the
inactive forms of COX-1 and -2 are more susceptible to degradation compared to the
native forms. It is well established that misfolded or structurally damaged proteins that
are present in the ER can be targeted for degradation by a process known as ER-
associated degradation (ERAD) (l64-166,182-186). It is possible that COX-2 might
selectively undergo suicide inactivation at endogenous hydroperoxide or AA levels that
would be insufficient to inactivate COX-1. If indeed the inactive protein gets degraded,
this could partly explain the short-lived nature of COX-2 protein. The present study
examined the impact of COX catalysis in vivo on the protein stabilities of COX-1 and
COX-2. We demonstrate that COX-2 protein degradation that is mediated by the C-
terrninal 19 amino acid insertion can be substantially enhanced by COX substrates in
NIH/3T3 and HEK293 cells in a proteasome-independent manner. We also provide
evidence that substrate-dependent degradation of COX-2 requires a functional COX

active site and proceeds from substrate-induced inactive forms of the enzyme.

120

Experimental Procedures

Materials. Dulbecco’s modified Eagle’s medium (DMEM), fetal bovine serum
(FBS), ponasterone A, and tetracycline were obtained from Gibco/Invitrogen. Bovine calf
serum (BCS) was from Hyclone. Arachidonic acid (AA), eicosapentaenoic acid (EPA),
eicosadienoic acid (EDA), oleic acid (OA), linoleic acid (LA), 2-arachidonoyl glycerol
ether (2-AG ether), dihomo-y-linolenic acid (DHLA), (R/S)-flurbiprofen, (S)-
flurbiprofen, (R)-flurbiprofen, and NS-398 were from Cayman Chemicals. (S)-
flurbiprofen and NS-398 are time-dependent, irreversible inhibitors of cyclooxygenase
(COX) activity whereas (R)-flurbiprofen is a competitive COX inhibitor. Cycloheximide,
puromcyin, and glutathione (GSH) were purchased from Sigma. Calcium ionophore
(A23187), bradykinin, and M0132 were purchased from Calbiochem. [l-‘4C]
arachidonic acid (55 mCi/mmol) was from American Radiolabeled Chemicals.
Microsomal PGES-l (mPGES-l) was kindly provided by Dr. Michael Garavito of
Michigan State University.

Construction of Plasmids for Transfection. Recombinant ovine (ov) COX-1
cDNA and N-terminal hexahistidine-tagged ovCOX-l cDNA were subcloned into pIND
(Invitrogen). cDNAs for human (hu) COX-2, N-terminal hexahistidine-tagged huCOX-2,
and N-terminal hexahistidine-tagged murine (mu) COX-2 were subcloned into
pcDNAS/FRT/TO (Invitrogen). pIND is ecdysone-inducible whereas pcDNAS/FRT/TO
is tetracycline-inducible. After subcloning, the Quick-ChangeTM site—directed
mutagenesis kit (Stratagene) was used to create the following huCOX-2 mutants: 0533A
huCOX-2, N594A huCOX-2, and delS95-612 huCOX-2. Correct cDNA orientation and

mutations were confirmed by sequencing.

121

Cell Culture and Transfection. NIH/3T3 fibroblasts at early passage (<6 passages)
were cultured in DMEM supplemented with 10% BCS and 100 u/ml
penicillin/streptomycin. To induce COX-2 expression, the cells were first made quiescent
by serum starvation for 48 h in DMEM containing 0.2% BCS and thereafter treated with
DMEM supplemented with 20% FBS for 4 h.

Stable HEK293 transfectants of native or mutant COX-2 constructs were
generated using the Flp-In T-Rex tetracycline-inducible expression system (Invitrogen).
Native ovCOX-l or His-tagged ovCOX-l were stably transfected into the ecdysone-
inducible expression system (Invitrogen). All these transfections were conducted
according to the manufacturer’s protocol. Stably transfected HEK293 cells were cultured
in DMEM supplemented with 10% FBS, 100 u/ml penicillin/streptomycin, and the
appropriate pharmacological selective reagents. Inducible expression was achieved by a
24 h serum starvation in serum-free medium followed by treatment with 10 pg/ml
tetracycline or 10 uM ponasterone A (ecdysone analog) for an additional 24 h in normal
culture medium.

Fatty @ and Drug Treatments. Quiescent 3T3 cells were serum-stimulated for 4
h then treated with 50 uM cycloheximide (CHX) or 50 11M puromycin for different times
in the presence or absence of 20 M arachidonic acid (AA), 100 M flurbiprofen (FB),
20 M NS-398, or a combination of 20 11M AA and one of the following: 100 1.1M
flurbiprofen (FB), 20 11M NS-398, or 20 11M MGl32. Alternatively, after serum-
stimulation the cells were treated for different times with 50 11M cycloheximide (CHX) or
50 M puromycin in the presence or absence of 10 M A23187 or 10 1.1M bradykinin

with or without 100 11M flurbiprofen , 20 11M NS-398, or 20 11M MGl32.

122

HEK293 cells inducibly expressing wild-type or mutant cyclooxygenase
constructs were grown to 80% confluency, serum-starved for 24 h, and then treated with
10 ug/ml tetracycline or 10 uM ponasterone A in complete culture medium for 24 h to
induce expression. Afterwards, the cells were incubated for various times with 50 12M
puromycin to block translation in the absence or presence of the polyunsaturated fatty
acids AA, EPA, EDA, 2-A0 ether, LA, DHLA, 0A, or the COX inhibitor flurbiprofen.
To examine the effect of COX, proteasome, lysosomal, and ER to Golgi trafficking
inhibitors on AA-induced protein turnover, cells expressing native huCOX—2, delS94-612
huCOX—2, or N594A huCOX-2 were treated for 4 h with 50 11M puromycin with or
without AA in combination with one of the following: 20 11M M0132, 100 M
flurbiprofen, 50 11M (S)-flurbiprofen, 50 11M (R)-flurbiprofen, 20 11M NS-398, 50 M
leupeptin, 25 11M E64, or 5 jig/ml brefeldin A.

Microsome Prepa_ration 3nd Trypsin Cleajvagg. HEK293 cells stably expressing
delS95-612 huCOX—2 or N594A huCOX-2 were lysed by sonication in 0.1 M Tris-Cl
buffer, pH 8.0. The lysates were centrifuged at 5,000 rpm for 10 min at 4°C to precipitate
unbroken cells. Thereafter, microsomes were prepared by ultraentrifugation at 55,000
rpm for l h at 4°C. The microsomes were resuspended in 0.1 M Tris-Cl buffer, pH 8.0
and 200 pig of this resuspension was incubated in a 100 ill reaction mixture containing
0.1 M Tris-Cl, pH 8.0, 10 11M hematin, and 1 mM phenol in the absence of substrate, or
with 100 “M AA, or a combination of 100 11M AA and 100 1.1M PB. The incubation was
performed at 37°C for 4 h. 100 pig of microsomes were treated with 5 pg trypsin before or
after solubilization with 1% Tween-20. Trypsin treatment was performed at 37°C for 20

min.

123

Western Trmfer Blotting. After the appropriate treatments, NIH/3T3 and
HEK293 cells were harvested by centrifugation at 2000 rpm for 3 min. Cell pellets were
stored at -80°C until needed. RIPA lysis buffer (10 mM Tris, pH 7.2, 150 mM NaCl,
SmM EDTA, 0.1% SDS, 1.0% Triton X-100, 1% deoxycholate) containing a protease
inhibitor cocktail (Roche Applied Science) was used for cell lysis. Protein concentrations
were determined using the BCA protein assay kit (Pierce). The NuPAGE system
(Invitrogen) was used to resolve the proteins in the whole cell lysates on a 7% tris-acetate
polyacrylamide gel. After transfer to a nitrocellulose membrane, immunoblotting was
performed with the appropriate primary antibody. Horseradish peroxidase-conjugated
anti-rabbit or anti-mouse IgG antibodies (Bio-Rad) were used as secondary antibodies.
Immunodetection was performed using the Western Lighting Chemiluminescent kit
(Amersham Biosciences) followed by exposure to X-ray film. Densitometry analysis was
performed using Quantity One software (Bio-Rad).

Cyclooxygenase Inactivation Assay. Cultured HEK293 cells stably expressing
His-tagged ovCOX-l or His-tagged muCOX-2 were incubated for 10 min with different
concentrations of AA ranging from 2.5 M to 50 12M. After incubation the cells were
washed with 1X PBS (phosphate buffered saline) and harvested by centrifugation at 2000
rpm for 3 min. Cell pellets were resuspended in 0.1 M Tris, pH 8.0 buffer containing 1
mM EDTA and a protease inhibitor cocktail and lysed by sonication. Protein
concentration in the lysates was determined by BCA assay (Pierce). COX activity assays
were performed at 37°C by monitoring the initial rate of oxygen uptake using an oxygen

electrode as described previously (268). Reactions were initiated by adding cell lysate to

124

the assay chamber containing 3 ml of 0.1 M Tris-H01, pH 8.0, 1 mM phenol, 5 11M
hematin, and 100 M AA.

Microsomes were prepared from the lysates of His muCOX-2-epxressing cells by
ultracentrifugation at 55,000 rpm for 1 h at 4°C. The microsomes were resuspended in 0.1
M Tris-HCl, pH 8.0 containing 1 mM EDTA and 20% glycerol and a BCA assay (Pierce)
was perfomed to determine protein concentration. The microsomal fractions were then
used to perform COX radioactive TLC assays to examine COX product formation as
previously described (269). Briefly, microsomes (0.5 mg) were incubated in a 100-ltl
reaction mixture containing 0.1 M Tris-HCl, pH 8.0, 1 mM phenol, 10 11M hematin, and

10 11M [1-14C]arachidonic acid at room temperature for 1 min. To assay for PGE2
formation microsomal POES-l (mPGES-l) and 60 “M GSH were included in some of
the reaction samples. The reactions were terminated by adding 300 pl of an ice-cold
mixture of ethylether, methanol, and 0.2 M citric acid (30:4:1). The organic phase of the
reaction mixture, containing the radioactive products, was isolated and applied directly
onto a TLC silica plate at 4°C. The plate was developed in ethyl acetate, 2,2,4-
trimethylpentane, acetic acid, and water (110:50:20:100) and exposed to x-ray film, and

radioactive products were visualized by autoradiography.

125

LESLIE

COX Inhibitors Retard COX-2 Degradation in Nil/3T3 cells. About nine years
ago Smith and co-workers made the interesting finding that serum-stimulating quiescent
NIH/3T3 cells in the presence of the non-selective NSAIDS flurbiprofen and aspirin, or
the COX-2 selective inhibitor NS-398 dramatically enhanced inducible COX—2 protein
expression without affecting COX-2 mRNA levels ((14), Arakawa and Smith,
unpublished results). In contrast, constitutive COX-1 protein expression was not affected.
These results suggested that COX inhibitors promote the stability of COX-2 protein in
NIH/3T3 cells by preventing the degradation of the enzyme. After successfully
reproducing these findings (Fig. 26a) 1 tested hypothesis that NSAIDS inhibit the rapid
degradation of COX-2. In a cycloheximide chase experiment (R/S)-flurbiprofen, (R)-
flurbiprofen, and (S)-flurbiprofen nearly completely retarded COX-2 degradation in
NIH/3T3 cells within the time frame of the experiment (Fig. 26b). There is no evidence
to date that NSAIDS directly inhibit the 26S proteasome. Since proteasome inhibitors
partly inhibit COX-2 degradation in 3T3 cells (257), my findings suggest the possibility
of a second non-proteasome—dependent pathway for COX-2 protein turnover. Moreover,
flurbiprofen inhibits COX-2 degradation in 3T3 cells to a greater extent than the
proteasome inhibitors M0132 and epoxomicin (data not shown). Therefore, it is likely
that COX inhibitors prevent both the proteasomal and non-proteasomal degradation of
COX-2 in 3T3 cells.

Proteasomal Degradation of COX-2 in HEK293 cells is not inhibited by NSAIDS.
We have previously reported that heterologously expressed COX-2 undergoes

proteasomal degradation in HEK293 cells in a manner that is dependent upon the C-

126

terminal 19 amino acid cassette of the protein (257). To determine if COX inhibitors
stabilize COX-2 in HEK293 cells, COX-2 protein degradation was examined in these
cells in the presence or absence of treatment with 20 11M NS-398 or 100 pM flurbiprofen.
We were surprised to find that, unlike 3T3 cells,ineither NS-398 nor flurbiprofen
significantly stabilized COX-2 in HEK293 cells (Fig 26c). Furthermore, inducible COX-
2 expression in these cells was not enhanced by treatment with 20 M NS-398 (Fig. 26d).

Serum-stimulation of quiescent 3T3 cells results in the coordinate induction of
COX-2 and the mobilization of free AA due to activation of cPLA20t (77). Therefore, it is
reasonable to expect that serum-treatment of quiescent 3T3 cells will stimulate COX-2
COX catalysis. During catalysis COX-2 may undergo suicide inactivation, a process that
could initiate the rapid degradation of the enzyme. Resting HEK293 cells lack detectable
phosholipase A2 activity so that heterologously expressed COX-2 is more likely to be in
its latent form (33,58). This may explain the discrepancy that is observed in the two cell
lines with respect to the effect of COX inhibitors on COX-2 protein turnover. It is
possible that COX-2 is degraded in serum-treated 3T3 cells by at least two distinct
mechanisms: (a) COX substrate-induced protein turnover which is inhibited by COX
inhibitors; and (b) a proteasome-dependent basal degradation of the resting enzyme that
is directed by the C—terminal 19 amino acid cassette.

Substrate-dependent Degraflttion of COX-2. To test the hypothesis that COX
fatty acid substrates can induce COX-2 protein turnover, we analyzed the effect of

treatment with exogenous AA on the degradation of native huCOX-2, delS95-612

127

26A.

 

 

 

 

 

 

100 M FB
0 2 4 6 10 2 ..4 7 6 7 10 '(h)
COX-2 Wait first. “mm A
Actin _ ,
V and amp fl Um. V’W" 7". a
1.6
l' \

E 1.2 4 I \ \-

W

g -

E

9

3

3- a

0 2 4 6 8 10 12

Time after induction (h)

Q—O No inhibitor

|___.| molt-MFR

128

 

26B' 100 FB

2Q“ 0‘ 2 8 2 8 TimewithCHX (h)

i. .4 -I—‘hu uni-q
.r

cox-2 ' a... 2. 4i 2......

Actin ..,,.- 2.,” .4...

~w-«w m

100 “M FB
0 2 8 2 8 (h)
Densitometry (%) 100 58 35 102 103

 

 

 

 

 

 

 

 

 

50 M 500 |.l.M 50 M
No inhibitor (Lu-FR (R)-FB (S)-FB
0 2 4 2 4 2 4 2 4 (h)

’L'Jk- “an" .‘ern 1!: W.» \‘L‘J ' -1124- ‘ .?;~.Wr+'” ’

. .L-mm.) :‘J‘ET‘J-I' 1' .‘ MeM‘i- - in"".‘-Lj_7'73

ow ' - - . 't .170;va rue-Te: ism-er» nu.‘ ma. iv-uv-vv M'\‘

rib-.2 - ~.--o nay-27:; mh,. — -.. Inn nun-urn... I- -

I ' '- wilful .‘p .-.I .bfimqu‘ '2 .rfi'ifflL‘r

ACtln 3“ _" ‘ a a“ (tori-.1 w “I W“

l
.
. "‘ynra ‘. - . '- .-~., "" . - u - ran—n!

No so M 500 11M so itM
Inhibitor (R)-FB (R)-FB (sire

0 ‘2 4| ‘2 4| ‘2 4“2 4‘(h)
Densitometry (%) 100 39 36 65 83 80 69 65 48

 

 

 

 

 

 

 

 

 

 

 

 

129

100 FB

26C.

  

100M“

 

 

2 24 (h)
I Densitometry(%) I 100 I 96 I 48 I 46 I 18 I 100 I 62 I 37 I 17 I
100 NS-398
0 2 4 12 24 2 4 12 24 (h)
huCOX-2 . “ w “P‘. W (gm, ~ ..
Actin r g u I!“ in; w ‘2! '5
.3 7 W M V A— “— A20MNS-398
o 2 4 12 24 2 4 12 24 (h)

 

[Densitomeu'y(%) I 100 I 86 I 77 I 50 I 21 87 I 77 I 47 I 30 I

 

D. l 2
huCOX-2 tiff 4'" 1. tet-lnduced

Actin -" 2. tet-induced + 20 “M PIS-938

Figure 26. Cyclooxygenase inhibitors prevent the rapid degradation of COX-2 in serum-
treated NIH/3T3, but not HEK293, cells. (A) NIH/3T3 cells were made quiescent by 48 h
serum-starvation and serum—induced for different times in the absence or presence of 100 11M
flurbiprofen (FB). (B) Quiescent (Q) 3T3 cells were serum-induced for 4 h then treated with 50
11M cycloheximide (CI-1X) with or without 100 M FB, 50 M (R)—FB, 500 uM (R)—FB, or 50
M (S)—FB. (C) HEK293 cells stably expressing huCOX—2 were serum—starved for 24 h and
treated with 10 jig/ml tetracycline in normal complete medium for an additional 24 h to induce
COX-2 expression. Upon COX—2 induction, cells were treated with 50 11M puromycin to block
translation with or without 100 M FB or 20 11M NS-398. (D) COX expression was induced in
HEK293 cells with tetracycline (tet) in the absence (1) or presence (2) of treatment with 20 11M
NS-398. After the treatments in (A )-(D) the cells were harvested, lysed, and analyzed by Western
blotting for COX-2 and actin.

130

huCOX-2, and N594A huCOX-2 in HEK293 cells. The basal degradation of native
huCOX-2 was dramatically enhanced by treatment with 20 M AA (Fig. 27a, panel I).
An AA dose response of COX-2 degradation revealed that protein turnover could be
substantially enhanced by fatty acid concentrations as low as 5 11M (Fig. 27b, panel I).
The mutants delS95-612 huCOX-2 and N594A huCOX-2, which are usually stable under
basal conditions, also rapidly degraded upon treatment with 10 M AA (Fig. 27a, panels
11 and III, and Fig. 27b, panel II). Therefore, substrate-induced degradation of COX-2 is
not mediated by the C-terminal l9-amino acid cassette. Since removal of the cassette
does not affect COX-2 catalytic activity, these results imply that substrate—induced
degradation could be a consequence of substrate turnover at the COX active site. To test
this hypothesis, we examined the effect of COX inhibitors on substrate-induced COX-2
degradation in HEK293 cells. We also analyzed a panel of polyunsaturated fatty acids
only some of which are COX-2 substrates. Flurbiprofen, (R)-flurbiprofen, (S)-
flurbiprofen, and NS-398 significantly blocked the AA-induced degradation of huCOX-Z
in 293 cells (Figs. 27c and d). Moreover, these NSAIDS selectively inhibited COX-2
degradation that was substrate-dependent without affecting the basal degradation of the
enzyme (Figs. 27c and d). In contrast, the degradation of delS95-612 huCOX-2 and
N594A huCOX—2, which was only observed upon exogenous AA treatment, was nearly
completely blocked by flurbiprofen and (S)-flurbiprofen (Fig. 27e).

In order to be utilized as substrate by COX-2, exogenous AA would have
to penetrate the cell, move to the ER membrane by diffusion or carrier-mediated
transport, and be delivered to COX-2 on the luminal side of the membrane. This transport

and delivery process should be very rapid since prostaglandin product formation is

131

known to reach a plateau within minutes after treatment with exogenous AA in cultured
cells expressing COX activity (270). Therefore, a brief challenge of cells with exogenous
AA should be sufficient to induce COX-2 degradation. HEK293 cells stably expressing
huCOX—2 were exposed to 20 11M AA for 5 min. or 4 h. After aspirating the AA-
containing medium, cells that were treated for 5 min with AA were washed with 1X PBS
then treated with the protein translation inhibitor puromycin with or without flurbiprofen
for an additional 4 h. It was observed that regardless of the duration of AA treatment (5
min or 4 h), COX-2 degradation was enhanced to a similar extent (Fig. 27f). Flurbiprofen
added after the 5 min treatment with AA did not protect the enzyme from degradation as
well as when the inhibitor was added together with substrate. These observations indicate
that a short-term incubation of cells with AA, during which prostaglandin products are
expected to be formed, is sufficient to further destabilize COX-2.

Like native huCOX-2, the N-terminal hexahistidine (His6)-tagged huCOX-2 was
susceptible to AA-induced degradation that could be inhibited by NSAIDS (Figs. 28a and
b). The IC50 values for the inhibition of huCOX—2 COX activity by flurbiprofen and NS-
398 are 0.51 M and 1.77 11M, respectively (271). I found that 10 M PB and 500 nM
NS-398 significantly inhibited the degradation of Hisn-tagged huCOX-2 induced by 15
11M AA (Fig. 28b). Just like AA, the COX-2 polyunsaturated fatty acid (PUFA)
substrates EPA and DHLA enhanced COX-2 degradation albeit at higher concentrations
than AA (Figs. 28c and d, panel 11). In contrast, 2-A0 ether, a derivative of the COX-2
substrate 2-arachidonyl glycerol, and OA did not affect COX-2 protein turnover (Fig.28d,
panels II and III). Similarly, LA and EDA which are known to be poor COX-2 substrates

did not induce COX-2 degradation even at concentrations as high as 50-100 M

132

M
27A.(1) 0 2 4 12 24 2 4 12 24 (h)

huCOX-2 ,

 

Actin

 

(11) 20mm
0 2 4 12 2 4 12 (11)

1-3.1:. 7 «In; ,. ,._ 6‘ :2“.— mm: .v‘fiau? 'qu
4‘

«1595-612 i .t .,,5 ,3
huCOX-2 i; ”'u‘ “‘ . . I "
mu: 9- ~_ ii". '11-'— main-um un- rub- tuna—Ind

say i.-. m racy. aerrnxxri-Ir,.n.u.vv‘";’,q
u.

it: ‘- l‘ M 1‘
Actin i '1'” m

.1: <5
#5:.rn'4 :2: ‘ .4 ’

c.41- -. Av

 

(In) 10 111! AA

024122424122401)

 

 

 

«a. _ "7

it '~ ‘ ‘ . I" I
N594A huCOX-Z {mfi ” gird-d
: . . T- ~ “le ‘ 2&1 _° , . 1' .y‘

°:"- .ru: » ‘3' 7 ' .,-:~— 4.1.1: ”3" "~52

 

 

 

Actin

 

133

27B. (1)

 

 

 

1 2 3 4 5
1. 24hinduction
huCOX-Z «fl- 4.. 2. “ + 4hpuromycin
3. “ + “ +5pMAA
4. “ + “ +10pMAA
Actin "fl““* 5, “ + “ +20IJMAA
120
2"
E 100
:E': 80
h 60
=
E 40
O
3: 20
e\° 0
(11M)
B. (II)
-1_.2....3 4 5.
delS95-612 , 7 .,,. . ,_ 1 24hinduction
huCOX-2 . T‘Vf‘fi" " ”' ° ' 2. “ + 4hpuromycin
2.. .. —.__.'.'.' _' v . -.- 1.: 3. u + ‘6 +5 "M AA
‘ ‘ " ' 4 “ + “ +10uMAA
Actin 5 “ + “ +20llMAA

 

 

 

0 5 10 20 (11M)

134

27C.

' huCOX-2

. vl 1 '23.. 1‘ . I ‘, 4 , .' hi. .
Actin

1 2 3 4 5 6 7
Densitometry (%)I 100 J 51.2 Il7.6 I 26.4 I 39.4 I 45.6 I 44.3 I

 

 

1. 24 h induction
2. “ + 4 h puromycin
3. “ + “ + 10 pM AA
4. “ + “ + “ + 20 nM M0132
5. “ + “ + “ + 50 11M R-FB
6. “ + “ + “ + 50 11M S-FB
7. “ + “ + “ + 100 M FB
D. 1 2 3 4 5 6
ram. .... .. _. 2......“ . 5+4

‘1. 'i

,,_ F I“. e- 4 an. on Iraq

 

if 1". P_-rr....'.::_.!!.-.l
1. 24 h induction
2. “ + 4 h puromycin
3. “ + “ + 20 M AA
4. “ + “ + “ + 20 11M NS-398
5. “ + “ + “ + 100 “M FB
6. “ + “ + “ + 20 M M0132

135

27E.
1234567

 

 

 

 

 

 

 

 

 

 

 

delS95-612 ..._, .2 M . ....
huCOX-2 l. 24hinduction
. 2. “ + 4 h puromycin
ACtIn ~~~hflflww 3. 66 + u + ZOILMAA
' 4. “ + “ + “ + 20uMMGl32
.. ..5. “ + “ + “ + 10011MFB
N594A "
6. “ + “ + “ + 50 R-FB
huCOX-z, ~-- ~- --«---~ ,7. « . .. + « + some.
Aetin -"' 1"” *‘V W H W w.
F.
1 2 3 4 5 6
hUCOX-Z ‘5‘“ race: ' q .4 ~_.
Actin ilr“
E- V -:_'_. ‘.'. . i' 1.1322“ ’1 j. .
l 2 3 4 5 6
Demimmetl'y (%) 100 78.3 50.7 77.0 36.3 37.4
1. 24h induction
2 “ + 4 h puromycin
3. “ + “ + 2011MAA
4. “ + “ + “ + 10011MFB
5 “ + 5minpuromycin+20pMAA; 1X PBS wash;4hpuromycin
6 6‘ + 66 + 66 66 o 66 + lmuMFB

Figure 27. COX-2 degradation in HEK293 cells is enhanced by exogenous AA
treatment. (A) After serum—starvation and challenge with tetracycline to induce
expression HEK293 cells stably expressing native huCOX-2 (I), delS95-612 huCOX-2
(H), or N594A huCOX—2 (III) were treated with 50 p.M puromycin for the indicated times
with or without the addition of 10 11M or 20 11M AA. Alternatively, the cells were treated
for 4 h with puromycin, a combination of puromycin and different concentrations of AA
(5 M, 10 11M, or 20 12M) (B), or a combination of puromycin, AA, and the indicated
inhibitors (C, D, and E). (F) After a 24 h COX-2 induction with 10 jig/ml tetracycline
cells were treated with 50 12M puromycin and 20 p.M AA for 5 min or 4 h in the presence
or absence of 100 12M FB. Following the 5 min treatment with AA, cells were washed
with PBS and treated for an additional 4 h with puromycin in the presence or absence of
100 “M PB. After the treatments in (A)-(F) protein levels were analyzed by Western
blotting and densitometry analysis. The results in HO) and B(II) are representative of
three independent experiments and the graph error bars denote 1 SE of the mean. The
result in F is representative of two different experiments.

136

(Fig. 28d, panels I, II, and 111). Thus, COX-2 degradation in 293 cells is selectively
induced by COX PUFA substrates. Collectively, these results suggest that substrate
binding to the COX active site is a prerequisite for the substrate-dependent degradation of
COX-2.

6533A huCOX-g is Refractory to Substrate-dependent Degrafltjgl. In a standard
in vitro oxygen electrode assay, GS33A huCOX-2 has less than 5% of the COX specific
activity of native enzyme with AA as substrate (17,272). We stably expressed 6533A
huCOX-2 in HEK293 cells and analyzed its degradation profile after treatment with
exogenous AA. Under basal conditions GS33A huCOX-2 degraded with a half-life that
was similar to that of the native huCOX-2 (Fig. 29a). However, this mutant was resistant
to degradation that was induced by COX substrate (Figs. 29b and c). Our findings suggest
that a functional COX active site is required for substrate-induced degradation of COX-2.

Substrate-dependent COX-2 Degragrtion in NIH/3T3 cells. We attempted to
reproduce the enhancement of COX-2 degradation by fatty acid substrate in a different
cell type. Serum-stimulation of quiescent 3T3 cells induced endogenous COX-2
degradation which was enhanced by treatment with 20 MM exogenous AA (Fig. 30a,
panels I and II). Flurbiprofen significantly blocked both the endogenous degradation of
COX-2 and AA-induced degradation of the protein (Fig. 30b). We also conducted
experiments to determine if further stimulating endogenous AA release with bradykinin
or A23187 in addition to serum-induction would also cause an enhancement of COX-2

degradation. In one experiment, treatment with 10 M A23187 modestly enhanced COX-

2 degradation albeit to a similar extent as treatment with 20 M AA (Fig. 30c). The

enhancement of COX-2 degradation by both A23187 and AA was inhibited by NS-398.

137

28A. MA... AA. _AA+FB_ AA+NS-398
024 24 24,214 a.)

HisshuCOX-Z a...” . “A w

B.
l 2 3 4 5 6 7 3
His; huCOX-2 ~n « ‘- 3 , fl.“
1. tet-induced
2 “ + 4 h puromycin
3 “ + “ + 20 “M AA
4. “ + “ + “ + 10 pM FB
5. u + ‘6 + “ + 30 ”M FB
6 u + u + u + 100 [M FB
7 “ + “ + “ + 500 nM NS-398
s “ + n + “ + 20 M NS-398
C. AGE)— EPA
q 2 __ 10 100 7, , 2 ‘10 50. 100
His, huCOX-Z.“ I" ,_ . _fi, _. ...., ..,_

L‘._‘:.-.,;J_-_.... 2 - _ __ _1..

Actin ‘.

EPA

 

9‘ Rundnlng

   

 

o 2 10 100(uM) o 2 10 50100 (m)

138

28D (1) AA LA

0 3.75 7.5 15 100 3.75 7.5 15 100 (11M)

 

 

 

o 3.75 7.5 15 100 3.75 75 15 10? (pM)
FDensitometry(%) I 100 I 75.2 I 39.7 I 43.7 I 31.6 I105.3 I100.0I123.7I105.7I

 

 

D(II) AA DHLA EDA 2-AG ether
,0 5 20 2_ .10 so 100 10 5010010 50 (11M)

HisshuCOX-Zhwg . ‘1.“ 1‘2?" rfl‘fl‘ an a a .1“
AA I DHLA EDA IL? 2-AG ethe'l'

0 {s 20‘6 2 10 so 100\ [10 so 100\ ’10 501 (11M)
IDensitometry(%)I 100I 48.2 I 17.3 I 68.5 I795 50.0 I58.0I 67.1 I 74.1 I 63.2 I 76.2 I 84.7 I

 

 

 

III 2-AG ether EDA 0A
( ) 0 2 5 10 20. 2 10 50 100 2 10 50 100 (11M)

';’_‘1 . . .1 . . .7, .4 . . _...“_. ___._.. __.

r
. fa»

2-AG ether EDA 0A

 

 

o 6 s 10 20\ f o 10? G 1 50 R
Densitometry(%)I 100I 126.l I 132.3I 111.7 I115.6I94.9 I89'9t‘7'3 163.7 I 146.2 I 145.0 I 110.3I12:7“IM

 

Figure 28. COX-2 degradation is selectively induced by PUFA COX substrates for the
enzyme. N-terminal Hiss-tagged huCOX-Z-expressing HEK293 cells were treated as in 27(A) to
induce protein expression. (A) Cells were treated for the indicated times with 50 11M puromycin
with or without 20 11M AA, or a combination of puromcyin, AA, and the indicated concentrations
of F8 or NS398. (B) After induction of His huCOX-2 (1), cells were treated for 4 h with 50 [AM
puromycin only (2), or in combination with 15 11M AA (3) and 10 M PB (4), 30 11M FB (5), 100
[AM FB (6), 500 nM NS-398 (7), or 20 M NS-398 (8). In (C) and (D) puromycin treatment was
followed by treatment with different concentrations of AA, linoleic acid (LA), eicosapentaenoic
acid (EPA), 2-arachidonyl glycerol ether (2-AG ether), dihomo-Y-linolenic acid (DHLA),
eicosadienoic acid (EDA), and oleic acid (0A). COX-2 protein levels were analyzed by Western
blotting and densitometry. The result in C is representative of at least three independent
experiments and the graph error bars denote 1 SE of the mean.

139

29A.
20 AA

0 2 4 12 24 2 4 12724 (b)

1' ..:r _. a. - an n a. .1 . —~ 1:: a. 1,3- ,. u- u.

G533A huCOX-2 "u .1. i— q... w I . U" U .11 a“-

B- 1 2 3 4 5
f 3‘" i " i ' "9‘3 1. 24hinduction
G533A huCOX-2 I P “h w w . i 2. “ + 4 h puromycin
r... ‘11" ‘ , #7:)": am; m1 3, “ + “ + 5 “1M AA
"11‘ *—‘ 1 ‘ 4. “ + “ + lOuMAA
Actin flflkg‘flal 5. “ + “ +20uMAA
C.

 

% Protein remaining

@ Native COX-2
l G533A cox-2 "'
0 5 10 20 (11M AA)

Figure 29. The G533A cyclooxygenase-inactive COX-2 mutant is not susceptible to
substrate-dependent degradation. (A) Degradation of GS33A huCOX—2 was examined
with or without addition of 20 11M AA. (B) Cells expressing G533A huCOX-2 were
treated for 4 h with 50 pLM puromycin plus different concentrations of AA (5 11M, 10 11M,
or 20 1.1M). Protein levels were analyzed by Western blotting. (C) Densitometry was
performed on the levels of native huCOX-2 or G533A huCOX-2 and normalized to those
of actin. Densitometry analysis for each protein is based on at least three independent
experiments. The error bars denote 1 SE of the mean.

 

 

 

In contrast, bradykinin at 10 MM did not facilitate COX-2 protein turnover. It is possible
that A23187 may cause more release of endogenous AA than bradykinin and that the
amount of bradykinin-induced free AA may not be sufficient to facilitate COX-2
degradation in serum-stimulated 3T3 cells.

SubstLate-dependent Degra_dz_1tion of COX-1. COX-l is known to be a very stable
protein (257). This is consistent with the observation that NIH/3T3 cells that have been
made quiescent by serum deprivation for 48 h continue to express the enzyme.
Furthermore, stimulating quiescent 3T3 cells with serum will induce COX-1 mRNA (22).
However, serum-stimulation of 3T3 cells in the presence of flurbiprofen does not lead to
an accumulation of COX-1 protein. Collectively, these results imply that the constitutive
expression of this COX isoform observed in many tissues is likely to be mainly due to the
stability of the enzyme rather than a continuous balance of protein synthesis and
degradation. Experiments were performed to determine if COX-l protein turnover could
be induced by COX fatty acid substrate. In NIH/3T3 cells, treatment with 20 11M AA did
not affect the stability of endogenous COX-1 protein (Fig. 31a). This was a surprising
result because treatment of quiescent 3T3 cells with 10 p.M AA results in prostanoid
product formation due to COX-1 activity. Therefore, at the AA concentrations that were
used in our experiments COX-1 is catalyzing oxygenation. In HEK293 cells, treatment
with _>_20 11M of AA resulted in a modest loss of COX-1 protein (Figs. 31b and c). If
indeed COX-2 degradation due to AA challenge is a consequence of suicide inactivation
during COX catalysis, it is possible that COX-l may be more resistant to COX substrate-
induced inactivation in vivo compared to COX-2. Alternatively, the inactive form(s) of

COX-1 formed during COX catalysis may be less susceptible to degradation.

141

30A. (1) 20”,“,
0 0.5 l 2 0.5 l 2 Time with puromycin (h)

COX-2

Actin

 

 

 

$6 Protein rornalnlng
I
F

 

 

0 0.5 1 1.5 2 2.5
Tlm 0 after purom yeln (h)

._. No substrate

I-——-I zopMAA

142

30A. (11) MA

 

 

 

 

 

 

0 1 2 1 2 Time with CHX (h)
COX-2 W Irv-‘1 g,»
Actin we: , ,, “a, .1. 1..»
120
100
O
.E
.E 80.
C
E .0.
E ‘0
s 40‘ \
a2 20. i— _ _ _l
0 . fi _ fl
0 0.5 1 1.5 2 2.5
Time after CHX (h)

P * No Substrate

 

F—_—. 201LMAA

143

30B.

 

 

_h
N
O

_A
O
o

804

60-

 

% Protein remainlng

40-

20-

 

 

 

0 0:5 1 1:5 2 2.5
Tlme after puromycin (h)
H No substrate
H 20 11M AA

l———-I ZOuMAA+100uMFB

A---i 100uMFB

M1 '2 3 4 5 6 7 8 9
30C. .3. ..___ 3................ .._3

COX-2 Ihufi. «it a“.

Actin 1- w

r 2 3 4 5 6 7 s 9
Densitometry(°/ o) 100 73 80 87 94 62 76 59 80

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 30. The enhancement of COX-2 degradation in NIH/3T3 cells by exogenous
and endogenous AA is inhibited by NSAIDS. (A and B) Serum-starved, quiescent 3T3
cells were serum-stimulated for 4 h to induce COX-2 expression. Thereafter, cells were
treated with 50 11M puromycin or 50 pM CHX for the indicated times in the presence or
absence of 20 MM AA, 100 M FB, or a combination of both AA and PB. After Western
analysis, densitometry was performed on the levels of COX-2 and normalized to those of
actin. (C) After inducing COX-2 expression (1) as in A and B, cells were treated with 50
12M CHX (2) in the presence or absence of the following: 20 M NS-398 (3), 10 11M
bradykinin (4) , 10 M bradykinin + 20 [AM NS398 (5), 10 11M A23187 (6), 10 |.l.M
A23187 + 20 uM NS-398 (7), 20 11M AA (8), or 20 11M AA + 20 [AM NS-398 (9).
Densitometry was performed on COX-2 protein levels and normalized to those of actin.

31A.
20 AA
0 l 2 l 2 Time with puromycin (h)
cox-1 m5} “'3 :4 3‘32. A

20 EM AA
0 1 2 1 2 Timewith CHX (h)

COX'I . , ,3, ‘ _..- “'3‘ “*0:- Av

we 4‘! ~- '9. 9..-.“ *M“ 33....

Actrn .. ,3. 3.... ,3. ...., .3...

145

 

31B.
0 2 10 25 50 (uMAA)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ovCOX—l
ACtin 9""?! $31-an ~3: ~13 . -e~. $3.3:
0 2 10 25 50
Densitometry (%) 100 81 78 74 69
C3 120
100 L; %
.‘s‘ T ‘ ~
5 80- \ ‘
0 ‘
Em i
3'
0 4° ‘
32
20 i
0 .
0 2 4 6
Time (h)

.__. No substrate
'- — " '" i 20 [1M AA

(LAM AA)

Figure 31. COX-1 stability is not significantly affected by exogenous AA. (A)
Quiescent 3T3 cells were serum-stimulated for 4 h then treated with 50 MM puromycin or
50 M CHX for the indicated times with or without the addition of 20 [1M AA. (B) 293
cells stably expressing ovCOX-l were treated with 50 11M puromycin for 4 h with or
without the addition of different concentrations of AA. (C) Densitometry of the effect of
20 mM AA on ovCOX—l stability in 293 cells. Quantitative analysis is based on six

independent experiments. The graph error bars denote 1 SE of the mean.

146

Substrate-dependent Degradation of COX-2 is not Proteasome- or Lvsosome—
dexndent. We tested an inhibitor of the 26S proteasome and selective inhibitors of
lysosomal degradation in an attempt to identify the pathway responsible for substrate-
dependent COX-2 protein turnover. The 268 proteasome inhibitor M0132 failed to
inhibit substrate-induced COX-2 degradation in both 3T3 and 293 cells (Figs. 27c, 27d,
27e, and 32a). Similarly, the lysosomal degradation inhibitors E64 and leupeptin, and
brefeldin A, an inhibitor of ER to Golgi trafficking, failed to prevent AA-induced COX-2
degradation in 293 cells (Figs. 32a and b). Therefore, substrate-dependent degradation of
COX-2 does not appear to require the proteolytic activity of the proteasome or lysosome.

Experiments were carried out to determine if the protease(s) responsible for
cleaving COX-2 in a substrate-dependent manner was ER-associated. Microsomes
prepared from 293 cells expressing delS95-612 huCOX—Z or N594A huCOX-2 were
incubated with or without 100 “M AA in a reaction buffer containing 10 uM hematin and
1 mM phenol. After 4 h at 37°C the samples were analyzed by Western blotting to
determine whether there was loss of COX-2 protein as a consequence of substrate
treatment. There was no observable change in the levels of delS95-612 huCOX-2 or
N594A huCOX-Z in microsomes that were exposed to AA (Fig. 32c). It is possible that
the microsomes that we prepared were not sufficiently intact to retain any ER luminal
protease(s) responsible for COX-2 proteolytic cleavage. To test the integrity of the
microsomes, I treated solubilized or non-solubilized microsomes with trypsin. Limited
digestion of COX-2 with trypsin yields a single cleavage at its C-terminus that produces a
~2-3 kDa C-terminal fragment. The C-terminal trypsin cleavage site is within the epitope

for the primary antibody used to detect COX-2. Since COX-2 is an ER luminal protein it

147

should not be cleaved by trypsin if the microsomes are intact. However, both delS95-612
huCOX-Z and N594A huCOX-2 were proteolytically cleaved by trypsin in non-
solubilized microsomes (Fig. 32d). These results indicated that the prepared microsomes
were permeable and/or wrong-side-out. Therefore, we cannot eliminate the possibility
that there could be an ER-associated protease involved in substrate-dependent COX-2
degradation. The protease is likely to be a luminal enzyme or it may be ER membrane-
associated but requires luminal co-factors for its catalytic activity.

COX Self-inactivation Preceeds Substgte-depenggnt COX-2 Degradation. Both
COX-1 and COX-2 are known to undergo irreversible, mechanism-based COX
inactivation during in vitro catalysis (3,156). It is possible that the self-inactivation of
COX-2 in HEK293 cells and NIH/3T3 cells treated with AA would accompany the
formation of cyclooxygenase oxygenation products. If indeed the enzyme self-inactivates
in vivo during catalysis, it would reflect damage at the COX active site. Structurally
damaged COX-2 would likely be more susceptible to degradation than the native
enzyme, which could explain why AA treatment facilitates COX-2 protein turnover in
HEK293 and NIH/3T3 cells.

I decided to examine whether COX-1 and COX-2 undergo self-
inactivation in vivo during COX catalysis. For this purpose, I utilized HEK293 cells
stably expressing N-terminal His-tagged versions of ovCOX-l (His ovCOX-l) or murine
COX-2 (His muCOX-2). These cells were treated for 10 min with varying concentrations
of AA (0, 2.5, 10, 25, and 50 MM). Thereafter, the cells were washed with PBS, pH 7.4
and centrifuged at 2000 rpm for 3 min. Cell pellets were resuspended in 0.1 M Tris, pH

7.4 and sonicated. The COX activity of the whole cell lysates was then assayed using an

148

32A.

  
 

 

 

 

 

delS95-6l2 . .
huCOX-Z tet-mductlon
“ + 4 h puromycin
Actin “ + “ +20pMAA
“ + “ + “ +50 Leu
N594A “ + “ + “ +25 ”um E64p
huCOX-2 “ + “ + “ +20 uMMGl32
“ + “ + “ +20 pMNS-398
Actin F3
B.
huCOX-2
Actin
qr “WWT‘CF ._. “
delS95-612 £533; 33. ' ”‘1. .' f3
huCOX-2 If. ‘ "5 J _ “i
h ' 1" aygffiéézsfi .1 A
. 5;?1'riaj-‘fifnf‘rfi-‘fi‘g
Actin fififlt-flg 3
,3. .... 3-..“, ,3 4 ‘_l
l. tet-induction
2. “ + 4 h puromycin
3. “ + “ + 20 [1M AA
4. “ + “ + “ + 5 ug/ml BFA
5. “ + “ + “ + 5 ug/ml BFA
6. “ + “ + “ +20uMNS-398

149

32C.

 

123456

N594A ""” ' -‘ 71
huCOX-2 Eh," I

      

TV, “ ' o

r ' 1
:ng93X6-1223 "HHHH— 3 1,4. No substrate
"7 2,5.100 [AM AA
3,6.100uMAA-l-100uMFB

32D.
1 2 3 4
3......3 _3 .3__... “.3. l. unsolubilized
N594A huCOX-2 ~ .. r 2. unsolubilized + trypsin tmt.
. 7 1 ' . 3. solubilized

4. solubilized + trypsin tmt.

Figure 32. Inhibitors of proteasomal and lysosomal degradation do not prevent
substrate-dependent COX-2 protein turnover. (A) HEK293 cells stably expressing
delS95-612 huCOX-2 and N594A huCOX-2 were treated as in 27A to induce protein
expression (1). Thereafter, cells were treated for 4 h with 50 pM puromycin (2) or in
combination with 20 pM AA (3), 20 pM AA + 50 pM leupeptin (4), 20 pM AA + 25 pM
E64 (5), 20 pM AA + 20 pM MGl32 (6), or 20 pM AA + 20 pM NS398 (7). (B) Cells
stably expressing native huCOX—2 and delS95-612 huCOX-2 were treated as in 27A to
induce protein expression (1). This was followed by treatment with 50 pM puromycin
only (2) or in combination with 20 pM AA (3), 20 pM AA + 5 pg/ml brefeldin A (4), 20
pM AA + 5 pg/ml brefeldin A (5), or 20 pM AA + 20 pM NSB98 (6). (C) Microsomes
were prepared from 293 cells stably expressing delS95-612 huCOX-2 or N594A huCOX-
2 as described in the text. Thereafter, the microsomes were resuspended in 0.1 M Tris-Cl
buffer, pH 8.0 and 200 pg of this resuspension was incubated in a 100 pl reaction mixture
containing 0.1 M Tris-Cl, pH 8.0, 10 pM hematin, and 1 mM phenol with no substrate
(1,4) or with 100 pM AA (2,5) and 100 pM FB (3,6). The incubation was performed at
37°C for 4 h. (D) After resuspension, 100 pg of microsomes were treated with 5 pg
trypsin before or after solubilization with 1% Tween-20. Trypsin treatment was
performed at 37°C for 20 min. After the treatments in A-D protein levels were analyzed
by Western blotting.

150

oxygen electrode. Microsomes were also prepared from the whole cell lystates and
assayed for prostaglandin product formation by radioactive TLC. There was a dramatic
concentration-dependent decrease in COX activity of His6-tagged muCOX-Z-expressing
whole cell lysates (Fig. 33a), which could not be a result of a decline in His6-tagged
muCOX—2 protein levels during the 10 min duration of AA pretreatment (data not
shown). In contrast, the decline in COX activity observed with Hisé—tagged ovCOX-l-
expressing HEK293 cells was less dramatic (Fig. 33a). For instance, His6-tagged
muCOX-2-expressing lysates pretreated with 10 pM had a ~40% decrease in specific
activity compared to only ~20% for Hing-tagged ovCOX-l-expressing lysates pretreated
with the same concentration of AA. The observations made with the oxygen electrode
assays are consistent with results from radioactive COX TLC assays which
demonstrated that prostaglandin product formation of Hisa-tagged muCOX-Z-expressing
microsomes prepared from cells that were treated with 10 pM or 25 pM AA was
significantly reduced (Fig. 33b). Overall, these findings indicate that, at least in HEK293
cells, both COX-1 and COX-2 undergo substrate-induced catalytic inactivation with

COX-2 appearing to be the more susceptible isoform.

151

33A. His ovCOX-l His mucox-z

 

spadflc athy (nmnlIn-Inlrm lynx.)
apodflc adlvlty (mrnllninlflu In“)

 

 

   

mPGES—l + GSH

 

B. AA.»
HETE‘ _ W .
HHI"~ *‘I-“ww
PGE2”
< 5 5 5
< < I
‘ 5 5 < a a.
Pretreatment -5 g 1 I: O o '0
~
1 2 3 4 5 6

Figure 33. COX activity and prostaglandin product formation of His ovCOX-l and His
muCOX—2 cells pretreated with exogenous AA. Upon induction of His ovCOX-l or His
muCOX-Z, cultured 293 cells stably expressing these enzymes were incubated with the indicated
concentrations of AA for 10 min. Afterwards, the cells were washed with 1X PBS, harvested and
centrifuged at 2000 rpm for 3 min. Cell pellets were resuspended in 0.1 M Tris-Cl, pH 8.0 and
lysed by sonication. (A) The COX activity of the whole cell lysates was measured by an oxygen
electrode assay in a standard reaction mixture containing 100 pM AA, 1 mM phenol, and 5 pM
hematin dissolved in 0.1 M Tris-Cl, pH 8.0. (B) Microsomes were isolated from the whole cell
lysates pretreated with 0 pM (1, 4), 10 pM (2, 5) or 25 pM AA (3, 6) and prepared for radioactive
TLC as described in the text. Equivalent amounts of microsomes (0.5 mg) were incubated in a
IOO-pl reaction mixture containing 0.1 M Tris-HCI, pH 8.0, 1 mM phenol, 10 pM hematin, and
10 pM [1-“C]arachidonic acid at room temperature for 1 min. To assay for PGE2 formation
microsomal PGES-l (mPGES-l) and 60 pM GSH were included in reaction samples 4, 5 and 6.
Afier the reactions were terminated as described in the text the radioactive products were isolated,
resolved by TLC, and analyzed by autoradiography. HETE—hydroxyeicosatetraenoic acid; HHT —
hydroxyheptadecatrienoic acid.

152

Discussion

The secondary and tertiary structures of COX-2 are more susceptible to
guadinium hydrochloride chemical denaturation compared with COX-1, demonstrating a
difference in the inherent structural stabilities of the COX isoforms (150). This difference
in protein stability parallels that observed in viva where COX-l is very stable under
conditions in which COX-2 is rapidly degraded (257). We have previously attributed the
short protein half-life of COX-2 to a unique C-terminal l9-amino acid cassette (l9-aa)
that is required for ER-associated degradation of the enzyme in HEK293 cells. In the
present study we have shown that COX-2 undergoes a second distinct form of
degradation that is induced by COX fatty acid substrate and is independent of the l9-aa.
We also demonstrate that the COX substrate-induced degradation of COX-2 requires a
functional COX active site because it can be blocked by: (a) both competitive and time-
dependent, irreversible COX inhibitors, and (b) a G533A point mutation that inactivates
COX catalysis. That the substrate-induced degradation of COX-2 is dependent on
substrate binding and turnover at the COX active site is suggested by the finding that
only known COX-2 COX fatty acid substrates are able to facilitate COX-2 protein
turnover. Furthermore, a 5 min brief challenge of COX-Z-expressing cells with AA,
during which prostaglandin products are expected to be formed, is sufficient to
significantly enhance COX-2 degradation. Overall, these findings enable us to postulate
that COX substrate-dependent degradation of COX-2 is initiated at the COX active site
by the binding of fatty acid substrate. It is not clear whether the inactivity observed with
GS33A COX-2 is due to a defect in substrate binding or substrate turnover. This point

mutant has some residual COX activity with AA as substrate (~5% relative to the native

153

enzyme) (17,272) suggesting that it may bind AA but inefficiently bis-oxygenate the
substrate to form PGGz.

Proteasome activity appears to be dispensable for the COX substrate-dependent
degradation of COX-2. We have previously shown that COX-2 degradation in NIH/3T3
cells is partly proteasome-independent in NIH/3T3 cells, but largely proteasome-
dependent in HEK293 cells (257). Also noteworthy is the finding that COX inhibitors
dramatically prevented the basal degradation of COX-2 in serum-induced NIH/3T3 cells
but did not affect the proteasome-dependent degradation of the enzyme in HEK293 cells.
Collectively, these findings suggest that COX inhibitors may not inhibit the proteasome-
dependent degradation of COX-2 that is initiated by the C-terminal 27 amino acid
instability element. Therefore, there appears to be a second distinct mechanism for COX-
2 protein turnover present in NIH/3T3 cells that is triggered by the endogenous release of
AA. HEK293 cells lack detectable phospholipase A2 activity whereas NIH/3T3 cells can
activate cPLAmI in response to serum treatment (33,58,77). This could explain why COX
inhibitors do not retard the basal, proteasome-dependent degradation of COX-2 in
HEK293 cells. In NIH/3T3 cells, endogenously released AA will be utilized as COX
substrate by COX-2, thereby, initiating the substrate-dependent degradation of the
enzyme. Addition of exogenous free AA or A23187, but not bradykinin, further enhanced
the NSAID-inhibitable degradation of COX-2 in serum-treated 3T3 cells. In this regard, it
is possible that a copious amount of endogenous AA was released upon serum-
stimulation that was sufficient to induce substantial degradation of COX-2, so that adding
an additional stimulus like bradykinin did not dramatically affect COX-2 protein levels. It

would be interesting to determine if the effect of bradykinin and A23187 on COX-2

154

degradation can be amplified in 3T3 cells that have been induced to express COX-2 with
phorbol ester instead of serum. Since phorbol ester is not a major activator of cPLAza in
fibroblast cells it should induce a relatively smaller release of AA compared to serum
resulting in a proportionate attenuation of the endogenous COX-2 degradation. If indeed
endogenous COX-2 degradation is partly induced by cPLAz-mobilized AA in serum-
treated NIH/3T3 cells, it would also be reasonable to test the effect of cPLAz inhibition
on the degradation of the protein.

COX-1 and COX-2 undergo irreversible suicide inactivation during COX
catalysis under cell-free conditions (3,156,162,163). Whether or not these enzymes
undergo suicide inactivation in intact cells has not yet been clarified. Riese et al. have
provided evidence to suggest that COX-2 may inactivate in RAW264.7 macrophage cells
that have been challenged with a combination of LPS and IFN-y (273). In their study,
they observed that while COX-2 expression levels remained at maximal levels 24—48 h
after induction with LPS and IFN-y , the specific activity of the enzyme was transient
during this time period. In vitro COX inactivation is accompanied by significant changes
in the structural conformation of the enzyme, suggesting that this process is a culmination
of the destruction of the native structure of heme-protein complex (162). The concept of
inactivation-coupled structural damage due to COX catalysis becomes important in the
context of COX protein stability since it is well known that the ER has a sophisticated
quality control system that recognizes and selectively eliminates structurally-damaged
ER-associated proteins (164-166,182-186). In the present study, evidence has been
provided to show that both COX-1 and -2 can inactivate in intact HEK293 cells upon

exposure to exogenous AA. From this finding we propose that the inactivation-coupled

155

structural damage to COX-2 as a consequence of substrate turnover at the COX active
site precedes the substrate-dependent degradation of the enzyme (Fig. 34). We are yet to
determine whether this model is also applicable to COX-1. It is interesting to observe that
although the COX substrate-dependent degradation of COX-1 in HEK293 is relatively

modest compared to COX-2, it appears to be proportionate to its COX substrate-induced

 

 

 

 

 

 

 

inactivation.
HE K293 NIH/3T3
FTPII'RCFCIIH? challengej [ Serum stimulation 1
2
Induction of (0&2 LCOX-Z induction 1 ELA2 activation 1 FCPLA,

 

 

 

 

lnhlbltors??
heterologous expression 1

A Mobilization of free
. . f g endogenous AA

, Basal ( OX-Z 1

“F I ERAD ‘

x

 

 

 

 

 

 

 

 

Proteolysis by l \‘n v
1 "‘ - 1
; -65 proteasome .‘ [COX 2 (‘OX ”my,“ I |__ NSAIDS

 

 

stem—4

 

 

 

Inactivation induced
structural damage to
cyclooxygenase site

 

 

 

 

 

Proteasome independent
degradation

 

 

 

Figure 34. Model for the basal and COX substrate-initiated degradation of COX-2
in HEK293 (red) and serum-treated NIH/3T3 cells (blue). Proteasome-dependent
glycoprotein ERAD is the predominant pathway for COX-2 degradation in HEK293 cells
which essentially lack phospholipase A2 (PLA2) activity. Addition of exogenous AA to
COX-Z-expressing HEK293 cells will activate a second distinct proteasome-independent,
NSAID-inhibitable degradation pathway that involves COX catalysis. The degradation of
endogenous inducible COX—2 in serum-treated NIH/3T3 cells may occur via both
pathways. If COX-2 degradation in these cells is further enhanced by endogenous AA
release, then phospholipase A2 inhibitors should stabilize the protein to the same extent as
NSAIDS.

156

It seems unlikely that the COX substrate-dependent loss of COX-2 protein as
determined by immunoblotting is due to the formation of COX reactive oxygenated side
products that covalently modify a region(s) of the COX-2 protein that serves as an
epitope for antibody detection. In examining the effect of COX substrate on the stability
of COX-2 we utilized two different antibodies for immunoblotting. In NIH/3T3 cells a
peptide-directed antibody against the epitope SS98-K612 of muCOX-2 was used to detect
serum-induced COX-2. In HEK293 cells an antibody against Q583-N594 of huCOX-2
was used to detect tetracycline-inducible stably expressed COX-2. Both antibodies
yielded similar results even though they are directed against different epitopes on the
protein. Therefore, we believe that the substrate-dependent decrease in COX-2
immunoreactivity in both cell lines reflects the degradation of the enzyme. We were
unable to identify the degradation pathway or the protease(s) responsible for the
substrate-induced proteolytic cleavage of COX-2. We have found that the proteolytic
activities of the 26S proteasome and the lysosome, and trafficking between the ER and
Golgi appear to be dispensable for this process. It is likely that the protease(s) involved in
substrate-induced cleavage of COX-2 may be resident in the ER. It may not have been
possible to identify the ER- or microsome-associated proteolytic activity because the
microsomes that were used in our studies were not intact; trypsin cleavage analysis
suggested that the microsomes were likely to be wrong-side-out.

Very few enzymes have been hitherto identified as ER-resident proteases. Signal
peptidase, which is known to be physically associated with the Sec6l translocon, cleaves
N-terminal ER targeting sequences of proteins as they are imported into the ER (274-

276). Signal peptide peptidase (SPP) is an ER membrane-associated aspartyl protease that

157

cleaves the signal peptide fragments generated by signal peptidase (277). Recently, SPP
has been shown to be essential for the retrotranslocation and subsequent proteasomal
degradation of major histocompatibility complex (MHC) Class I heavy chains that is
induced by the US2 protein of cytomegalovirus (278). However, this study did not clarify
whether the protease activity of SPP per se was required for this process. ER
aminopeptidases-l and -2 (ERAP-l and -2) are ER luminal, interferon-y—inducible
metallopeptidases that are believed to be involved in the final proteolytic processing of
short proteasome-generated peptide fragments that will be loaded onto MHC Class I
molecules for antigen presentation on the surface of an antigen presenting cell (279-282).
There are no reports thus far implicating ERAP-l and -2 in the degradation of whole
proteins. The membrane remodeling enzyme A9 steroyl-CoA desaturase is an ER integral
membrane protein that is rapidly degraded (tug ~3-4 h) under basal conditions in a
proteasome-independent manner (283,284). Heinemann et al have purified and
characterized a 90 kDa plasminogen-related protease with Arg/Lys specificity that it is
responsible for the rapid degradation of steroyl-CoA desaturase (284,285). This protease
is unlikely to act in the ER lumen because the sequence determinant of steroyl-CoA
desaturase that directs its rapid turnover is found on the cytosolic N-terminal portion of
the enzyme (286). Although the mechanism for steroyl-CoA proteolysis is still unclear,
there is evidence to show that the protein gets cleaved within a putative membrane
binding region, most likely by the ER plasminogen-related protease (283). It has been
suggested that the N-terminal instability sequence of steroyl-CoA desaturase may help to
bring the membrane protease to close vicinity with the enzyme (286). Recently, Donoso

et al. have shown that the degradation of a misfolded mutant form of BiP is initiated in

158

the ER lumen by a serine protease that is yet to be identified (287). In this regard, it is
likely that a proteolytic system exists in the ER that may serve as an alternative or a
complement to the ERAD-proteasome pathway in the quality control of structurally
defective ER-associated proteins. Therefore, we speculate that the catalytically
inactivated form(s) of COX-2 is degraded by this putative ER protease system.

In conclusion, we have found that AA, the principal PUFA COX-2 substrate,
induces both the COX catalytic inactivation and the rapid degradation of the enzyme. We
have also observed that inhibition of COX-2 COX catalysis using site-directed
mutagenesis and pharmacological approaches prevent AA-dependent degradation. Based
on these observations we have proposed that the AA-dependent degradation of COX-2 is
a consequence of structural damage to the COX active site. In the future, it will be
important to measure and compare the kinetic constants for the catalytic inactivation of
the COX isoforms in intact cells. The proposition that COX-2 degradation in NIH/3T3
cells is enhanced by serum-induced mobilization of free AA will need to be tested further
using inhibitors of cPLAz. A protocol for preparing intact microsomes will also be
designed in an effort to identify the ER-associated protease(s) that degrades catalytically
inactive COX-2. Indeed, the unconventional COX substrate-dependent quality control of
COX-2 may serve as an additional physiological mechanism for the metabolic regulation

of prostanoid synthesis.

159

CONCLUSION

The structural basis for the selection of normal short-lived proteins for rapid
proteolysis in an intracellular environment where they coexist with many stable proteins
is not clearly defined. We have identified a 27 amino acid instability element close to the
C-terminal end of COX-2 that regulates the proteasomal degradation of this ER-resident
membrane-bound protein by mediating its entry into the ERAD system. We believe that
COX-2 is the only known ER luminal integral membrane protein that is degraded via
ERAD from its native or properly folded form. The ERAD pathway is an important
protein quality control pathway that is not well characterized in mammalian cells. ERAD
components in the ER lumen that participate in the initial selective recognition of an
ERAD substrate and the mechanism by which they do so are not clearly defined. The
identity of the ER membrane retrotranslocon is also yet to be determined. Therefore,
since COX-2 is one of the very few mammalian proteins that are known to undergo
proteasomal degradation from the ER it could serve as a suitable reporter substrate for
studying ERAD in a mammalian system.

We have also identified a second distinct mechanism for COX-2 degradation that
is preceded by the COX substrate (AA)-induced suicide inactivation of the enzyme.
Intracellular COX-1 appears to be less susceptible to substrate-induced suicide
inactivation and degradation. COX-l is a stable protein that is constitutively expressed in
many cells. It is also able to utilize endogenous AA at higher concentrations than COX-2.
Therefore, based on our experimental data we reason that the housekeeping function of
COX-1 might extend to include regulating the intracellular levels of free AA. The

proteolysis of COX-2 in response to two distinct signals, namely a structural and a

160

metabolic signal, reflects the tight and elaborate control of COX-2 expression at all three
levels of gene regulation. Since the differential expression profiles of COX-1 and COX-2
are thought to contribute to their specialized biological functions, our observations may

provide a clue to the unsolved mystery of why mammals have two COX isoforms.

161

APPENDIX

Table 3. C-terminal mutants designed by QuickChangeTM site-directed mutagenesis

 

 

 

 

 

 

 

Mutation Primers Template
G533 A Forward: 5 ’CT CCTT GAAAGCACTT A huCQX-2
hUCOX-2 TGGGTAATG 3’

Reverse: 5’CATTACCCATAAGTGCTTT

CAAGGAG 3’
N594 A Forward: 5’ GTCACCATCGCAGCAAGTT huCOX-2
huCOX-2 CTTCCCGC 3’

Reverse: 5’ GCGGGAAGAACTTGCTG

CGATGGTGAC 3’
P607 A T608 A Forward: 5’ CTA GAT GAT ATC AAT GCA huCOX_2
huCOX-2 GCA GTA CT A CTA AAA GAA 3’

Reverse: 5’ 'I'I’C 'I'I'I‘ TAG TAG TAC TGC

TGC ATT GAT ATC ATC TAG 3’
V609 A Forward: 5’ GAT GAT ATC AAT CCC ACA huCQX-2
huCOX-2 GCA CT A CTA AAA GAA CGT 3’

Reverse: 5’ ACG TTC T'I'I‘ TAG TAG TGC

TGT GGG A'I'I’ GAT ATC ATC 3’
L610 A Forward: 5’ GAT ATC AAT CCC ACA GTA huCOX-2
hUCOX-2 GCA CT A AAA GAA CGT TCG 3’

 

Reverse: 5’ CGA ACG TTC 'I'I'I‘ TAG TGC
TAC TGT GGG ATT GAT ATC 3’

 

 

162

 

Table 3 (cont’d)

 

 

 

 

 

 

 

 

Reverse: 5"I‘TCAGTCGAACGTTCA
TTGATATCATC 3’

 

L61 1A Forward: 5’ ATC AAT CCC ACA GTA CT A huCOX-2
hUCOX-2 GCA AAA GAA CGT TCG ACT 3’
Reverse: 5’ AGT CGA ACG TTC TTT TGC
TAG TAC TGT GGG ATT GAT 3’ .
K6 1 2A Forward: 5 ’ CCCACAGTACTACT A huCOX-2
hUCOX-2 GCAGAACGTTCGACT 3'
Reverse: 5’ AGTCGAACGTTCTGCTAGT
AGTACTGTGGG 3’
dd 5 9 5-6 1 2 Forward: 5 ’ AC AGTCACCATCAAT huCQX-2
huCOX-2 GAACGTTCGAC 3’
Reverse: 5’ CT ACAG'I'I’CAGTC
GAACGTTCATTG 3’
(131597-61 2 Forward: 5 ’ ACCATCAATGCA huCOX-2
huCOX-2 AGTGAACGTTCGACT 3’
Reverse: 5’ AGTCGAACGTTCACTTGCATT
GATGGTGAC 3’
de1602-61 2 Forward: 5 ’CTTCCCGCTCCGGAG huCQX-2
hUCOX-2 AACGTTCGACTGAA 3’
Reverse: 5’ TTCAGTCGAACG'I'I’CT C
CGGAGCGGGAAG 3’
(131607 -6 1 2 Forward: 5 ’GATGATATCAAT huCOX_2
hUCOX-2 GAACGTTCGACTGAA 3’

 

163

 

Table 3 (cont’d)

 

 

 

 

 

 

 

MuCOX-2 Forward: 5’ AGT GT'I’ CCA GAT CCA CAG huCQX-2
UPSTRM8 CCT ACC AAA ACA GCC ACC ATC AAT
huCOX-2 GCA AGT 3’
Reverse: 5’ ACT TGC ATT GAT GGT GGC
TGT TTT GGT AGG CTG TGG ATC TGG
AAC ACT 3’
ovCOX- 1 Forward: 5' AGT GTT CCA GAT CCA CGT huCOX-2
UPSTRM8 CAG GAG GAC AGG CCT GGG GTG AAT
h COX 2 GCA AGT TCT TCC 3’
u ' Reverse: 5’ GGA AGA ACT TGC ATT CAC
CCC AGG CCT GTC CT C CT G ACG TGG
ATC TGG AAC ACT 3’
V591P T592G Forward: 5’ GAG CTC ATT AAA ACA CCC huCOX-2
huCOX-2 GGC ATC AAT GCA AGT TCT 3’
Reverse: 5’ AGA ACT TGC ATT GAT GCC
GGG TGT TTT AAT GAG CTC 3’
insS94-596 Forward: 5’ GACAGGCCTGGGGTGGAGA ovCOX-l
ovCOX-l ATGCAAGTCGGCCACCCACA 3'
Reverse: 5’ TGTGGGTGGCCGACTTGCA
TTCTCCACCCCAGGCCTGTC 3’
{113594-601 Forward: 5’ TCCCGCTCCGGAGAGCGG in3594-612
ovCOX-l CCACCC 3’ ovCOX-l
Reverse: 5’ GGGTGGCCGCTCTCCGGAGC
GGGA 3’
in5594-612 Forward: 5 ’ GACAGGCCTGGGGTGGCAGC in5594-612
(N594A) AAGTTCTTCCCGC 3’ OVCOX-l
ovCOX-l

 

Reverse: 5’ GCGGGAAGAACTTGCTGC
CACCCCAGGCCTGTC 3’

 

 

164

 

Table 4. C-terminal mutants desimed by overlap-extension PCR mutagenesis

 

 

Mutation Primers Template
ins594-612 Mutation Primers ovCOX-l
OVCOX-l Forward: 5 ’ GACAGGCCTGGGGTGAATGC

AAG'I'I’CTTCCCGCTCCGGACTAGATGAT
ATCAATCCC ACAGTACT ACT AAAAGAG
CGGCCACCCACA 3’

Reverse: 5’ TGTGGGTGGCCGCT CTTTT AG
TAGTACTGTGGGATTGATATCATCI’AGT
CCGGAGCGGGAAGAACTTGCA’I‘TCACC
CC AGGCCTGTC 3 ’

Common Primers
Forward: 5’ GCG'ITTAAAC’I'I‘AAGCTT 3’

Reverse: 5’ GAGCTCGGTACCAAGCTT 3’

 

Sin5594-612 Mutation primers huCOX-2
huCOX-2 Forward: 5’ ACAGTCACCATCGCT GCT A

GTATTAGAGGTCCTTT AACT AAACTT AG
TGATAATAGTCTTG’I‘TAGTGATGAACGT
TCGACTGAACT G 3’

Reverse: 5’ CAGTTCAGTCGAACGTTCATC
ACTAACAAGACTATTATCACTAAGTTI’A
GTTAAAGGACCTCTAATACTAGCAGCG
ATGGTGACT GT 3’

Common primers
Forward: 5’ GC'I‘TGGTACCGAGCTCGGA
TCC 3’

Reverse: 5’ CGGGCCCTCTAGACTCGAG
CGGCC 3’

 

 

 

 

 

165

Table 4 (cont’d)

 

Sins597-6 12
huCOX-2

 

 

Mutation primers

Forward: 5’ GTCACCATCAA'I‘GCAAGTAT

TAGAGGTCCTTI‘AACTAAACTTAGTGAT
AATAGTC’I'FGTTAGTGATGAACGTTCGA
CTGAACTG 3’

Reverse: 5’ CAGTTCAGTCGAACGTTCAT
C ACTAAC AAGACTATTATC ACT AAGT'I'I’
AGT'I’AAAGGACCTCTAATACTTGCATTG
ATGGTGAC 3’

Common primers
Forward: 5’ GCTTGGTACCGAGCTCGGA
TCC 3’

Reverse: 5’ CGGGCCCTCTAGACTCGAG
CGGCC 3’

huCOX-2

 

 

A. Conditions for QuickChangeTM site-directed mutagenesis

 

 

 

 

 

Cycles Temperature (°C) Time (min)
1 94 3
20 94 l
52 l
68 20
1 72 7

 

 

166

 

B. Site-directed mutagenesis by overlap-extension PCR
1) Mutation step

a) PCR reaction set-up:

50 ng DNA template

5 11.1 10X Vent DNA polymerase reaction buffer

- 25 picomoles each of forward common primer* and reverse mutation
primer (Reaction A), or 25 picomoles each of reverse common primer*
and forward mutation primer (Reaction B)

- 1.25 m of40deNTPmix

- 1 ul Vent DNA polymerase (2.5 U/ul)

Add ddeO to final volume of 50 pd

* The common primers carry the restriction sites

 

 

 

 

b) PCR conditions:
Cycles Temperature (°C) Time (min)
1 94 3
3O 94 l
52 l
72 5
l 72 5

 

 

 

167

- After the PCR reaction is completed, treat with Reactions A and B with l u] DpnI
for at least 37°C for at least 4 h

- Purify the PCR fragments by gel extraction and measure DNA concentration

2) Extension step
a) Reaction set-up:
- Equal amounts of the purified PCR fragments from Reactions A and B
- 5 pl Vent DNA 10X polymerase reaction buffer
-1.25 11.140 mM dNTP mix
- 1 11.1 Vent DNA polymerase

- Bring total volume to 50 111 with ddeO

 

 

 

b) PCR conditions
Cycles Temperature (°C) Time (min)
1 94 3
6 94 I
52 1
72 5

 

 

 

168

3) Amplification step
a) Reaction set-up:
To the reaction obtained from the extension step, add:
- 25 picomoles of each of the common primers

- 1 ul Vent DNA polymerase

 

 

 

 

b) PCR conditions
Cycles Temperature (°C) Time (min)
1 94 3
30 94 1
52 l
72 5
l 72 5

 

 

 

- Purify the PCR product using a PCR purification kit
- Perform a restriction digestion with the appropriate enzymes and li gate

onto vector of interest.

169 '

10.

ll.

12.

13.

BIBLIOGRAPHY

Smith, W. L., DeWitt, D. L., and Garavito, R. M. (2000) Annu Rev Biochem 69,
145—182

Narumiya, S., and FitzGerald, G. A. (2001) J Clin Invest 108(1), 25-30

Rouzer, C. A., and Mamett, L. J. (2003) Chem Rev 103(6), 2239-2304

Wang, D., Mann, J. R., and DuBois, R. N. (2005) Gastroenterology 128(5), 1445-
1461

Grosser, T., Fries, S., and FitzGerald, G. A. (2006) J Clin Invest 116(1), 4-15

Honda, T., Segi-Nishida, E., Miyachi, Y., and Narumiya, S. (2006) J Exp Med
203(2), 325-335

van der Donk, W. A., Tsai, A. L., and Kulmacz, R. J. (2002) Biochemistry 41(52),
15451-15458

Leslie, C. C. (2004) Biochem Cell Biol 82(1), l-l7

Murakami, M., Kambe-Ohkura, T., and Kudo, I. (2002) Adv Exp Med Biol 507,
15-19

Kudo, I., and Murakami, M. (2002) Prostaglandins Other Lipid Mediat 68-69, 3-
58

Picot, D., Loll, P. J ., and Garavito, R. M. (1994) Nature 367(6460), 243-249

Kurumbail, R. G., Stevens, A. M., Gierse, J. K., McDonald, J. J., Stegeman, R.
A., Pak, J. Y., Gildehaus, D., Miyashiro, J. M., Penning, T. D., Seibert, K.,
Isakson, P. C., and Stallings, W. C. (1996) Nature 384(6610), 644-648

Otto, J. C., and Smith, W. L. (1994) J Biol Chem 269(31), 19868—19875

170

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

Spencer, A. G., Woods, J. W., Arakawa, T., Singer, II, and Smith, W. L. (1998) J
Biol Chem 273(16), 9886-9893

MirAfzali, Z., Leipprandt, J. R., McCracken, J. L., and DeWitt, D. L. (2006) J
Biol Chem 281(38), 28354-28364

Spencer, A. G., Thuresson, E., Otto, J. C., Song, 1., Smith, T., DeWitt, D. L.,
Garavito, R. M., and Smith, W. L. (1999) J Biol Chem 274(46), 32936-32942

Yuan, C., Rieke, C. J ., Rimon, G., Wingerd, B. A., and Smith, W. L. (2006) Proc
Natl Acad Sci U S A 103(16), 6142-6147

Munro, 8., and Pelham, H. R. (1987) Cell 48(5), 899-907

Song, I., and Smith, W. L. (1996) Arch Biochem Biophys 334(1), 67-72

Tanabe, T., and Tohnai, N. (2002) Prostaglandins Other Lipid Mediat 68-69, 95-
l 14

Perkins, D. J ., and Kniss, D. A. (1997) Biochem J 321 ( Pt 3), 677-681

DeWitt, D. L., and Meade, E. A. (1993) Arch Biochem Biophys 306(1), 94-102

Rich, G., Yoder, E. J ., and Moore, S. A. (1998) J Cell Physiol 176(3), 495-505

Zaric, J ., and Ruegg, C. (2005) J Biol Chem 280(2), 1077-1085

Funk, C. D. (2001) Science 294(5548), 1871-1875

Zhang, Y., Shaffer, A., Portanova, J ., Seibert, K., and Isakson, P. C. (1997) J
Pharmacol Exp Ther 283(3), 1069-1075

Samad, T. A., Moore, K. A., Sapirstein, A., Billet, S., Allchome, A., Poole, S.,
Bonventre, J. V., and Woolf, C. J. (2001) Nature 410(6827), 471-475

171

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

Cao, C., Matsumura, K., Yamagata, K., and Watanabe, Y. (1998) Brain Res Mol
Brain Res 56(1-2), 45-56

Leahy, K. M., Omberg, R. L., Wang, Y., Zweifel, B. S., Koki, A. T., and
Masferrer, J. L. (2002) Cancer Res 62(3), 625-631

Sinicrope, F. A. (2006) Mol Carcinog 45(6), 447-454

Yedgar, S., Cohen, Y., and Shoseyov, D. (2006) Biochim Biophys Acta 1761(11),
1373-1382

Ghosh, M., Tucker, D. E., Burchett, S. A., and Leslie, C. C. (2006) Prog Lipid
Res 45(6), 487-510

Mounier, C. M., Ghomashchi, F., Lindsay, M. R., James, 8., Singer, A. G.,
Parton, R. G., and Gelb, M. H. (2004) J Biol Chem 279(24), 25024-25038

Ni, Z., Okeley, N. M., Smart, B. P., and Gelb, M. H. (2006) J Biol Chem 281(24),
16245-16255

Leslie, C. C. (1997) J Biol Chem 272(27), 16709-16712

Lin, L. L., Lin, A. Y., and DeWitt, D. L. (1992) J Biol Chem 267(33), 23451-
23454

Hoeck, W. G., Ramesha, C. 8., Chang, D. J ., Fan, N., and Heller, R. A. (1993)
Proc Natl Acad Sci U S A 90(10), 4475-4479

Maxwell, A. P., Goldberg, H. J ., Tay, A. H., Li, Z. G., Arbus, G. S., and Skorecki,
K. L. (1993) Biochem J 295 ( Pt 3), 763-766

Wu, T., Levine, S. J ., Lawrence, M. G., Logun, C., Angus, C. W., and Shelhamer,
J. H. (1994) J Clin Invest 93(2), 571-577

Murakami, M., Matsumoto, R., Urade, Y., Austen, K. F., and Ann, J. P. (1995) J
Biol Chem 270(7), 3239-3246

172

41.

42.

43.

45.

46.

47.

48.

49.

50.

51.

52.

53.

Miyakawa, T., Kojima, M., and Ui, M. (1998) Biochem J 329 (Pt 1), 107-114

Clapham, D. E. (1995) Nature 375(6533), 634-635

Bootman, M. D., and Berridge, M. J. (1995) Cell 83(5), 675-678

Reynolds, L. J ., Hughes, L. L., Louis, A. I., Kramer, R. M., and Dennis, E. A.
(1993) Biochim Biophys Acta 1167(3), 272-280

Loo, R. W., Conde-Frieboes, K., Reynolds, L. J ., and Dennis, E. A. (1997) J Biol
Chem 272(31), 19214-19219

Clark, J. D., Lin, L. L., Kriz, R. W., Ramesha, C. S., Sultzman, L. A., Lin, A. Y.,
Milona, N ., and Knopf, J. L. (1991) Cell 65(6), 1043-1051

Das, S., Rafter, J. D., Kim, K. P., Gygi, S. P., and Cho, W. (2003) J Biol Chem
278(42), 41431-41442

Bonventre, J. V., Huang, Z., Taheri, M. R., O'Leary, E., Li, B, Moskowitz, M. A.,
and Sapirstein, A. (1997) Nature 390(6660), 622-625

Uozumi, N., Kume, K., Nagase, T ., Nakatani, N., Ishii, S., Tashiro, F., Komagata,
Y., Maki, K., Ikuta, K., Ouchi, Y., Miyazaki, J ., and Shimizu, T. (1997) Nature
390(6660), 618-622

Hirabayashi, T., Murayama, T., and Shimizu, T. (2004) Biol Pharm Bull 27(8),
1 168-1 173

Hegen, M., Sun, L., Uozumi, N., Kume, K., Goad, M. E., Nickerson-Nutter, C. L.,
Shimizu, T., and Clark, J. D. (2003) J Exp Med 197(10), 1297-1302

Takaku, K., Sonoshita, M., Sasaki, N., Uozumi, N., Doi, Y., Shimizu, T., and
Taketo, M. M. (2000) J Biol Chem 275(44), 34013-34016

Hong, K. H., Bonventre, J. C., O'Leary, E., Bonventre, J. V., and Lander, E. S.
(2001) Proc Natl Acad Sci U S A 98(7), 3935-3939

173

54.

55.

56.

57.

58.

59.

60.

61.

62.

63.

65.

66.

Nagase, T., Uozumi, N., Ishii, S., Kita, Y., Yamamoto, H., Ohga, E., Ouchi, Y.,
and Shimizu, T. (2002) Nat Med 8(5), 480—484

Nagase, T., Uozumi, N., Ishii, S., Kume, K., Izumi, T., Ouchi, Y., and Shimizu, T.
(2000) Nat Immunol 1(1), 42-46

Kalyvas, A., and David, S. (2004) Neuron 41(3), 323-335

Oikawa, Y., Yamato, E., Tashiro, F., Yamamoto, M., Uozumi, N., Shimada, A.,
Shimizu, T., and Miyazaki, J. (2005) FEBS Lett 579(18), 3975-3978

Murakami, M., Shimbara, S., Kambe, T., Kuwata, H., Winstead, M. V.,
Tischfield, J. A., and Kudo, I. (1998) J Biol Chem 273(23), 14411-14423

Murakami, M., Koduri, R. S., Enomoto, A., Shimbara, S., Seki, M., Yoshihara,
K., Singer, A., Valentin, E., Ghomashchi, F., Lambeau, G., Gelb, M. H., and
Kudo, I. (2001) J Biol Chem 276(13), 10083-10096

Satake, Y., Diaz, B. L., Balestrieri, B., Lam, B. K., Kanaoka, Y., Grusby, M. J .,
and Arm, J. P. (2004) J Biol Chem 279(16), 16488-16494

Murakami, M., and Kudo, I. (2004) Biol Pharm Bull 27(8), 1158-1164

Alaoui-El-Azher, M., Wu, Y., Havet, N ., Israel, A., Lilienbaum, A., and Touqui,
L. (2002) Mol Pharmacol 61(4), 786-794

Nakano, T., and Arita, H. (1990) FEBS Lett 273(1-2), 23-26

Crowl, R. M., Stoller, T. J ., Conroy, R. R., and Stoner, C. R. (1991) J Biol Chem
266(4), 2647-2651

Pfeilschifter, J ., Schalkwijk, C., Briner, V. A., and van den Bosch, H. (1993) J
Clin Invest 92(5), 2516-2523

Kuwata, H., Sawada, H., Murakami, M., and Kudo, I. (1999) Adv Exp Med Biol
469, 183-188

174

67.

68.

69.

70.

71.

72.

73.

74.

75.

76.

77.

78.

79.

Pruzanski, W., Vadas, P., Stefanski, E., and Urowitz, M. B. (1985) J Rheumatol
12(2), 211-216

Pruzanski, W., Keystone, E. C., Stemby, B., Bombardier, C., Snow, K. M., and
Vadas, P. (1988) J Rheumatol 15(9), 1351-1355

Vadas, P. (1984) J Lab Clin Med 104(6), 873-881

Green, J. A., Smith, G. M., Buchta, R., Lee, R., Ho, K. Y., Rajkovic, I. A., and
Scott, K. F. (1991) Inflammation 15(5), 355-367

Vadas, P., and Pruzanski, W. (1993) Circ Shock 39(2), 160-167

Bidgood, M. J ., Jamal, O. S., Cunningham, A. M., Brooks, P. M., and Scott, K. F.
(2000) J Immunol 165(5), 2790-2797

Gijon, M. A., Spencer, D. M., Siddiqi, A. R., Bonventre, J. V., and Leslie, C. C.
(2000) J Biol Chem 275(26), 20146-20156

Balsinde, J ., Balboa, M. A., and Dennis, E. A. (1998) Proc Natl Acad Sci U S A
95(14), 7951-7956

Reddy, S. T., and Herschman, H. R. (1997) J Biol Chem 272(6), 3231-3237

Murakami, M., Kambe, T., Shimbara, S., and Kudo, I. (1999) J Biol Chem 274(5),
3103-31 15

Reddy, S. T., and Herschman, H. R. (1994) J Biol Chem 269(22), 15473-15480

Murakami, M., Kuwata, H., Amakasu, Y., Shimbara, S., Nakatani, Y., Atsumi,
G., and Kudo, I. (1997) J Biol Chem 272(32), 19891-19897

Roshak, A., Sathe, G., and Marshall, L. A. (1994) J Biol Chem 269(42), 25999-
26005

175

80.

81.

82.

83.

84.

85.

86.

87.

88.

89.

90.

Langenbach, R., Morham, S. G., Tiano, H. F., Loftin, C. D., Ghanayem, B. I.,
Chulada, P. C., Mahler, J. F., Lee, C. A., Goulding, E. H., Kluckman, K. D., Kim,
H. S., and Smithies, O. (1995) Cell 83(3), 483-492

Kobayashi, T., and Narumiya, S. (2002) Prostaglandins Other Lipid Mediat 68-
69, 557-573

Naraba, H., Murakami, M., Matsumoto, H., Shimbara, S., Ueno, A., Kudo, I., and
Oh-ishi, S. (1998) J Immunol 160(6), 2974-2982

Moncada, S., and Vane, J. R. (1979) Fed Proc 38(1), 66-71

Liou, J. Y., Shyue, S. K., T sai, M. J ., Chung, C. L., Chu, K. Y., and Wu, K. K.
(2000) J Biol Chem 275(20), 15314-15320

Bustos, M., Coffman, T. M., Saadi, S., and Platt, J. L. (1997) J Clin Invest 100(5),
1150-1158

Okahara, K., Sun, B., and Kambayashi, J. (1998) Arterioscler Thromb Vasc Biol
18(12), 1922-1926

Caughey, G. E., Cleland, L. G., Penglis, P. S., Gamble, J. R. and James, M. J.
(2001) J Immunol 167(5), 2831-2838

Solomon, S. D., McMurray, J. J ., Pfeffer, M. A., Wittes, J ., Fowler, R., Finn, R,
Anderson, W. F., Zauber, A., Hawk, E., and Bertagnolli, M. (2005) N Engl J Med
352(11), 1071-1080

Bresalier, R. S., Sandler, R. S., Quan, H., Bolognese, J. A., Oxenius, B., Horgan,
K., Lines, C., Riddell, R., Morton, D., Lanas, A., Konstam, M. A., and Baron, J.
A. (2005) N Engl J Med 352(11), 1092-1102

DeWitt, D. L., Day, J. S., Sonnenburg, W. K., and Smith, W. L. (1983) J Clin
Invest 72(6), 1882-1888

Schildknecht, S., Bachschmid, M., Baumann, A., and Ullrich, V. (2004) FASEB J
18(6), 757-759

176

92.

93.

94.

95.

96.

97.

98.

99.

100.

101.

102.

103.

1 ()4.

Brock, T. G., McNish, R. W., and Peters-Golden, M. (1999) J Biol Chem 274(17),
1 1660-1 1666

Urade, Y., and Hayaishi, O. (1999) Biochim Biophys Acta 1436(3), 606-615

Eguchi, N., Minarni, T., Shirafuji, N., Kanaoka, Y., Tanaka, T., Nagata, A.,
Yoshida, N., Urade, Y., Ito, S., and Hayaishi, O. (1999) Proc Natl Acad Sci U S A
96(2), 726-730

Heavey, D. J., Ernst, P. B., Stevens, R. L., Befus, A. D., Bienenstock, J ., and
Austen, K. F. (1988) J Immunol 140(6), 1953-1957

Diaz, B. L., Satake, Y., Kikawada, E., Balestrieri, B., and Ann, J. P. (2006)
Biochim Biophys Acta 1761(12), 1489-1497

Kabashima, K., and Narumiya, S. (2003) Prostaglandins Leukot Essent Fatty
Acids 69(2-3), 187-194

Lehmann, J. M., Moore, L. B., Smith-Oliver, T. A., Wilkison, W. 0., Willson, T.
M., and Kliewer, S. A. (1995) J Biol Chem 270(22), 12953-12956

Forman, B. M., Tontonoz, P., Chen, J ., Brun, R. P., Spiegelman, B. M., and
Evans, R. M. (1995) Cell 83(5), 803-812

Scher, J. U., and Pillinger, M. H. (2005) Clin Immunol 114(2), 100-109

Ide, T., Egan, K., Bell-Parikh, L. C., and FitzGerald, G. A. (2003) Thromb Res
110(5-6), 311-315

Chinetti, G., Griglio, S., Antonucci, M., Torra, I. P., Delerive, P., Majd, Z.,
Fruchart, J. C., Chapman, J ., Najib, J ., and Staels, B. (1998) J Biol Chem 273(40),
25573-25580

Azuma, Y., Shinohara, M., Wang, P. L., and Ohura, K. (2001) Biochem Biophys
Res Commun 283(2), 344-346

Urade, Y., and Eguchi, N. (2002) Prostaglandins Other Lipid Mediat 68-69, 375-
382

177

105.

106.

107.

108.

109.

110.

111.

112.

113.

114.

115.

Grill, M., Peskar, B. A., Schuligoi, R., and Amann, R. (2006)
Neuropharmacology 50(2), 165-173

Murakami, M., Matsumoto, R., Austen, K. F., and Ann, J. P. (1994) J Biol Chem
269(35), 22269-22275

Harada, Y., Hatanaka, K., Kawamura, M., Saito, M., Ogino, M., Majima, M.,
Ohno, T., Ogino, K., Yamamoto, K., Taketani, Y., Yamamoto, S., and Katori, M.
(1996) Prostaglandins 51(1), 19-33

Li, S., Wang, Y., Matsumura, K., Ballou, L. R., Morham, S. G., and Blatteis, C.
M. (1999) Brain Res 825(1-2), 86-94

Ueno, N., Takegoshi, Y., Kamei, D., Kudo, I., and Murakami, M. (2005) Biochem
Biophys Res Commun 338(1), 70-76

Tanioka, T., Nakatani, Y., Semmyo, N ., Murakami, M., and Kudo, I. (2000) J
Biol Chem 275(42), 32775-32782

Murakami, M., and Kudo, I. (2004) Prog Lipid Res 43(1), 3-35

Kamei, D., Yamakawa, K., Takegoshi, Y., Mikami-Nakanishi, M., Nakatani, Y.,
Oh-Ishi, S., Yasui, H., Azuma, Y., I-Iirasawa, N., Ohuchi, K., Kawaguchi, H.,
Ishikawa, Y., Ishii, T., Uematsu, S., Akira, S., Murakami, M., and Kudo, I. (2004)
J Biol Chem 279(32), 33684-33695

Myers, L. K., Kang, A. H., Postlethwaite, A. E., Rosloniec, E. F., Morham, S. G.,
Shlopov, B. V., Goorha, S., and Ballou, L. R. (2000) Arthritis Rheum 43(12),
2687-2693

Trebino, C. E., Stock, J. L., Gibbons, C. P., Naiman, B. M., Wachtrnann, T. S.,
Umland, J. P., Pandher, K., Lapointe, J. M., Saha, S., Roach, M. L., Carter, D.,
Thomas, N. A., Durtschi, B. A., McNeish, J. D., Hambor, J. E., Jakobsson, P. J .,
Carty, T. J ., Perez, J. R., and Audoly, L. P. (2003) Proc Natl Acad Sci U S A
100(15), 9044-9049

Engblom, D., Saha, S., Engstrom, L., Westman, M., Audoly, L. P., J akobsson, P.
J ., and Blomqvist, A. (2003) Nat Neurosci 6(11), 1137-1138

178

116.
117.

118.

119.
120.
121.
122.
123.
124.
125.
126.

127.

128.

Wang, L. H., Hajibeigi, A., Xu, X. M., Loose-Mitchell, D., and Wu, K. K. (1993)
Biochem Biophys Res Commun 190(2), 406-411

Xu, X. M., Tang, J. L., Chen, X., Wang, L. H., and Wu, K. K. (1997) J Biol Chem
272(11), 6943-6950

Duquette, M., and Laneuville, O. (2002) J Biol Chem 277(47), 44631-44637

DeLong, C. J ., and Smith, W. L. (2005) Biochem Biophys Res Commun 338(1),
53-61

Topper, J. N ., Cai, J ., Falb, D., and Gimbrone, M. A., Jr. (1996) Proc Natl Acad
Sci U SA 93(19), 10417-10422

Xie, W. L., Chipman, J. G., Robertson, D. L., Erikson, R. L., and Simmons, D. L.
(1991) Proc Natl Acad Sci U S A 88(7), 2692-2696

Kujubu, D. A., Reddy, S. T., Fletcher, B. S., and Herschman, H. R. (1993) J Biol
Chem 268(8), 5425-5430

Fletcher, B. S., Kujubu, D. A., Penin, D. M., and Herschman, H. R. (1992) J Biol
Chem 267(7), 4338-4344

Herschman, H. R., Fletcher, B. S., and Kujubu, D. A. (1993) J Lipid Mediat 6(1-
3), 89-99

Asano, K., Lilly, C. M., and Drazen, J. M. (1996) Am J Physiol 271(1 Pt 1),
L126-131

Kaufmann, W. E., Worley, P. F., Taylor, C. V., Bremer, M., and Isakson, P. C.
(1997) Brain Dev 19(1), 25-34

Harris, R. C., and Breyer, M. D. (2001) Am J Physiol Renal Physiol 281(1), Fl-ll

Bagga, D., Wang, L., Farias-Eisner, R., Glaspy, J. A., and Reddy, S. T. (2003)
Proc Natl Acad Sci U S A 100(4), 1751-1756

179

129.

130.

131.

132.

133.

134.

135.

136.

137.

138.

139.

140.

141.

Chen, C. Y., and Shyu, A. B. (1994) Mol Cell Biol 14(12), 8471-8482

Kang, Y. J ., Wingerd, B. A., Arakawa, T., and Smith, W. L. (2006) J Immunol
177(11), 8111-8122

Kang, Y. J ., Mbonye, U. R., Delong, C. J., Wada, M., and Smith, W. L. (2007)
Prog Lipid Res

Hwang, D., Jang, B. C., Yu, G., and Boudreau, M. ( 1997) Biochem Pharmacol
54(1), 87-96

Mestre, J. R., Mackrell, P. J ., Rivadeneira, D. E., Stapleton, P. P., Tanabe, T., and
Daly, J. M. (2001) J Biol Chem 276(6), 3977-3982

Scheinman, R. I., Cogswell, P. C., Lofquist, A. K., and Baldwin, A. S., Jr. (1995)
Science 270(5234), 283-286

Auphan, N., DiDonato, J. A., Rosette, C., Helmberg, A., and Karin, M. (1995)
Science 270(5234), 286-290

Farrajota, K., Cheng, S., Martel-Pelletier, J ., Afif, H., Pelletier, J. R, Li, X.,
Ranger, P., and Fahmi, H. (2005) Arthritis Rheum 52(1), 94-104

Kapoor, M., Kojima, F., Qian, M., Yang, L., and Crofford, L. J. (2006) J Biol
Chem

Subbaramaiah, K., Lin, D. T., Hart, J. C., and Dannenberg, A. J. (2001) J Biol
Chem 276(15), 12440-12448

Chen, C. Y., and Shyu, A. B. (1995) Trends Biochem Sci 20(11), 465-470

Chen, C. Y., Gherzi, R., Ong, S. E., Chan, E. L., Raijmakers, R., Pruijn, G. J .,
Stoecklin, G., Moroni, C., Mann, M., and Karin, M. (2001) Cell 107(4), 451-464

Mukherjee, D., Gao, M., O'Connor, J. P., Raijmakers, R., Pruijn, G., Lutz, C. S.,
and Wilusz, J. (2002) EMBO J 21(1-2), 165-174

180

142.

143.

144.

145.

146.

147.

148.

149.

150.

151.

152.

153.

154.

Stoecklin, G., Mayo, T., and Anderson, P. (2006) EMBO Rep 7(1), 72-77

Lasa, M., Mahtani, K. R., Finch, A., Brewer, G., Saklatvala, J ., and Clark, A. R.
(2000) Mol Cell Biol 20(12), 4265-4274

Lasa, M., Brook, M., Saklatvala, J ., and Clark, A. R. (2001) Mol Cell Biol 21(3),
771-780

Dixon, D. A., Tolley, N. D., King, P. H., Nabors, L. B., McIntyre, T. M.,
Zimmerman, G. A., and Prescott, S. M. (2001) J Clin Invest 108(11), 1657-1665

Murmu, N ., Jung, J ., Mukhopadhyay, D., Houchen, C. W., Riehl, T. E., Stenson,
W. F., Morrison, A. R., Arumugam, T., Dieckgraefe, B. K., and Anant, S. (2004)
Proc Natl Acad Sci U S A 101(38), 13873-13878

Mukhopadhyay, D., Houchen, C. W., Kennedy, 8., Dieckgraefe, B. K., and
Anant, S. (2003) Mol Cell 11(1), 113-126

Rockwell, P., Yuan, H., Magnusson, R., and Figueiredo-Pereira, M. E. (2000)
Arch Biochem Biophys 374(2), 325-333

Figueiredo-Pereira, M. R, Li, Z., Jansen, M., and Rockwell, P. (2002) J Biol
Chem 277(28), 25283-25289

Xiao, G., Chen, W., and Kulmacz, R. J. (1998) J Biol Chem 273(12), 6801-6811

Otto, J. C., DeWitt, D. L., and Smith, W. L. (1993) J Biol Chem 268(24), 18234-
18242

Garavito, R. M., Malkowski, M. G., and DeWitt, D. L. (2002) Prostaglandins
Other Lipid Mediat 68-69, 129-152

Malkowski, M. G., Ginell, S. L., Smith, W. L., and Garavito, R. M. (2000)
Science 289(5486), 1933-1937

Rieke, C. J ., Mulichak, A. M., Garavito, R. M., and Smith, W. L. ( 1999) J Biol
Chem 274(24), 17109-17114

181

155.

156.

157.

158.

159.

160.

161.

162.

163.

164.

165.

166.

167.

Kulmacz, R. J ., and Wang, L. H. (1995) J Biol Chem 270(41), 24019-24023

Smith, W. L., and Song, I. (2002) Prostaglandins Other Lipid Mediat 68-69, 115-
128

Kiefer, J. R., Pawlitz, J. L., Moreland, K. T., Stegeman, R. A., Hood, W. F.,
Gierse, J. K., Stevens, A. M., Goodwin, D. C., Rowlinson, S. W., Mamett, L. J .,
Stallings, W. C., and Kurumbail, R. G. (2000) Nature 405(6782), 97-101

Dietz, R., Nastainczyk, W., and Ruf, H. H. (1988) Eur J Biochem 171(1-2), 321-
328

Lu, G., Tsai, A. L., Van Wart, H. E., and Kulmacz, R. J. (1999) J Biol Chem
274(23), 16162-16167

Landino, L. M., Crews, B. C., Gierse, J. K., Hauser, S. D., and Mamett, L. J.
(1997) J Biol Chem 272(34), 21565-21574

Laneuville, O., Breuer, D. K., Xu, N., Huang, Z. H., Gage, D. A., Watson, J. T.,
Lagarde, M., DeWitt, D. L., and Smith, W. L. (1995) J Biol Chem 270(33),
19330-19336

Mevkh, A. T., Miroshnikov, K. A., Igumnova, N. D., and Varfolomeev, S. D.
(1993) FEBS Lett 321(2-3), 205-208

Mevkh, A. T., Miroshnikov, K. A., Igumnova, N. D., and Varfolomeev, S. D.
(1993) Biokhimiia 58(10), 1573-1579

Meusser, B., Hirsch, C., J arosch, E., and Sommer, T. (2005) Nat Cell Biol 7(8),
766-772

McCracken, A. A., and Brodsky, J. L. (2003) Bioessays 25(9), 868-877

Spiro, R. G. (2004) Cell Mol Life Sci 61(9), 1025-1041

Stoka, V., Turk, B., and Turk, V. (2005) IUBMB Life 57(4-5), 347-353

182

168. Rubinsztein, D. C. (2006) Nature 443(7113), 780-786

169. Ciechanover, A. (2005) Cell Death Differ 12(9), 1178-1190

170. Finn, P. F., and Dice, J. F. (2006) Nutrition 22(7-8), 830-844

171. Zwickl, P., Voges, D., and Baumeister, W. (1999) Philos Trans R Soc Lond B
Biol Sci 354(1389), 1501-1511

172. Peters, J. M., Cejka, 2., Harris, J. R., Kleinschmidt, J. A., and Baumeister, W.
(1993) J Mol Biol 234(4), 932-937

173. Adams, J. (2003) Cancer Treat Rev 29 Suppl 1, 3-9

174. Wang, J ., and Maldonado, M. A. (2006) Cell Mol Immunol 3(4), 255-261

175. Miller, J ., and Gordon, C. (2005) FEBS Lett 579(15), 3224-3230

176. Yao, T., and Cohen, R. E. (2002) Nature 419(6905), 403-407

177. Braun, B. C., Glickman, M., Kraft, R., Dahlmann, B., Kloetzel, P. M., Finley, D.,
and Schmidt, M. (1999) Nat Cell Biol 1(4), 221-226

178. Coffino, P. (2001) Nat Rev Mol Cell Biol 2(3), 188-194

179. Newman, R. M., Mobascher, A., Mangold, U., Koike, C., Diah, S., Schmidt, M.,
Finley, D., and letter, B. R. (2004) J Biol Chem 279(40), 41504-41511

180. Zhang, M., Pickart, C. M., and Coffino, P. (2003) EMBO J 22(7), 1488-1496

181. Kahana, C., Asher, G., and Shaul, Y. (2005) Cell Cycle 4(11), 1461-1464

182. Hampton, R. Y. (2002) Curr Opin Cell Biol 14(4), 476-482

183

183.

184.

185.

186.

187.

188.

189.

190.

191.

192.

193.

194.

195.

196.

197.

McCracken, A. A., and Brodsky, J. L. (2005) Curr Top Microbiol Immunol 300,
17—40

Pety de Thozee, C., and Ghislain, M. (2006) ScientificWorldJoumal 6, 967-983

Hirsch, C., Jarosch, E., Sommer, T., and Wolf, D. H. (2004) Biochim Biophys
Acta 1695(1-3), 215-223

Trombetta, E. S., and Parodi, A. J. (2003) Annu Rev Cell Dev Biol 19, 649-676

Guo, W., Shea, J. E., and Berry, R. S. (2005) Ann N Y Acad Sci 1066, 34-53

Carvalho, P., Goder, V., and Rapoport, T. A. (2006) Cell 126(2), 361-373

Ismail, N., and N g, D. T. (2006) Cell 126(2), 237-239

Ahner, A., and Brodsky, J. L. (2004) Trends Cell Biol 14(9), 474-478

Mueller, B., Lilley, B. N., and Ploegh, H. L. (2006) J Cell Biol 175(2), 261-270

Nishikawa, S. I., Fewell, S. W., Kato, Y., Brodsky, J. L., and Endo, T. (2001) J
Cell Biol 153(5), 1061-1070

Brodsky, J. L., Werner, E. D., Dubas, M. E., Goeckeler, J. L., Kruse, K. B., and
McCracken, A. A. (1999) J Biol Chem 274(6), 3453-3460

McCracken, A. A., and Brodsky, J. L. (1996) J Cell Biol 132(3), 291-298

Pollock, S., Kozlov, G., Pelletier, M. F., Trempe, J. F., Jansen, G., Sitnikov, D.,
Bergeron, J. J ., Gehring, K., Ekiel, I., and Thomas, D. Y. (2004) Embo J 23(5),
1020-1029

Caramelo, J. J ., Castro, 0. A., Alonso, L. G., De Prat-Gay, G., and Parodi, A. J.
(2003) Proc Natl Acad Sci U S A 100(1), 86-91

Trombetta, E. S., and Helenius, A. (2000) J Cell Biol 148(6), 1123-1129

184

198.

199.

200.

201 .
202.
203.

204.

205.

206.

207.

208.

209.

210.

Jakob, C. A., Burda, P., Roth, J ., and Aebi, M. (1998) J Cell Biol 142(5), 1223-
1233

Tokunaga, F., Brostrom, C., Koide, T., and Arvan, P. (2000) J Biol Chem
275(52), 40757-40764

Hosokawa, N., Tremblay, L. 0., You, 2., Herscovics, A., Wada, I., and Nagata,
K. (2003) J Biol Chem 278(28), 26287-26294

Denic, V., Quan, E. M., and Weissman, J. S. (2006) Cell 126(2), 349-359

Kim, W., Spear, E. D., and Ng, D. T. (2005) Mol Cell 19(6), 753-764

Gnann, A., Riordan, J. R., and Wolf, D. H. (2004) Mol Biol Cell 15(9), 4125-4135

Oda, Y., Hosokawa, N., Wada, I., and Nagata, K. (2003) Science 299(5611),
1394-1397

Molinari, M., Calanca, V., Galli, C., Lucca, P., and Paganetti, P. (2003) Science
299(5611), 1397-1400

Plemper, R. K., Bohmler, S., Bordallo, J., Sommer, T., and Wolf, D. H. (1997)
Nature 388(6645), 891-895

Kabani, M., Kelley, S. S., Morrow, M. W., Montgomery, D. L., Sivendran, R.,
Rose, M. D., Gierasch, L. M., and Brodsky, J. L. (2003) Mol Biol Cell 14(8),
3437-3448

Fagioli, C., Mezghrani, A., and Sitia, R. (2001) J Biol Chem 276(44), 40962-
40967

Tortorella, D., Story, C. M., Huppa, J. B., Wiertz, E. J ., Jones, T. R., Bacik, I.,
Bennink, J. R., Yewdell, J. W., and Ploegh, H. L. (1998) J Cell Biol 142(2), 365-
376

Kalies, K. U., Allan, S., Sergeyenko, T., Kroger, H., and Romisch, K. (2005)
EMBO J 24(13), 2284-2293

185

211.

212.

213.

214.

215.

216.

217.

218.

219.

220.

221.

222.

223.

Wiertz, E. J., Tortorella, D., Bogyo, M., Yu, J ., Mothes, W., Jones, T. R.,
Rapoport, T. A., and Ploegh, H. L. (1996) Nature 384(6608), 432-438

Pariyarath, R., Wang, H., Aitchison, J. D., Ginsberg, H. N., Welch, W. J .,
Johnson, A. E., and Fisher, E. A. (2001) J Biol Chem 276(1), 541-550

Pilon, M., Schekman, R., and Romisch, K. (1997) EMBO J 16(15), 4540-4548

Van den Berg, B., Clemons, W. M., Jr., Collinson, I., Modis, Y., Hartmann, E.,
Harrison, S. C., and Rapoport, T. A. (2004) Nature 427(6969), 36-44

Clemons, W. M., Jr., Menetret, J. F., Akey, C. W., and Rapoport, T. A. (2004)
Curr Opin Struct Biol 14(4), 390-396

Wirth, A., Jung, M., Bies, C., Frien, M., Tyedmers, J ., Zimmermann, R., and
Wagner, R. (2003) Mol Cell 12(1), 261-268

Hamman, B. D., Chen, J. C., Johnson, E. E., and Johnson, A. E. (1997) Cell
89(4), 535-544

Lilley, B. N., and Ploegh, H. L. (2004) Nature 429(6994), 834-840

Ye, Y., Shibata, Y., Yun, C., Ron, D., and Rapoport, T. A. (2004) Nature
429(6994), 841-847

Jarosch, E., Taxis, C., Volkwein, C., Bordallo, J ., Finley, D., Wolf, D. H., and
Sommer, T. (2002) Nat Cell Biol 4(2), 134-139

J arosch, E., Geiss-Friedlander, R., Meusser, B., Walter, J ., and Sommer, T. (2002)
Traffic 3(8), 530-536

Oda, Y., Okada, T., Yoshida, H., Kaufman, R. J ., Nagata, K., and Mori, K. (2006)
J Cell Biol 172(3), 383-393

Pye, V. E., Beuron, F., Keetch, C. A., McKeown, C., Robinson, C. V., Meyer, H.
H., Zhang, X., and Freemont, P. S. (2007) Proc Natl Acad Sci U S A 104(2), 467-
472

186

224.

225.

226.

227.

228.

229.

230.

231.

232.

233.

234.

235.

236.

237.

Chen, B., Mariano, J ., Tsai, Y. C., Chan, A. H., Cohen, M., and Weissman, A. M.
(2006) Proc Natl Acad Sci U S A 103(2), 341-346

Fang, S., Ferrone, M., Yang, C., Jensen, J. P., Tiwari, S., and Weissman, A. M.
(2001) Proc Natl Acad Sci U S A 98(25), 14422-14427

Schulze, A., Standera, S., Buerger, E., Kikkert, M., van Voorden, S., Wiertz, E.,
Koning, F., Kloetzel, P. M., and Seeger, M. (2005) J Mol Biol 354(5), 1021-1027

Zhong, X., Shen, Y., Ballar, P., Apostolou, A., Agami, R., and Fang, S. (2004) J
Biol Chem 279(44), 45676-45684

Schuberth, C., and Buchberger, A. (2005) Nat Cell Biol 7(10), 999-1006

Bar-Nun, S. (2005) Curr Top Microbiol Immunol 300, 95-125

Katiyar, S., Joshi, S., and Lennarz, W. J. (2005) Mol Biol Cell 16(10), 4584-4594

Li, G., Zhao, G., Zhou, X., Schindelin, H., and Lennarz, W. J. (2006) Proc Natl
Acad Sci U S A 103(22), 8348-8353

Ravid, T., Doolman, R., Avner, R., Harats, D., and Roitelman, J. (2000) J Biol
Chem 275(46), 35840-35847

Murray, B. P., and Correia, M. A. (2001) Arch Biochem Biophys 393(1), 106-116

Webster, J. M., Tiwari, S., Weissman, A. M., and Wojcikiewicz, R. J. (2003) J
Biol Chem 278(40), 38238-38246

Alzayady, K. J ., Panning, M. M., Kelley, G. G., and Wojcikiewicz, R. J. (2005) J
Biol Chem 280(41), 34530-34537

Sever, N., Song, B. L., Yabe, D., Goldstein, J. L., Brown, M. S., and DeBose-
Boyd, R. A. (2003) J Biol Chem 278(52), 52479-52490

Sever, N., Yang, T., Brown, M. S., Goldstein, J. L., and DeBose-Boyd, R. A.
(2003) Mol Cell 11(1), 25-33

187

238.

239.

240.

241.

242.

243.

244.

245.

246.

247.

248.

249.

250.

251.

Song, B. L., Javitt, N. B., and DeBose-Boyd, R. A. (2005) Cell Metab 1(3), 179-
189

Song, B. L., Sever, N., and DeBose-Boyd, R. A. (2005) Mol Cell 19(6), 829-840

Evett, G. E., Xie, W., Chipman, J. G., Robertson, D. L., and Simmons, D. L.
(1993) Arch Biochem Biophys 306(1), 169-177

Herschman, H. R. (1996) Biochim Biophys Acta 1299(1), 125-140

Shao, J., Sheng, H., Inoue, H., Morrow, J. D., and DuBois, R. N. (2000) J Biol
Chem 275(43), 33951-33956

Tsai, B., Ye, Y., and Rapoport, T. A. (2002) Nat Rev Mol Cell Biol 3(4), 246-255

Ellgaard, L., and Helenius, A. (2001) Curr Opin Cell Biol 13(4), 431-437

Robertson, R. P. (1998) Diabetes 47(9), 1379-1383

Ristimaki, A., Honkanen, N., Jankala, H., Sipponen, P., and Harkonen, M. (1997)
Cancer Res 57(7), 1276-1280

Nemeth, J. F., Hochgesang, G. P., Jr., Mamett, L. J ., and Caprioli, R. M. (2001)
Biochemistry (Mose) 40(10), 3109-31 16

Fagioli, C., and Sitia, R. (2001) J Biol Chem 276(16), 12885-12892

Hosokawa, N ., Wada, I., Hasegawa, K., Yorihuzi, T., Tremblay, L. O.,
Herscovics, A., and Nagata, K. (2001) EMBO Rep 2(5), 415-422

Kitzmuller, C., Caprini, A., Moore, S. E., Frenoy, J. P., Schwaiger, E.,
Kellermann, O., Ivessa, N. E., and Ermonval, M. (2003) Biochem J 376(Pt 3),
687-696

Tsuda, H., Tokunaga, F., Nagarnitsu, H., and Koide, T. (2006) Thromb Res
117(3), 323-331

188

252.

253.

254.

255.

256.

257.

258.

259.

260.

261.

262.

263.

264.

265.

Yu, Y., Fan, J ., Hui, Y., Rouzer, C. A., Mamett, L. J ., Klein-Szanto, A. J .,
FitzGerald, G. A., and Funk, C. D. (2007) J Biol Chem 282(2), 1498-1506

Cok, S. J., and Morrison, A. R. (2001) J Biol Chem 276(25), 23179-23185

Dixon, D. A., Kaplan, C. D., McIntyre, T. M., Zimmerman, G. A., and Prescott,
S. M. (2000) J Biol Chem 275(16), 11750-11757

Kostova, Z., and Wolf, D. H. (2003) EMBO J 22(10), 2309-2317

Spear, E. D., and Ng, D. T. (2005) J Cell Biol 169(1), 73-82

Mbonye, U. R., Wada, M., Rieke, C. J., Tang, H. Y., Dewitt, D. L., and Smith, W.
L. (2006) J Biol Chem 281(47), 35770-35778

Sevigny, M. B., Li, C. F., Alas, M., and Hughes-Fulford, M. (2006) FEBS Lett
580(28-29), 6533-6536

Wu, Y., Swulius, M. T., Moremen, K. W., and Sifers, R. N. (2003) Proc Natl
Acad Sci U S A 100(14), 8229-8234

Wu, Y., Terrnine, D. J ., Swulius, M. T., Moremen, K. W., and Sifers, R. N.
(2007) J Biol Chem 282(7), 4841-4849

Cabral, C. M., Liu, Y., and Sifers, R. N. (2001) Trends Biochem Sci 26(10), 619-
624

Kelleher, D. J ., Kreibich, G., and Gilmore, R. (1992) Cell 69(1), 55-65

Nilsson, I. M., and von Heijne, G. ( 1993) J Biol Chem 268(8), 5798-5801

Bolt, G., Kristensen, C., and Steenstrup, T. D. (2005) Glycobiology 15(5), 541-
547

Tanaka, N., Sato, T., Fujita, H., and Morita, I. (2004) Arterioscler Thromb Vasc
Biol 24(3), 607-612

189

266.

267.

268.

269.

270.

271.

272.

273.

274.

275.

276.

277.

278.

Miyamoto, T., Ogino, N., Yamamoto, S., and Hayaishi, O. (1976) J Biol Chem
251(9), 2629-2636

Xu, X. M., Tang, J. L., Hajibeigi, A., Loose-Mitchell, D. S., and Wu, K. K.
(1996) Am J Physiol 270(1 Pt 1), C259-264 }

Song, 1., Ball, T. M., and Smith, W. L. (2001) Biochem Biophys Res Commun
289(4), 869-875

Wada, M., Yokoyama, C., Hatae, T., Shimonishi, M., Nakamura, M., Irnai, Y.,
Ullrich, V., and Tanabe, T. (2004) J Biochem (Tokyo) 135(4), 455-463

Habenicht, A. J ., Goeri g, M., Grulich, J ., Rothe, D., Gronwald, R., Loth, U.,
Schettler, G., Kommerell, B., and Ross, R. (1985) J Clin Invest 75(4), 1381-1387

Barnett, J ., Chow, J ., Ives, D., Chiou, M., Mackenzie, R., Osen, E., Nguyen, B.,
Tsing, S., Bach, C., Freire, J ., and et al. (1994) Biochim Biophys Acta 1209(1),
130-139

Rowlinson, S. W., Crews, B. C., Lanzo, C. A., and Mamett, L. J. (1999) J Biol
Chem 274(33), 23305-23310

Riese, J ., Hoff, T., Nordhoff, A., DeWitt, D. L., Resch, K., and Kaever, V. (1994)
J Leukoc Biol 55(4), 476-482

Martoglio, B., and Dobberstein, B. (1998) Trends Cell Biol 8(10), 410-415
Sanders, S. L., and Schekman, R. (1992) J Biol Chem 267(20), 13791-13794
Schmitz, A., and Herzog, V. (2004) Eur] Cell Biol 83(10), 501-509

Sato, T ., Nyborg, A. C., Iwata, N., Diehl, T. S., Saido, T. C., Golde, T. E., and
Wolfe, M. S. (2006) Biochemistry (Mose) 45(28), 8649-8656

Loureiro, J ., Lilley, B. N., Spooner, E., Noriega, V., Tortorella, D., and Ploegh, H.
L. (2006) Nature 441(7095), 894-897

190

279.

280.

281.

282.

283.

284.

285.

286.

287.

Serwold, T., Gonzalez, F., Kim, J ., Jacob, R., and Shastri, N. (2002) Nature
419(6906), 480-483

Saric, T., Chang, S. C., Hattori, A., York, I. A., Markant, 8., Rock, K. L.,
Tsujimoto, M., and Goldberg, A. L. (2002) Nat Immunol 3(12), 1169-1176

York, I. A., Chang, S. C., Saric, T., Keys, J. A., Favreau, J. M., Goldberg, A. L.,
and Rock, K. L. (2002) Nat Immunol 3(12), 1177-1184

Saveanu, L., Carroll, 0., Lindo, V., Del Val, M., Lopez, D., Lepelletier, Y., Greer,
F., Schomburg, L., Fruci, D., Niedermann, G., and van Endert, P. M. (2005) Nat
Immunol 6(7), 689-697

Heinemann, F. S., and Ozols, J. (1998) Mol Biol Cell 9(12), 3445-3453

Heinemann, F. S., Mziaut, H., Korza, G., and Ozols, J. (2003) Biochemistry
(Mose) 42(23), 6929-6937

Heinemann, F. S., Korza, G., and Ozols, J. (2003) J Biol Chem 278(44), 42966-
42975

Mziaut, H., Korza, G., and Ozols, J. (2000) Proc Natl Acad Sci U S A 97(16),
8883-8888

Donoso, G., Herzog, V., and Schmitz, A. (2005) Biochem J 387(Pt 3), 897-903

191

 

111111111111111111