The isoprostanes, a new class of natural products: Synthesis and biosynthesis

Article abstract of DOI:10.1055/s-1998-5926

A new class of natural products called isoprostanes are formed in humans as a result of non-enzymatic free-radical-catalyzed lipid peroxidation. Isoprostanes possess biological activity and can be used as an index of five- radical lipid peroxidation and as a marker of oxidative stress. We report here a new and general synthetic methodology for the syntheses of isoprostanes, using the four key lactones which are constructed from D- and L-glucose by a thionocarbonate-mediated radical cyclization reaction. These lactose possess the required stereochemistry and the right functional groups for the syntheses of isoprostanes.

Full text of DOI:10.1055/s-1998-5926

April 1998
SYNTHESIS
569
The Isoprostanes, A New Class of Natural Products: Synthesis and Biosynthesis
a
1
2
b
Joshua Rokach, Subhash P. Khanapure, Seong-Woo Hwang, Mustafa Adiyaman, Laurent Schio, Garret A. FitzGerald
a
Claude Pepper Institute and Department of Chemistry, Florida Institute of Technology, 150 West University Boulevard, Melbourne,
Florida 32901, USA
Fax +1(407)9521818; E-mail: jrokach@fit.edu
Center for Experimental Therapeutics, University of Pennsylvania, Philadelphia, Pennsylvania 19104, USA
b
Received 26 August 1997
Abstract: A new class of natural products called isoprostanes are
formed in humans as a result of non-enzymatic free-radical-cata-
lyzed lipid peroxidation. Isoprostanes possess biological activity
and can be used as an index of free-radical lipid peroxidation and
as a marker of oxidative stress. We report here a new and general
synthetic methodology for the syntheses of isoprostanes, using the
four key lactones which are constructed from D- and L-glucose by
a thionocarbonate-mediated radical cyclization reaction. These
lactones possess the required stereochemistry and the right func-
tional groups for the syntheses of isoprostanes.
Endoperoxide Mechanism for Isoprostane Formation
An endoperoxide mechanism of formation of the isopros-
tanes is proposed in Scheme 1. This is somewhat similar
to the one proposed for the enzymatic formation of PGG2,
the prostaglandin endoperoxide, with the following major
differences:
1) Regioselectivity of the Oxidation Process
Key words: isoprostanes, lactones, free radicals, oxidant stress,
prostaglandins
.
.
Reactive oxygen species (ROS), such as HO , HOO ,
ROO , etc., abstract a hydrogen atom from the doubly al-
.
lylic position in arachidonic acid to generate a radical at
positions 7, 10 and 13. These are trapped by oxygen to yield
six hydroperoxy radicals at positions 5, 8, 9, 11, 12 and 15.
Only four of these, 8-, 9-, 11- and 12-HPETE radicals, 9, 8,
1
. Introduction and Background
An increasing number of recent reports point to beneficial
effects of vitamin E, a free radical scavenger/inhibitor, in
a number of medical ailments including coronary arterial
diseases, Alzheimer’s, cancer, etc. Other reports point to
the improvement in the incidence of cardiovascular disor-
ders due to the consumption of red wine, which contains
flavanoids that presumably act as antioxidants in the same
way as vitamin E. What all the diseases benefitting from
vitamin E appear to have in common is the chronic degen-
erative nature of the process, indicating a slow buildup of
tissue damage, possibly due to a free-radical oxidative
process. There is some debate about how effective vita-
min E is. At issue is the lack of an objective measurement
linking the free-radical process to the clinical symptoms.
1
1 and 10, respectively, can lead to isoprostane formation
by the endoperoxide mechanism shown in Scheme 1. Four
classes of isoprostanes are formed this way and we have
named them with the Roman numerals I to IV.
2
) Stereochemistry of the Oxygen Capture
Scheme 2 shows the details we have proposed for the pro-
cess. The capture of the initial oxygen can occur on both
faces of the arachidonic acid molecule leading to race-
mates. The capture of the second molecule of oxygen will
most likely occur on the two faces. However, the presence
of the four asymmetric centers already in the molecule can
cause a preference for one face over the other. The prefer-
ence for S or R is not yet known and hard to predict.
The isoprostanes, a new class of natural products isomeric
with the well-known prostaglandins, have been uncovered
1
,2
recently in vivo. These are formed by a non-enzymatic
free-radical-induced peroxidation of polyunsaturated fat-
ty acids, e.g. arachidonic acid. Two of these isoprostanes,
IPF -I 2 and 8-iso-PGF 1, have been isolated and iden-
3
) Cyclization Step
As can be seen, 1 and 2 have the two side chains cis to
each other in contrast to the trans geometry present in all
prostaglandins, e.g. PGF2α 3. This is one of the notable
differences between the enzymatic process, yielding the
trans product, and the free-radical processes, affording
the cis structures. It was previously known that an in-
tramolecular radical addition to a double bond affords
2α
2α
tified in human urine using synthetic standards which we
had prepared beforehand by total synthesis (Figure 1).3
,4
We have also reported recently on the total synthesis of
5
,6
1
2-iso-PGF2α and IPF -III.
2α
7
mainly gem cis-substituted products. Depending on the
case, an appreciable amount of the minor trans product
can be formed. The outcome of the cyclization of the more
complex case, such as the one leading to the isoprostanes,
is largely unknown. Three biomimetic studies describing
the cyclization of fatty acid hydroperoxy radicals have
shed some light on the process and report very minor
amounts of trans product formed.8 Due to space ratio-
nalization, we have not incorporated into Scheme 2 the
possible formation of the trans isoprostanes.
–10
The analysis presented above as well as the one in Scheme
3 uses free arachidonic acid. While arachidonic acid exists
Figure 1

5
70
Papers
SYNTHESIS
Scheme 1
as the free acid in many situations in vivo, it is believed Biological Relevance of Isoprostane Formation
that in normal cells it is mainly esterified to the phospho-
The importance of isoprostanes, in our view, encompasses
lipids. The phospholipids have an asymmetric center on
four areas.
carbon 2 of the glycerol moiety. Whether such an asym-
metry will induce some chiral induction in the peroxida- The first, and perhaps the most urgent application of the
tion process described in Schemes 2 and 4 remains to be recent discovery of isoprostanes, is that they are stable
seen.
endpoint products of the free-radical-catalyzed peroxida-
tion of polyunsaturated fatty acids and as such can be
measured as an index of the free-radical process in vivo.
What is making this proposition attractive and promising
is the unique structural nature of the isoprostanes. This
makes them less susceptible to analytical interference by
other natural products generated in vivo. Preliminary re-
ports have already appeared correlating levels of isopros-
Dioxetane Mechanism for Isoprostane Formation
Recently 15-HPETE has been transformed by a free-radi-
cal process to 12-iso-PGF 36 methyl ester as a major
2α
9
product. We have adapted this mechanism to explain the
formation of isoprostanes from four of the six hydroper-
oxy radicals derived from arachidonic acid and the results
are seen in Scheme 3. As can be noted, by this mechanism
the 15-hydroperoxy radical gives rise to Group IV iso-
prostanes and the 5-hydroperoxy to Group I isoprostanes.
If we compare the endoperoxide mechanism (Scheme 1)
11,12
tane 8-epi-PGF 1 to oxidation of LDL,
the presumed
2α
primary step in the generation of atherosclerotic
13
14,15
plaques, to pulmonary hypertension,
2h0epatorenal
16
17–19
syndrome, smoking
and liver diseases.
with the dioxetane one (Scheme 3), we notice another dif- The second area of involvement of isoprostanes is their
ference: the hydroperoxy radicals at positions 8, 9, and 12, mechanism of formation. Arachidonic and other polyun-
1
0, lead to Groups II and III by the first mechanism and saturated fatty acids are mostly present in the syn-2 posi-
the reverse by the other. The 9-hydroperoxy radical 8 and tion of phospholipids, which are the main cell membrane
the 11-hydroperoxy radical 11 cannot lead to isoprostanes constituents. These phospholipids are stacked in bilayers
by this mechanism. The reason that it could in time be in- and are responsible for maintaining the cell integrity.
teresting to find the possible contribution to the generation Free-radical peroxidation of polyunsaturated fatty acids
of isoprostanes by the two mechanisms is the possibility esterified to phospholipids will yield, among other oxy-
that the relative amounts of isomers can be different under genation products, isoprostanes which contain three hy-
the two mechanistic options.
droxyl groups. This polarity, generated in the midst of a

April 1998
SYNTHESIS
571
Scheme 2

5
72
Papers
SYNTHESIS
Scheme 3: Dioxetane Mechanism for the Formation of Isoprostanes
highly hydrophobic milieu, will cause a change in mem- in urine; and which of the individual isoprostanes or
brane shape, disruption of the bilayer, cell leaks and even- groups of isoprostanes are more suitable to measure in
tual cell death.
urine as a non-invasive index of free-radical peroxidation
and tissue damage. The second part of the objective was
the development of a quantitative assay, e.g. GC/MS and
LC/MS to measure isoprostanes in biological fluids. We
have accomplished substantial progress and developed
The third area of importance involving the isoprostanes is
their biological activity: 8-iso-PGF2α 1 and 8-iso-PGE2
have been shown to be potent vasoconstrictors. The exis-
tence of a dedicated or incidental isoprostane receptor has
been put forward.
prostanes possess biological activity similar and/or differ-
2
1,22
two GC/MS assays for the measurement of 8-iso-PGF2α
It is still too early to tell if other iso-
24,25
and IPF -I.
For meaningful measurement it is impor-
2
α
tant to use as internal standard the isotopically labelled
isoprostane corresponding to the isoprostane to be mea-
ent to that observed for 8-iso-PGF .
2α
The fourth area of involvement is the possibility that the sured and avoid if at all possible the use of labelled pros-
isoprostanes interacting with the known prostaglandin re- taglandins and other isomers as an approximation.
ceptors FP (PGF ), EP (PGE ), DP (PGD ), IP (PGI )
2α
2
2
2
and TP (TXA ). Isoprostane production in the body is re- From our perspective, our number one priority when we
2
1
ported to be several fold higher than the prostaglandins.
started this program was the synthesis of isoprostanes.
The isoprostanes, being structurally close to the prosta- Our strategy, which consisted of synthesizing the isopros-
glandins, may interact to some degree with these recep- tane first and then looking for it in biological fluid, was
tors. This interaction can produce proinflammatory decided upon from the start of our involvement in the
conditions in the same way prostaglandins do. For exam- field. It became evident from the beginning, looking at our
ple, we have shown that 12-iso-PGF activates the PGF
receptor in the eye.
analysis of isoprostane formation (Schemes 2 and 3), that
several isoprostanes could be formed, and since they are
all isomeric and isomeric with prostaglandins of the F2
type, overlapping of different isomers was bound to hap-
pen, irrespective of the analytical tool used. This would
render individual isolation and identification very diffi-
2α
2α
2
3
2. Results and Discussion
Our general approach to the isoprostanes and free-radical cult. In addition, the amounts of the individual isopros-
oxidant stress and its involvement in disease states was to tanes produced were suspected to be in the low nanogram
try to understand the mechanism of formation of the iso- to picogram range, which would make the task of structur-
prostanes: which isoprostanes are formed in vivo and in al proof of the individual isoprostanes without synthetic
what proportions; which are eliminated more efficiently standards difficult.

April 1998
SYNTHESIS
573
Radical Cyclization Approach to the Synthesis
of Isoprostanes
A precursor for the key radical cyclization step was pre-
pared as shown in Scheme 6. This intermediate has a
thionocarbonate group as a radical precursor. This group
Since radical cyclizations are known to give 5-membered has the advantage that it can be prepared directly from the
7
rings with cis-oriented substituents, we decided to follow diol as opposed to the traditional thioxanthate, which re-
at first a synthetic strategy based on radical cyclization quires silylating the primary alcohol, introducing the
3
(
Scheme 4).
thioxanthate on the secondary hydroxy group and, at the
end of the sequence, deprotecting the silyl group. We have
used extensively the thionocarbonate group which was
The formation of the cyclopentane ring by generating a
bond between carbons 1 and 5 in 70 was at the center of
this strategy. We thought that if we could control the cy-
clization outcome to yield the syn-anti-syn intermediate
3–6
originally pioneered by Barton et al., who also described
the exclusive formation of the radical on the secondary
26,27
carbon.
We also found that this group is more sturdy
6
8 or the all-syn intermediate 69, we would obtain half
than the thioxanthate in carrying out the sequence de-
scribed in Scheme 6 as reflected by the slightly higher
yield across the board (not shown).
the isoprostanes in Scheme 2 from one single precursor.
In addition, using D-glucose, the same chemistry obvi-
ously applies, will yield the other half of the isopros-
tanes. So with one methodology and L- and D-glucose, The cyclization reaction on 81 was carried out using tribu-
we can have access to all 32 cis isoprostanes shown in tyltin hydride and a radical initiator, 2,2′-azobis(isopropi-
Scheme 2.
onitrile), in refluxing benzene. A 44% yield of the all-syn
-membered lactone 75 was obtained (Scheme 7) along
with a 19% yield of another isomeric 5-membered ring
lactone 104, which was identified as the Corey lactone
when compared with a commercial sample.
5
Scheme 5 shows the correlation between the four target
lactones derived from L- and D-glucose and 16 isopros-
tanes. For the sake of clarity, the other 16 epimers of the
hydroxy on the side chains are not shown but are obvious-
ly accessible from the same lactones.
The hydrogen bonded conformation 83 which stabilizes
the two OHs in the 1,3 diaxial position is an attempt to ex-
plain the preferential formation of the all-syn product 75.
In order to obtain the syn-anti-syn product 73, we have
prepared the bis acetyl derivative 81a to disrupt or elimi-
nate the hydrogen bonding in the transition state and per-
formed the cyclization reaction on this derivative. As can
be seen, lactone 73a was the major product of the reaction.
Three other minor lactones were also obtained (see sec-
tion on lactones). A chair conformation, such as the equa-
torial 1,3-dihydroxy arrangement, 84, can possibly
account for the formation of the major syn-anti-syn lac-
tone.
Another aspect of our strategy is the use of carbohydrate
as our chiral pool. This decision was based on the follow-
ing premises: 1) An assurance of absolute chiral purity,
which we think can be an advantage when these natural
products are made for the first time and are to be used,
among other things, to identify the naturally occurring
products. 2) We require two hydroxyl groups on the 5-
membered ring, a third one to generate a radical, and two
other functionalities at positions 1 and 6 of the sugar for
the subsequent introduction of the two sides of the iso-
prostanes. 3) Affordability of the starting material. Based
on the preceding criteria, we selected D- and L-glucose as
our chiral precursors. Throughout this article we have pro- It is worth mentioning that before the successful attempt
vided information on the chemistry and transformation of of the cyclization reaction leading to the syn-anti-syn lac-
L-glucose. All reactions have been performed with D-glu- tone 73a, as described above, we had attempted another
cose but not shown here except when they were per- experiment (Scheme 8) designed to eliminate the hydro-
formed with derivatives from the D-glucose series only gen bonding in intermediate 83 and also to provide some
(Schemes 8 and 10).
bulk on the OHs and force the two side chains anti to the
Scheme 4: Retrosynthetic Analysis

5
74
Papers
SYNTHESIS
Scheme 5
Scheme 6: Preparation of Radical Cyclization Precursor from L-Glu-
cose
Scheme 7: Radical Cyclization To Give syn-anti-syn and all-syn Lac-
tones

April 1998
SYNTHESIS
575
OH group. Such a cyclic derivative, in addition, was ex-
pected to eliminate the formation of the trans products.
Scheme 8: Radical Cyclization of Isopropylidene Thionocarbonate
Scheme 9
(from D-Glucose Series)
We selected the isopropylidene derivative 85 for that ex- product 92 as such, the crude Wittig reaction mixture of
periment. Conformation 86, because of crowding, is most 81 and 92 could be converted to the TBDMS derivatives
likely not favored and we were hoping that the more fa- which on purification afforded 71% of 81b and 11% of
vored conformation 87 on cyclization would afford the 92b. This two-step, one-pot procedure is the one now used
desired isopropylidene derivative 90. As can be seen from since it produces superior yields, and both products can be
Scheme 8, no cyclization products were observed. Instead used for the cyclization reaction. When the cis derivative
the only product isolated was 91, indicating that the equa- 92b was subjected to the HSnBu -mediated cyclization
3
torial conformation 88 in which the radical is too far from reaction, the reaction spectrum was very much the same as
the double bond is the predominant one. The radical was the one described for the trans. Perhaps it is not surpris-
reduced instead to afford 91. At that point we decided not ing, considering that monitoring of the reaction indicates
to substitute the OHs with a cyclic derivative and proceed- that 92b is very quickly isomerized to the trans product
ed to acetylate the OHs as described above. It is worth 81b under the reaction conditions.
adding that we also derivatized the two OHs in 81 with
In an attempt to prepare the trans derivative 81 uncontam-
TBDMS groups (Scheme 6) so as to have bulkier substit-
inated with the cis product, we performed the reactions
uents on the OH groups in order to improve on the yield
shown in Scheme 10. D-Glucose derivatives were used in
of the syn-anti-syn compound. The results were marginal-
this reaction. The trans product was obtained as intended
ly better and 73b was obtained in 43% yield along with
with little or no contamination of the cis product. Howev-
three other lactones, 101b (7%), 102b (17%) and 103b
er, other impurities appeared which we suspect are due to
(
4%) (Figure 2). The silyl group was kept throughout the
the stronger base/nucleophile 98 used which probably
synthetic transformation, being more stable under basic
conditions.
In the Wittig reaction between 80 and the methyl (triphe-
nylphosphoranylidene)acetate 82 (Scheme 6), in addition
to the trans compound 81, we obtained the cis derivative
92 (~20%), based on NMR (Scheme 9).
On silica gel purification, however, the cis derivative 92
was completely converted to the furano lactone 94. Trace
amounts, ~1%, of 96 were also isolated, indicating a
Michael addition occurs first, as shown, to yield 96 and
probably 95, which was not detected, presumably because
it is transformed to 94. The hydroxy furan 96 could have
also been formed from the trans derivative 81 by a Micha-
Scheme 10: Horner–Wittig Reaction of Thionocarbonate (from D-
el addition. Whereas we could not isolate the cis Wittig Glucose Series)

5
76
Papers
SYNTHESIS
damaged some of the thionocarbonate functional groups. tyro lactone and valero lactone are δ 178.1 and δ 171.2, re-
This reaction, however, was only attempted once. With spectively. Lactone 104 is identical to the commercial
excess reagent, the Michael product 99 was mainly Corey lactone. The only other 5-membered ring lactone
formed. On balance, this procedure is not an advantage possible is 75. This confirms the syn relationship of the
over the Wittig reaction described in Scheme 9.
hydroxy and the side chain and the identity of 75 and 104.
The structure of lactone 102b has been assigned by spec-
troscopic evidence and by treatment with fluoride fol-
lowed by hydrolysis to the Corey lactone 104. In a similar
fashion 101b, on fluoride deprotection and hydrolysis, af-
forded the 5-membered ring lactone 75. The syn-anti-syn
lactone 73a, on hydrolysis, afforded 73. No 5-membered
ring lactone was obtained in this case, confirming the
trans relationship of the OH and the lactone function.
NOE experiments show strong interaction between the
two bridgehead protons in 73c, prepared from 73, and in
Lactones: Structure Elucidation and Correlation
Figure 2 shows the six lactones which can possibly be
formed from L-glucose in the two cyclization reactions.
7
5 between the two cis protons 3a and 4, indicating the cis
relationship of the side chains. As can be seen in the NOE
correlations of 73c and 75b, other NOE correlations con-
firm the structure assigned. Lactone 101, the 6-membered
ring version of 75, because of the close chemical shift for
the bridgehead proton, was unsuitable for the NOE study.
The Corey lactone derivative 104c shows no NOE interac-
tion between the protons 3a and 4. The structure of lactone
103b at this point has not been completely characterized.
As mentioned earlier in the article, all the reactions of L-
glucose have been duplicated with D-glucose. Conse-
quently lactones 74 and 76 (Scheme 5) and their deriva-
tives were also prepared 3 and used in several of the
isoprostane syntheses (Scheme 12). An independent syn-
thesis of lactone 76 has been reported recently.29
,28
On the Mechanism of the Radical Cyclization
In the course of the cyclization of intermediate 81 to af-
ford the lactone 75, as well as a small amount of the trans
lactone 104, we have isolated and assigned a structure to
several byproducts. Scheme 11 shows the products ob-
tained and our explanation for their formation.
This exercise was useful as it allowed us to trace the for-
mation and the fate of the initial radical formed, 105,
which in addition to the intended cyclization reaction
leading to the lactones, followed other courses to yield the
observed byproducts. Trace amounts of 112 are formed,
presumably through the intermediacy of the thio ortho-
ester 108. Intermediate 108 can generate radical 109, a
good candidate for the formation of the observed 110 and
Figure 2
We have obtained all of them. Four of them are 6-mem-
bered ring lactones and are the four possible lactones in
the radical cyclization of 81b. We have in fact in this cy-
clization, in addition to the desired major lactone 73b
(
43%), isolated minor quantities of the three other lac-
1
11. Alternatively, 110 and 111 could be formed by direct
tones 101b (7%), 102b (17%) and 103b (4%) (see Exper-
imental). The two 5-membered ring lactones 75 and 104
are the products of the radical cyclization of the unsubsti-
tuted dihydroxy precursor 81 (Scheme 7). These lactones
are the only two possible 5-membered ring compounds
because of the syn relationship of the hydroxy and the lac-
tone groups. In this cyclization 73 and 103 were not ob-
served. The structural proof of these lactones is based on
cyclization of 105, followed by a loss of the S(SnBu ) on
3
2
an attack by •SnBu as shown. Dioxymethylene radicals
3
have been suggested as intermediate species in the reduc-
30
tion of thionocarbonates to methylenedioxy derivatives.
The primary radical 105, after cleavage of the carbon–oxy-
gen bond yields 106, which goes on to form the desired
lactones. To explain the formation of 107, it is assumed
that the radical in 106 attacks the sulfur as shown and that
the double bond is reduced followed by lactone formation.
1
13
1
1
extensive H, C, H– H COSY, NOE and DEPT NMR
studies as well as chemical transformations. For example
the carbonyl of the 5-membered ring lactones appears at a As can be seen, the transformation of 105 to 106 is not as
1
3
very characteristic δ 177–178 in the C NMR and the cor- efficient and automatic as we would have thought and the
responding value for the 6-membered ring lactone was δ breakage of the C–O bond requires extra activation ener-
1
71–172 (see Experimental). The literature values for bu- gy. In retrospect, it was not so surprising since radical 105

April 1998
SYNTHESIS
577
Scheme 11
Scheme 12

5
78
Papers
SYNTHESIS
13C NMR (CDCl3): δ = 111.1, 109.5, 105.5, 80.3, 78.5, 76.7, 67.0,
is stabilized by two oxygens and one sulfur and the radical
in 106 is not. In a preliminary experiment at higher tem-
perature (reflux in toluene) we found that the formation of
35.1, 26.6, 26.3, 26.0 and 25.0 (2 × C).
3
-Deoxy-5,6-dihydroxy-1,2-O-isopropylidene-â-L-glucofuranose
110 and 111 is substantially reduced but without an in-
(78):
crease in the formation of the desired lactones. Isomeric
thiocarbonate 107, however, is formed in larger amounts.
3-Deoxydiisopropylidene-L-glucose (77) (80 g, 0.327 mol) was dis-
solved in 30% aq HOAc (700 mL), and the solution was stirred at r.t.
for 24 h. The mixture was adjusted to pH 6 by adding 50% aq NaOH,
and extracted with EtOAc (8 × 500 mL). The aqueous portion was
evaporated, and the residue was extracted with EtOAc (2 × 500 mL).
The combined organic layers were washed with brine, dried
(Na SO ), and concentrated. The residue was dissolved in Et O
Synthesis of Isoprostanes
We have used the lactones 74b and 119b, derived from D-
glucose, and 73b and 75, derived from L-glucose, for the
preparation of isoprostanes (Scheme 12). Our initial ob-
jective was to prepare a representative of each of the four
classes of isoprostanes of the syn-anti-syn type, IPF -I 2,
2
4
2
(900 mL) and upon cooling to –20°C solidified to afford 54.13 g
(
81%) of diol 78 as a white fluffy powder, mp 81–82°C, [α] +10.73
D
(c = 1.0, MeOH).
2α
C H O
5
calc
found
C
52.93
52.76
H 7.90
7.90
IPF -II 27, IPF -III 31 and IPF -IV (8-iso-PGF ) 1, in
9
16
2α
2α
2α
2α
(204.2)
which the two side chains are β and the 2 OHs are α to the
3
–1
IR (KBr): ν = 3375 cm (br, OH).
plane of the ring. We also prepared representatives of the
1
H NMR (CDCl ): δ = 5.81 (d, J = 3.5 Hz, 1 H), 4.76 (t, J = 4.2 Hz, 1
all-syn isoprostanes, 8-,12-iso-IPF -IV (12-iso-PGF )
3
2
α
2α
5
H), 4.19 (dt, J = 10.7 and 4.3 Hz, 1 H), 3.89 (m, 1 H), 3.71 (dd, J =
3
6
and 8-,12-iso-IPF -I 25. An enantiomer of 8-iso-
PGF2α 53 was also prepared by total synthesis. Indepen-
dent syntheses of 12-iso-PGF , (±)12-iso-PGF
±)8-epi-PGF2α ethyl ester have appeared recently.
2α
11.5 and 2.9 Hz, 1 H), 3.55 (dd, J = 1.3 and 16.9 Hz, 1 H), 3.26 (br s,
3
2 H), 2.05 (dd, J = 13.4 and 4.4 Hz, 1 H), 1.84 (m, 1 H), 1.51 (s, 3 H),
3
1
32
and
2α
2α
1.32 (s, 3 H).
3
3
13
(
C NMR (CDCl ): δ = 111.3, 105.1, 80.5, 78.4, 72.2, 63.5, 33.7, 26.6
3
and 26.0.
So far we have identified two of these isoprostanes in
urine, 8-iso-PGF2α 1 of Class IV isoprostanes and the
more abundant IPF -I 2 of Class I isoprostanes. The pos-
3
-Deoxy-1,2-O-isopropylidene-5,6-O-thionocarbonate-â-L-gluco-
2α
furanose (79):
itive identification and isolation of IPF -I of Class I iso-
2α
To a solution of 3-deoxy-1,2-O-isopropylidene-L-glucose (78)
(23.84 g, 0.116 mol) in anhyd 1,2-dichloroethane (0.5 L) was added
1,1′-thiocarbonyldiimidazole (25 g, 0.139 mol) at r.t. under Ar. The
mixture was refluxed for 1 h and then was allowed to cool to r.t., and
the solvent was removed under reduced pressure. The oily residue
was dissolved in EtOAc (400 mL) and partitioned with 5% aq HCl
prostanes, represents the first proof of the existence of
more than one class of isoprostanes, reinforcing the anal-
ysis presented in Scheme 2 and 3.
NMR spectra were recorded on a Bruker AMX-360 NMR spectrom-
eter. Proton chemical shifts are reported in parts per million (ppm)
downfield from TMS unless noted otherwise. Coupling constants (J)
are reported in Hertz (Hz). Carbon chemical shifts are reported in
(200 mL). The resulting mixture was stirred for 20 min at r.t., and the
organic layer was separated. The aqueous phase was extracted with
EtOAc (3 x 500 mL). The combined organic layers were washed with
H O, brine, dried (Na SO ), and concentrated. The product was cry-
parts per million (ppm) with reference to internal solvent CDCl3
2
2
4
(
77.00 ppm), C D (128.39 ppm) or CD COCD (29.92 ppm) unless
6 6 3 3
stallized from EtOAc (200 mL) to yield 26.2 g (91%) of the pure
product 79 as a light yellow powder, mp 130–131°C, [α]D +14.1
1
noted otherwise. Assignments in the H NMR spectra were made on
the basis of homonuclear decoupling experiments or two-dimensional
correlation spectroscopy (COSY) experiments. Relative stereochem-
ical assignments were made on the basis of NOE experiments. As-
signments in the C NMR spectra were confirmed by distortionless
enhancement by polarization transfer (DEPT) spectroscopy experi-
ments.
(c = 1.0, MeOH).
C H O S calc
C
48.77
48.74
H 5.73
5.86
10
14
5
1
3
(246.28)
found
IR (KBr): υ = 1297 cm–1 (C=S).
1
H NMR (CDCl ): δ = 5.84 (d, J = 3.5 Hz, 1 H), 4.92 (m, 1 H), 4.81
3
(t, J = 4.1 Hz), 4.74 (t, J = 8.7 Hz, 1 H), 4.58 (dd, J = 9.0 and 6.7 Hz,
IR spectra were recorded on a Nicollet 550 FTIR spectrometer. Opti-
cal rotations were recorded on a JASCO model DIP-360 digital pola-
rimeter. Electron impact (EI), chemical ionization (CI) or fast atom
bombardment (FAB) high-resolution mass spectra (HRMS) were ob-
tained from Midwest Center for Mass Spectrometry, University of
Nebraska, Lincoln.
1
H), 4.40 (q, J = 10.5 and 4.9 Hz, 1 H,), 2.33 (dd, J = 13.4 and 4.5 Hz,
H), 1.68 (m, 1 H), 1.52 (s, 3 H), 1.33 (s, 3 H).
1
13
C NMR (CDCl ): δ = 191.41, 112.3, 105.9, 81.9, 80.3, 76.9, 71.2,
3
35.4, 26.9 and 26.2.
3-Deoxy-1,2-dihydroxy-5,6-O-thionocarbonate-L-glucofuranose
All nonaqueous reactions were carried out under Ar with freshly dis-
tilled solvents under anhydrous conditions unless otherwise noted.
Most reagents were purchased from commercial suppliers and used
without further purification unless otherwise noted.
(
3
(
80):
-Deoxy-1,2-O-isopropylidene-5,6-O-thionocarbonate-L-glucose
79) (43 g, 0.175 mol) was dissolved in THF (700 mL) at r.t.. To the
stirred solution, aq 4% H SO (500 mL) was added, and the mixture
2
4
was refluxed for 3 h. The mixture was allowed to cool to r.t., and to it
3
-Deoxy-1,2:5,6-di-O-isopropylidene-â-L-glucofuranose (77):
was added sat. aq NH Cl (300 mL). The organic layer was separated,
4
L-Glucose was converted to diacetonide L-glucose in a 91% yield as
and the aqueous layer was extracted with EtOAc (4 × 300 mL). The
combined THF solution and EtOAc extracts were concentrated under
reduced pressure to give the crude product that was dissolved in
3
4
described for the D-glucose. The diacetonide L-glucose was then
converted to 77 in a two-step procedure in 84% yield as described in
2
6
the literature for diacetonide D-glucose.
EtOAc (300 mL), washed with H O (2 × 200 mL), brine, dried
2
C H O
calc
found
C
59.00
59.46
H 8.25
8.61
(Na SO ), and concentrated. The product was crystallized from
EtOAc to yield 30.5 g (85%, as a mixture of diastereomer) of lactol
12
20
5
2
4
(
244.28)
1
80 as a white solid, mp 140–141°C.
IR (KBr): ν = 3355 and 3271 (br, 2 × OH), 1298 cm (C=S).
H NMR (CDCl ): δ = 5.82 (d, J = 3.6 Hz, 1 H), 4.76 (t, J = 4.2 Hz, 1
3
–1
H), 4.20–4.08 (m, 3 H), 3.83 (m, 1 H), 2.18 (dd, J = 13.2 and 3.9 Hz,
1
3
H), 1.77 (m, 1 H), 1.52 (s, 3 H), 1.43 (s, 3 H), 1.36 (s, 3 H), 1.32 (s,
H).
C H O S
(206.22)
calc
found
C
40.77
41.18
H 4.89
5.12
7
10
5

April 1998
SYNTHESIS
579
1
H NMR (CD COCD ): δ = 5.84 (d, J = 4.0 Hz, 1 H), 4.38 (d, J = 4.0
of Bu SnH (361 µL, 1.146 mmol) and AIBN (65 mg, 0.4 mmol) in an-
3
3
3
Hz, 1 H), 4.31 (m, 1 H), 4.05 (dd, J = 8.4 and 6.0 Hz, 1 H), 3.88 (t, J
8.6 Hz, 1 H), 3.76 (m, 1 H), 3.46 (d, J = 3.6 Hz, 1 H), 3.24 (t, J =
hyd degassed toluene (20 mL) by a syringe addition pump during a
period of 3 h under Ar. The mixture was allowed to cool to r.t., and
=
4
.0 Hz, 1 H), 1.08 (dd, J = 13.4 and 6.0 Hz, 1 H), 0.91 (m, 1 H).
C NMR (CD COCD ): δ = 191.5, 104.5, 84.7, 79.2, 76.9, 71.2, 34.7.
H O (0.5 mL) was added and the solvent was removed under reduced
2
13
pressure. The purification of the product by column chromatography
using 2% EtOAc in hexane afforded the less polar desired lactone 73b
(113.5 mg, 43%) as a white solid followed by all-syn 6-membered
ring lactone 101b (18 mg, 7%), and followed by a mixture of 102b
and 103b (54 mg, 21%).
3
3
Methyl (E)-3,5-Dideoxy-4,6-dihydroxy-7,8-O-thionocarbonate-L-
glucooct-2-enoate (81):
A mixture of 80 (2.0 g, 9.7 mmol) and methyl (triphenylphosphor-
anylidene)acetate 82 (6.47 g, 19.4 mmol) in THF (130 mL) was
stirred under Ar at r.t. for 24 h. The solvent was removed under re-
duced pressure, and the residue was dissolved in EtOAc (150 mL).
The resulting solution was washed with H O (2 × 40 mL), brine (40
mL), dried (Na SO ), and concentrated. Purification of the product by
column chromatography using 3% MeOH in CH Cl yielded the less
The data for 73b: mp 96–97°C.
C H O Si
2
calc
found
C
59.95
59.55
H 10.06
9.98
20
40
4
(400.7)
2
–1
IR (KBr): υ = 1749 (C=O) cm .
2
4
1
H NMR (C D ): δ = 3.95–3.86 (m, 1 H), 3.89 (dd, J = 11.9 and 1.7
6
6
2
2
Hz, 1 H), 3.68–3.61 (m, 1 H), 3.44 (dd, J = 11.9 and 3.8 Hz, 1 H), 2.38
dd, J = 15.3 and 2.4 Hz, 1 H), 2.13 (m, 1 H), 2.05 (m, 1 H), 1.87 (m,
polar furano lactone 94 (335 mg, 19%) yield followed by the desired
E) alkenic compound 81 (1.55 g, 61%) as a colorless oil.
(
(
–
1
2 H), 1.72 (q, J = 10.9 Hz, 1 H), 0.94 (s, 9 H), 0.92 (s, 9 H), 0.08 (s, 3
IR (neat): υ = 3420 (br, OH), 1723 (C=O), 1657 cm (C=C).
1
H), 0.00 (s, 3 H), 0.03 (s, 3 H), 0.04 (s, 3 H).
H NMR (CDCl ): δ = 6.91 (dd, J = 15.6 and 4.6 Hz, 1 H), 6.08 (d, J
3
13
C NMR (C D ): δ = 171.7, 75.6, 72.2, 67.4, 45.3, 44.13, 42.5, 32.7,
6
6
=
15.7 Hz, 1 H), 4.85 (m, 1 H), 4.75 (m, 2 H), 4.65 (m, 1 H), 4.43 (d,
26.2, 18.4, 18.5, –4.1, –4.2, 4.4 and –4.5.
J = 3.3 Hz, 1 H), 4.25 (m, 1 H), 3.96 (d, J = 3.0 Hz, 1 H), 1.88 (m, 1
H), 1.61 (m, 1 H).
The spectral data for (4aR,5S,7R,7aS)-hexahydro-3-oxo-5,7-
bis[(tert-butyldimethylsilyl)oxy]cyclopentapyran (101b):
H NMR (C6D6): δ = 4.42 (t, J = 7.4 Hz, 1 H), 3.86 (t, J = 3.4 Hz, 1
1
3
C NMR (CDCl ): δ 192.1, 167.2, 148.9, 120.6, 83.8, 70.3, 70.2,
3
1
6
9.8, 52.2, 37.6.
H), 3.72 (dd, J = 10.0 and 7.3 Hz, 1 H), 3.59 (dd, J = 10.0 and 7.7 Hz,
1 H), 2.58 (dd, J = 18.0 and 4.8 Hz, 1 H), 2.34 (m, 1 H), 2.08 (dd, J =
Spectral data for furano lactone (94):
1
H NMR (CDCl ): δ = 5.2 (t, J = 4.4 Hz, 1 H), 5.0–4.86 (m, 2 H), 4.54
3
1
3
0
8.0 and 11.8 Hz, 1 H), 1.88 (d, J = 15.0 Hz, 1 H), 1.55 (dq, J = 7.7,
.1 Hz, 1 H), 1.14 (ddd, J = 10.7, 6.8 and 3.8 Hz, 1 H), 0.97 (s, 18 H),
.05 (s, 6 H), 0.03 (s, 3 H), -0.02 (s, 3 H).
(t, J = 8.8 Hz, 1 H), 4.51 (dd, J = 6.7 and 9.0 Hz, 1 H), 4.38 (dt, J =
5
1
9
.4 and 10.2 Hz, 1 H), 2.8 (dd, J = 5.4 and 18.7 Hz, 1 H), 2.73 (d, J =
8.5 Hz, 1 H), 2.62 (dd, J = 5.6 and 14.2 Hz, 1 H), 1.97 (ddd, J = 5.0,
.8, 14.5 Hz, 1 H)
1
3
C NMR (C D ): δ = 176.7, 84.0, 74.3, 60.6, 50.8, 42.6, 38.8, 30.7,
6
6
1
3
26.4, 26.3, 18.7, 18.6, –4.2, –4.9, –5.1.
C NMR (CDCl ): δ = 191.1, 83.9, 81.9, 79.6, 79.6, 70.9, 36.7, 35.0.
3
Conversion of Lactone 101b to 75:
Lactone 101b was treated with TBAF (7 equiv) in THF solution at r.t.
overnight. The mixture was then treated with aq sat. NH Cl solution,
extracted with EtOAc and upon usual workup gave a product which
Methyl (E)-3,5-Dideoxy-4,6-bis[(tert-butyldimethylsilyl)oxy]-7,8-
O-thionocarbonate-L-glucooct-2-enoate (81b):
4
This is a preferred two-step one-pot procedure. The preceding prepa-
ration was performed on 6 g of 80. The crude mixture, before chro-
matography, was dissolved in anhyd DMF (90 mL) and was treated
with imidazole (15.8 g, 0.232 mol) and tert-butyldimethylsilyl chlo-
ride (21.85 g, 0.145 mol) at r.t. under Ar. The mixture was stirred for
was identical to 75.
Conversion of 102b to 104:
1
The R values for the trans lactones 102b and 103b are the same. H
f
1
8 h at 60°C, then cooled to r.t. and was quenched with H O (50 mL),
2
13
and C NMR analyses revealed 102b and 103b to be a mixture of ap-
and extracted with EtOAc (3 × 50 mL). The combined extracts were
proximately 80:20. On rechromatography, some of the less polar frac-
tions gave a pure sample of 102b. The treatment of 102b with excess
TBAF, as mentioned above, gave a product identical to the Corey lac-
tone 104.
washed with brine, dried (Na SO ), and concentrated. The purifica-
2
4
tion by flash column chromatography (10% EtOAc in hexane) gave
the less polar Z-isomer 92b (1.55 g, 11%) as an oil followed by the
major E-isomer 81b (10.1 g, 71%) as a white solid: mp 109–110°C;
[
α] –26.1 (c = 1.0, MeOH).
D
(
3aR,4R,5R,6aS)-Hexahydro-5-hydroxy-4-(hydroxymethyl)-2H-
C H O SSi calc
C
53.84
53.96
H 8.63
8.93
22
42
6
2
cyclopenta[b]furan-2-one (75):
(490.81)
found
In a three-neck 500 mL round-bottom flask was placed thionocarbon-
ate 81 (1.0 g, 3.81 mmol) in anhyd degassed benzene (320 mL), and
the mixture was refluxed under Ar. To the refluxing solution was add-
ed a solution of Bu SnH (2 mL, 7.62 mmol) and 2,2′-azobis(2-meth-
ylpropionitrile) (375 mg, 2.29 mmol) in anhyd degassed benzene
–1
IR (KBr): υ = 1729 (C=O) cm .
1
H NMR (C D ): δ = 6.9 (dd, J = 15.5 and 4.5 Hz, 1 H), 6.16 (dd, J =
6
6
1
5.5, 1.6 Hz, 1 H), 4.24 (m, 1 H), 4.1 (t, J = 8.0 Hz, 1 H), 4.01 (m, 1
3
H), 3.92 (m, 1 H), 3.62 (t, J = 8.4 Hz, 1 H), 1.46 (m, 1 H), 1.04 (m, 1
H), 0.94 (s, 9 H), 0.84 (s, 9 H), 0.21 (s, 3 H), 0.04 (s, 3 H), -0.09 (s, 3
H), -0.11 (s, 3 H).
(
60 mL) by a syringe addition pump during a period of 3 h. The mix-
ture was allowed to cool to r.t., and H O (1.5 mL) was added. The sol-
vents were removed under reduced pressure, and the product was
purified by column chromatography using 2% MeOH in CH Cl to
yield 110, 120 mg (13%), 107, (<1%), 111, 85 mg (8%), the major de-
sired lactone 75, 288 mg (44%), Corey lactone 104, 124 mg, (19%)
and 112, (1–2%).
1
3
2
C NMR (C D ): δ = 191.8, 166.5, 149.4, 121.5, 83.1, 68.8, 68.2,
6
6
6
7.9, 51.7, 41.6, 26.3, 26.2, 18.5, 18.4, –4.1, –4.2, –4.3 and –4.5
2
2
(4 × C).
Spectral data for Z-isomer (92b):
1
H NMR (CDCl ): δ = 6.3 (dd, J = 6.8 and 11.5 Hz, 1 H), 5.8 (d, J =
3
1
1.5 Hz, 1 H), 5.3 (t, J = 7.8 Hz, 1 H), 5.12 (t, J = 8.3 Hz, 1 H), 4.73
d, J = 7.8 Hz, 2 H), 4.28 (m, 1 H), 3.7 (s, 3 H), 1.88 (ddd, J = 3.2, 9.7,
and 13.9 Hz, 1 H), 1.48 (m, 1 H), 0.9 (ds, 18 H), 0.8–0.0 (4 × s, 12 H).
Data of 75: A white solid; mp 126–127°C.
(
C H O
4
(172.18)
calc
found
C
55.81
55.88
H 7.02
7.10
8
12
13
C NMR (CDCl ): δ = 192.2, 166.2, 154.3, 117.7, 82.9, 68.5, 67.3,
3
–1
6
6.6, 51.5, 41.2, 25.7 and 25.7 (6 × C), 17.9 (2 × C), (–4.4)–(–4.7)
IR (KBr) υ = 1744 (C=O) cm .
1
(4 × C).
H NMR (D -DMSO) δ = 5.01 (t, J = 7.0 Hz, 1 H), 4.62 (d, J = 1.5 Hz,
6
1
H), 4.46 (t, J = 4.8 Hz, 1 H), 4.07 (d, J = 1.5 Hz, 1 H), 3.67 (m, 1
H), 3.51 (m, 1 H), 2.99 (m, 1 H), 2.46 (m, 2 H), 1.93 (m, 2 H), 1.81
(m, 1 H).
(
4aS,5S,7R,7aS)-Hexahydro-3-oxo-5,7-bis[(tert-butyldimethyl-
silyl)oxy]cyclopentapyran (73b):
To a stirred refluxed solution of bis-silyloxy thionocarbonate 81b
13
C NMR (D -DMSO) δ = 178.2, 85.0, 71.6, 58.0, 49.3, 41.7, 38.1,
6
(326 mg, 0.665 mmol) in anhyd toluene (46 mL) was added a solution
30.4.

SYNTHESIS
5
80
Papers
The spectral data of 104 (Corey lactone) was identical to that reported
in the literature. The compound was authenticated with a commer-
cial sample.
(6) Adiyaman, M.; Li, H.; Lawson, J. A.; Hwang, S.-W.; Khana-
pure, S. P.; FitzGerald, G. A.; Rokach, J. Tetrahedron Lett.
1997, 38, 3339.
3
5
(
(
(
7) Beckwith, A. L. J.; Schiesser, C. H. Tetrahedron 1985, 41,
The spectral data of 110:
1
3925.
H NMR (CDCl ) δ = 5.33 (s, 1 H), 4.89 (t, J = 6.0 Hz, 1 H), 4.58 (m,
3
8) O’Connor, D. E.; Mihelich, E. D.; Coleman, M. C. J. Am. Chem.
Soc. 1984, 106, 3577.
9) Corey, E. J.; Wang, Z. Tetrahedron Lett. 1994, 35, 539.
1
H), 3.91 (m, 1 H), 3.84 (m, 2 H), 2.84 (dd, J = 17.0 and 10.1 Hz, 1
H), 2.78 (dd, J = 10.3 and 6.0 Hz, 1 H), 2.68 (m, 1 H), 2.55 (d, J =
1
6.8 Hz, 1 H), 2.17 (dd, J = 15.8, 2.1 Hz, 1 H).
C NMR (DEPT 135) (CDCl ) δ = 177.7, 107.5, 81.0, 80.6, 73.7,
1
3
(10) Porter, N. A.; Zuraw, P. J.; Sullivan, J. A.; Gross, P. M. Tetra-
hedron Lett. 1984, 25, 807.
(11) Lynch, S. M.; Morrow, J. D.; Roberts II, L. J.; Frei, B. J. Clin.
Invest. 1994, 93, 998.
12) Gopaul, N. K.; Nourooz-Zadeh, J.; Mallet, A. I.; Anggard, E. E.
Biochem. Biophys. Res. Commun. 1994, 20, 338.
13) Pratico, D.; Juliano, L.; Spagnvoli, L.; Mauriello, A.; Maclouf,
J.; Violi, F.; FitzGerald, G. A. Circulation (Suppl) 1996, 94,
77.
14) Takahashi, K.; Badr, K. F. Seminars in Nephrology 1993, 13,
29.
15) Baberheem, M.; Kang, K. H.; Morrow, J. D.; Roberts II, L. J.;
Newman, J. H. In American Physiological Society; 1992; pp
H660–H663.
(16) Morrow, J. D.; Moore, K. P.; Awad, J. A.; Ravenscraft, M. D.;
Marini, G.; Badr, K. F.; Williams, R.; Roberts II, L. J. J. Lipid
Mediators 1993, 6, 417.
(17) Morrow, J. D.; Frei, B.; Longmire, A. W.; Gaziano, J. M.;
Lynch, S. N.; Shyr, Y.; Strauss, W. E.; Oates, J. A.; Roberts II,
L. J. N. Engl. J. Med. 1995, 332, 1198.
3
6
7.1, 44.5, 32.3, 31.4.
+
+
FABMS m/z 223.0 (M + Na) ; 207.0 (M + Li) .
The spectral data of 111:
(
1
H NMR (CDCl ) δ = 5.38 (s, 1 H), 4.49 (t, J = 3.0 Hz, 1 H), 4.09 (m,
3
1
2
H), 3.88 (m, 2 H), 3.71 (s, 3 H), 2.62 (dd, J = 16.3, 7.6 Hz, 1 H),
.48 (dd, J = 16.4, 6.2 Hz, 1 H), 2.34 (m, 1 H), 2.11–1.93 (m, 3 H).
C NMR (DEPT 135) (CDCl ) δ = 174.7, 103.4, 79.9, 74.7, 68.5,
(
1
3
3
2
6
0.8, 52.3, 46.0, 35.0, 34.7.
(
(
HREIMS m/z calcd for C H O (M–OH)+ 215.0915, found
1
0
16
6
1
+
2
15.0916, (M–OCH ) 201.0770.
3
+
FABMS m/z 255.1 (M + Na) .
Spectral data for 107:
1
H NMR (CDCl ): δ = 4.2 (dt, J = 5.7 Hz, 1H), 3.93 (m, 1H), 3.38 (m,
3
2
H), 3.2 (dd, J = 5.2 and 11.5 Hz, 1H), 2.75 (dd, J = 11.6 and 17.6 Hz,
H), 2.35 (dd, J = 8.2 and 17.7 Hz, 1H), 2.15 (m, 1H), 1.83 (dt, J =
.8 Hz, 1H), 1.58 (m, 1H), 1.3 (m, 1H).
1
3
13
C NMR (CDCl ): δ = 174.8, 170.9, 81.4, 79.1, 65.0, 47.0, 33.0,
3
3
2.8, 27.9.
+
LRFAB m/z (M + Na) 255.
(18) Reilly, M.; Delanty, N.; Lawson, J. A.; FitzGerald, G. A. Circu-
lation 1996, 94, 19.
19) Bachi, A.; Zuccato, E.; Baraldi, M.; Fanelli, R.; Chiabrando, C.
Spectral data for 112:
–
1
(
(
(
IR (film) υ = 3388 (br OH), 1730 (C=O), 1633 cm .
1
Free Radic. Biol. Med. 1996, 20, 619.
H NMR (CDCl /CD OD) δ = 6.88 (dd, J = 15.6, 4.6 Hz, 1 H), 6.0 (d,
3
3
20) Meagher, E. A.; Lucey, M.; Bensinger, S.; FitzGerald, G. A.
J. Invest. Med. 1996, 44, 223.
21) Takahashi, K.; Nammour, T. M.; Fukunaga, M.; Ebert, J.; Mor-
row, J. D.; Roberts II, L. J.; Hoover, R. L.; Badr, K. F. J. Clin.
Invest. 1992, 90, 136.
J = 15.6 Hz, 1 H), 4.46 (m, 1 H), 3.79 (m, 1 H), 3.67 (s, 3 H), 3.62 (d,
J = 4.7 Hz, 2 H), 3.44 (m, 1 H), 3.13 (s, 4 H), 1.8 (dt, J = 14.6 Hz, 1
H), 1.59 (dt, J = 14.5 Hz, 1 H).
1
3
C NMR (CDCl ) δ = 167.0, 150.3, 119.3, 74.1, 72.0, 67.0, 63.1,
3
5
1.6, 38.3.
(
22) Pratico, D.; Smyth, E. M.; Violi, F.; FitzGerald, G. A. J. Biol.
Chem. 1996, 271, 14916.
(23) Kunapuli, P.; Lawson, J. A.; Rokach, J.; FitzGerald, G. A. J. Bi-
ol. Chem. 1997, 272, 27147.
24) Pratico, D.; FitzGerald, G. A. J. Biol. Chem. 1996, 271, 8919.
25) Pratico, D.; Barry, O. P.; Lawson, J. A.; Adiyaman, M.; Hwang,
S. W.; Khanapure, S. P.; Iuliano, L.; Rokach, J.; FitzGerald, G.
A. Proc. Natl. Acad. Sci. USA 1998, in press.
We have recently proposed a new nomenclature system in which the
major change is the classification system. This nomenclature was
not ready at the time this article was written.
3
6
(
(
We wish to thank the NIH for support under Grant DK-44730, the
NSF for an AMX-360 NMR instrument (Grant CHE-90-13145) and
the Turkish Ministry of Education for the doctoral fellowship to M.A.
(
(
(
(
(
26) Barton, D. H. R.; McCombie, S. W. J. Chem. Soc., Perkin
Trans. 1 1975, 1574.
27) Barton, D. H. R.; Subramanian, R. J. Chem. Soc., Perkin Trans.
1 1977, 1718.
28) Rondot, B.; Durand, T.; Girard, J. P.; Rossi, J. C.; Schio, L.;
Khanapure, S. P.; Rokach, J. Tetrahedron Lett. 1993, 34, 8245.
29) Mulzer, J.; Kermanchahi, A. K.; Buschmann, J.; Luger, P.
Liebigs Ann. Chem. 1994, 531.
1
Current address: UMDNJ-Robert Wood Johnson Medical School,
6
75 Hoes Lane, Piscataway, NH 08854, USA
2
Current address: Roussel Uclaf, Centre de Recherche de Romain-
ville, 102 route de Noisy, 93230 Romainville, France
30) Williams, D. R.; Moore, J. L. Tetrahedron Lett. 1983, 24, 339.
(
1) Morrow, J. D.; Hill, K. E.; Burk, R. F.; Nammour, T. M.; Badr,
(31) Larock, R. C.; Lee, N. H. J. Am. Chem. Soc. 1991, 113, 7815.
(32) Vionnet, J.-P.; Renaud, P. Helv. Chim. Acta 1994, 77, 1781.
(33) Taber, D. F.; Herr, R. J.; Gleave, D. M. J. Org. Chem. 1997, 62,
194.
K. F.; Roberts II, L. J. Proc. Natl. Acad. Sci. USA 1990, 87,
9
383.
2) Morrow, J. D.; Harris, T. M.; Roberts II, L. J. Anal. Biochem.
990, 184, 1.
(
(
(
(
1
(34) Schmidt, O. T. Methods Carbohydr. Chem., II, 318. 3
(35) Pouchert, C. J.; Behnke, J. The Aldrich Library of 1 C and H
FT-NMR Spectra; Aldrich Chemical Co.: Milwaukee, 1992.
(36) Rokach, J.; Khanapure, S. P.; Hwang, S. W.; Adiyaman, M.;
Lawson, J. A.; FitzGerald, G. A. Prostaglandins 1987, in press.
1
3) Hwang, S. W.; Adiyaman, M.; Khanapure, S.; Schio, L.; Ro-
kach, J. J. Am. Chem. Soc. 1994, 116, 10829.
4) Adiyaman, M.; Lawson, J. A.; Hwang, S. W.; Khanapure, S. P.;
FitzGerald, G. A.; Rokach, J. Tetrahedron Lett. 1996, 37, 4849.
5) Hwang, S. W.; Adiyaman, M.; Khanapure, S. P.; Rokach, J.
Tetrahedron Lett. 1996, 37, 779.