вход по аккаунту


Olefin Alkylation in Biosynthesis.

код для вставкиСкачать
D E C E M B E R 1968
PAGES 903-964
Olefin Alkylation in Biosynthesis
Stereochemical studies, made possible by asymmetric labeling with hydrogen isotopes,
have led to the hypothesis that enzymic association of C5 units in polyisoprenoid synthesis is not a concerted process but proceeds in two steps: a trans 1,2-addition to an
olefin followed by a trans 1,2-elimination. It is shown that similar mechanisms can also
explain the cyclization of squalene epoxide to tetracyclic and pentacyclic triterpenoids.
1. Introduction
It is difficult, though perhaps not impossible, to
imagine a chemical basis for life in which connected
carbon atoms play no part. Recent experiments on the
genesis of “life-like” molecules from substances, such
as methane and hydrogen cyanide, which might be
presumed to have been abundant on the prebiotic
earth, have indicated several natural non-biotic
mechanisms for carbon-carbon bond formation. The
hypothesis underlying such experiments is that the
first life on earth was organized from existing, complex chemical species, and a necessary consequence
is that mechanisms for biosynthesis evolved before
the available supply was exhausted: mechanisms
for carbon-carbon bond formation along with the
In the biosynthetic reactions leading to proteins,
carbohydrates, nucleic acids, and most lipids, condensations of the aldol and Claisen type predominate:
an enolate anion, formed by removal of hydrogen (or
sometimes carboxyl) cc to a carbonyl group, attacks
another carbonyl group. The alkylation of enols, so
common in organic synthesis, is curiously rare in biosynthesis; it is probably true that the nucleophilic
carbonyl group and the electrophilic enolate ion
generate, by reaction with each other, most of the
carbon-carbon bonds found in nature.
[*I Professor J. W. Cornforth, F.R.S.,
“Shell” Research Limited,
Milstead Laboratory of Chemical Enzymology,
Kent (England)
[I] An excellent review of terpenoid biosynthesis is that by R . B.
Cluyton, Quart. Rev. (chem. SOC. London) 19, 168, 201 (1965).
Angew. Chem. internat. Edit.
Vol. 7 (1968) 1 No. 12
This is one reason why the enzymes of terpenoid biosynthesis ~1 are of especial interest: although (R)mevalonic acid ( I ) , the starting-point for all terpenoid
biosynthesis, is formed by Claisen condensation of
three molecules of acetyl-coenzyme A, neither the aldol
nor the Claisen condensation is of any further importance in forming the many additional carboncarbon bonds of the terpenoids. Instead, we have
a type of reaction which can be classed as olefin alkylation.
The “leaving” group for the first alkylation [(4) +
(7)] is pyrophosphate ion. Mevalonic acid 5pyrophosphate ( 2 ) is produced by two successive
phosphoryl transfers to (R)-mevalonic acid ( I ) from
adenosine triphosphate. The double bond required
for further alkylation is then generated by a concerted
elimination reaction, also requiring adenosine triphosphate; the four products obtained in this step
are inorganic phosphate, adenosine diphosphate, carbon dioxide, and isopentenyl pyrophosphate (3). The
leaving group is then activated further by a prototropic
exchange which produces dimethylallyl pyrophosphate ( 4 ) . The increased electrophilic activity of the
allylic pyrophosphate (4), compared with the homoallylic isomer ( 3 ) , is seen for example in its sensitivity
to aqueous acid. Below pH 5 pyrophosphate ion is
rapidly eliminated and water is alkylated; the major
product is the alcohol (5) presumably formed from
the intermediate allylic carbonium ion (6). The sequence of enzymic reactions leading from ( I ) to (4)
was first indicated in soluble preparations from yeast,
and then from liver. The four enzymes responsible
have been extensively purified. Each of them can be
separated from the other three, though none of them
has yet been obtained completely pure.
Combination of the intermediates (3) and (4) is the
first and characteristic step in the biosynthesis of polyisoprenoids. In this enzymic reaction, a molecule of
dimethylallyl pyrophosphate and one of isopentenyl
pyrophosphate react, with loss of a proton and of a
pyrophosphate anion, to form a new allylic pyrophosphate. In the soluble liver and yeast preparations
which so far have been the most extensively studied,
the product is geranyl pyrophosphate (7). This allylic
pyrophosphate can then react, apparently on the
same enzyme, with another molecule of isopentenyl
pyrophosphate ( 3 ) to produce the Cis-derivative
farnesyl pyrophosphate (8). Further chain-lengthening to geranylgeranyl pyrophosphate does not proceed
at a measurable rate with this enzyme system.
HO~C, ,c<, ;CH,OH
The formation of (7) can be interpreted as proceeding
by way of two carbonium-ion intermediates, thus:
+ R+
+ OPP-
The isomerization of isopentenyl pyrophosphate to
dimethylallyl pyrophosphate can also be interpreted
by equations (2) and (3), if H+ replaces R+.
To know what actually happens on the enzyme, more
information is needed than is ascertainable by present
methods. Some insight can be gained, however, by
studying the stereochemistry of the condensation; and
during the last few years Popjak and the writer, at
Milstead Laboratory, have examined the stereochemistry of polyisoprenoid, and particularly of squalene,
Cp3 .OH
HO~C, /c< /CH~OP~O;@
2. Stereochemical Considerations
It is not immediately apparent that anything can be
learned in this way, for the optical asymmetry present
in mevalonic acid and its phosphates is lost in isopentenyl and dimethylallyl pyrophosphate. However, in isopentenyl pyrophosphate there are three
methylene groups. Assuming that it is possible to
distinguish between the two hydrogen atoms on each
methylene group, as in ( 9 ) , and also to trace the fate
of each hydrogen atom in the products of the enzymic
reactions leading, for example, to farnesyi pyrophosphate (IO), then several conclusions are possible
concerning what happens on the enzymes. Let us
suppose that in the stereochemical sequence rules [21
H, ranks higher than Hb and He ranks higher than
HF. Then
1) If the hydrogen-ion eliminations in the isomerization and in the two coupling reactions are stereospecific, positions 2, 6, and 10 in (10) will be occupied
by H, or Hd according to the stereochemistry. Nonstereospecific reactions would lead to mixtures of H,
and Hd at these positions.
2) Displacement of pyrophosphate anion is necessary
to form the carbon-to-carbon bonds between positions 4-5 and 8-9. If these displacements occur with
inversion of configuration the absolute configurations
at C-5 and C-9 in ( I O ) will be ( S ) ; if retention of configuration is the rule, then these absolute configurations will be ( R ) . If these displacements are nonstereospecific (as carbonium-ion reactions often are),
then (RS)-configurations will result at C-5 and C-9.
3) The two sides (above and below the plane of the
projection ( 9 ) ) of the double bond in isopentenyl
pyrophosphate are not equivalent in an asymmetric
environment. If the new carbon-carbon bond is formed
from below the plane of this double bond, the absolute
configurations at C-4 and C-8 in (10) will be (S); if it
R. S . Cahn, C. K . Ingold, and V. Prelog, Experientia I 2 , 81
Angew. Chem. internat. Edit. J Vol. 7 (1968)
No. 12
is formed from above that plane, the absolute configurations will be (R).
These three potentially stereospecific processes define
the stereochemical behavior of all atoms in the substrates which undergo permanent changes in bonding
as a result of the coupling reaction. The isornerization
is not so completely defined, for the transformation of
isopentenyl pyrophosphate (3) into dimethylallyl
pyrophosphate ( 4 ) requires the uptake of a hydrogen
ion. The terminal methylene group in (3) can take
up the ion from above or below the plane of the
double bond, but the resulting pattern of substitution (e.g. the C-12 methyl group in (10)) has no
chirality unless the entering hydrogen can be distinguished from both Ha and Hb. Since the only practical
way of distinguishing between the two hydrogen atoms
in each pair (Hab, Hcd, Her) is to label one of them
with a hydrogen isotope, it follows that chirality could
only be introduced at the methyl group of dimethylally1 pyrophosphate by using all three known isotopes of hydrogen, and the stereochemistry of hydrogen addition could only be solved by devising some
method of recognizing this chirality.
2.1. Preparation of Isotopically Labeled Compounds
Experimentally, it is not convenient to label isopentenyl
pyrophosphate (3) directly with hydrogen isotopes.
Its precursor mevalonic acid ( I ) is a better subject,
for the principal reason that mevalonic acid already
possesses a center of asymmetry at C-3, and that only
(R)-mevalonic acid is accepted as a substrate by the
enzyme that phosphorylates it [3,41.
If a hydrogen atom at the 4-or 5-position of mevalonic acid
is replaced stereospecifically by deuterium or tritium, this configuration will persist at the corresponding position (2 and 1
respectively) of the isopentenyl pyrophosphate ( 3 ) formed
from it enzymically, since at neither of these carbon atoms is
the bonding altered. If, however, mevalonic acid is labeled
stereospecifically at the 2-position with isotopic hydrogen,
the relationship between the chirality at C-2 of mevalonate
with the geometry at C-4 in isopentenyl pyrophosphate depends on the mechanism of the elimination reaction (2) +
( 3 ) . Therefore, it is necessary to determine the orientation
of isotopic hydrogen at C-4 of the isopentenyl pyrophosphate - it is not sufficient, as it is with C-1 and C-2, t o know
the configuration of the parent mevalonate.
Introduction of labeled hydrogen (both deuterium
and tritium) into C-5 of mevalonic acid was done
enzymically 151. An enzyme extractable from pig liver,
mevaldate reductase 161, transfers hydrogen from
NAD2H and NAD3H to mevaldic acid ( I I ) , the aldehyde of mevalonic acid; NADH itself can easily
be labeled with deuterium or tritium. Curiously, the
enzyme reduced both the (R)-and the (S)-forms of
- - - _
[3] F. Lynen and M. Grass/, Hoppe-Seylers 2. physiol. Chem.
313, 291 (1958).
141 R. H. Cornforth, J . W . Cornforth, and G . Popjak, Tetrahedron 18, 1351 (1962).
[5] C. Donninger and G. Popjak, Proc. Roy. SOC. (London),
Ser. B 163, 465 (1966).
[6] M . J . Schlesinger and M . J . Coon, J. biol. Chemistry 236,
2421 (1961).
Angew. Chem. internat.
Edit. J Val. 7 (1968) No. 12
mevaldic acid, but this lack of stereospecificity is at Li
site remote from the center of action. Tritium-labeled
mevalonic acid, prepared in this way, was converted by
a preparation of soluble enzymes from rat liver into
farnesyl pyrophosphate, which could be hydrolyzed
enzymically (alkaline phosphatase) to farnesol. When
this labeled farnesol (12) was oxidized to farnesal (13)
by NAD+ in the presence of liver alcohol dehydrogenase, one-third of the tritium, i.e. all the tritium at
C-1, was lost.
The orientation of the hydrogen removed from primal y
alcohols by liver alcohol dehydrogenase has been determined, for ethanol, as pro-R. If the stereospecificity of alcohol dehydrogenase is the same for farnesol
as for ethanol (as would be expected) then the isotopic hydrogen introduced into mevalonate by mevaldate reductase should also produce this (5R) configuration. Confirmation of this was obtained 171
from experiments with geraniol as a model for farnesol,
showing that liver alcohol dehydrogenase removes the
pro-R hydrogen from geraniol; later, a direct correlation of the specificities for geraniol and for ethanol
was obtained [*I by chemical degradation of geraniol,
labeled with tritium by reduction of geranial with
NAD3H in the presence of liver alcohol dehydrogenase, to ethanol containing the same primary hydroxyl group. On liver (or yeast) alcohol dehydrogenase this ethanol lost all its tritium.
For the stereospecific labeling at the 4- and 2-positions of mevalonic acid, chemical methods were used.
The acid (14) and the lactone (15) 181 were first converted into the geometrically isomeric benzhydrylamides (16) and (17). Each of these was epoxidized
and subsequently reduced by LiBD4 to give, after alkaline hydrolysis, two specimens of mevalonic acid
labeled with deuterium at position 4.
Because of the method of synthesis, these two specimens of mevalonic acid must be composed of different molecular species. The epoxidation of a double
bond by a peroxy-acid is a cis addition of oxygen; the
[7] C.Donninger and G . Ryback, Biochem. J. 91, 10 P (1964).
[*I D . Arigoni and H. Weber, private communication.
181 J . W . Cornforth, R . H . Cornforth, C. Popjak, and I . Y. Gore,
Biochem. J. 69, 146 (1958).
H O z C P O H
reduction of an epoxide by a metal hydride produces
inversion of configuration at the center from which
oxygen is displaced. Hence the benzhydrylamide (16)
produced a racemic mixture of the enantiomeric epoxides (IBa) and (18b), and from this by reduction
and hydrolysis the racemic mixture (2Oa) and (20b)
was formed; whereas the benzhydrylamide (17) gave
the racemic mixture (21a) and (21b) via the racemic
epoxide (19a) and (196). Further, because mevalonic
kinase phosphorylates only the (3R) form of mevalonic
acid, the racemic mixtures (20) and (21) will behave
enzymicallyas though each consisted exclusively of one
molecular species (20a) and (21a). When the sequence of enzymic reactions has proceeded as far as
isopentenyl pyrophosphate, the asymmetry at C-3 of
mevalonate is destroyed but the asymmetric labeling
persists, (20a) giving exclusively the (2s) enantiomer
(22) and (21a) giving exclusively the (2R) (23)Egl.
H 0 z C G o H
has become C-2 and a (3R)-mevalonic acid has become (3s). Thus the enzymically inactive component
(206) gives enzymically active (3R) [2R-2-Dl]mevalonic
acid (24). Similar treatment of the racemic mixture
(21) will give (25) as the sole enzymically active
molecular species.
The necessary inversion of functional groups was effected 1101
in three stages: 1) C-4 labeled mevalonates were esterified by
reaction of methyl iodide with the silver salts; 2) the methyl
mevalonates were oxidized by zinc permanganate in cold
acetone to methyl hydrogen 3-hydroxy-3-methylglutarates;
3) reduction of the lithium salts with lithium borohydride
then gave C-2 labeled mevalonates.
The diastereoisomeric mevalonic acids (20) and (21)
were also useful in solving the problem of labeling
position 2 of mevalonic acid asymmetrically with
hydrogen. For if the carboxyl and hydroxymethyl
functions of mevalonic acid are exchanged, the product is stiIl mevalonic acid with the difference that C-4
The enzymic reaction leading from mevalonic acid
5-pyrophosphate (2) to isopentenyl pyrophosphate
( 3 ) is a trans elimination of water and carbon dioxide; so that (3R)[2R-2-Dl]mevalonate generates
cis- [4-D1]- isopentenyl pyrophosphate (26) and
(3R)[2S-2-Dl]mevalonate generates tuans-[4-D1]-isopentenyl pyrophosphate (27). However, this was not
known a priovi and the geometry of the two deuterioisopentenyl pyrophosphates, prepared enzymically
from the two mevalonates, had to be ascertained [lo].
This was done by enzymic hydrolysis of both compounds to isopentenol (26a) and (27a), addition of
[9] J. W. Cornforth, R. H. Cornforth, C. Donninger, and G.
Popjrik, Proc. Roy. SOC. (London), Ser. B 163,492 (1966).
[ l o ] J. W. Cornforth, R . H . Cornforth, G . Popjak, and L. S.
Yengoyan, J. biol. Chemistry 241, 3970 (1966).
Angew. Chem. infernat. Edit.
/ VoI. 7 (1968) / No. I2
' ' S O H
(+ Enantiomer)
bromine to the double bond, and elimination of hydrogen bromide by methanolic alkali in conditions
favoring the Ez reaction. The main product is a mixture of trans-4-bromoisopentenol with a lesser amount
of the cis isomer. Since addition of bromine to a double
bond is specifically trans, and the E2 elimination preferentially trans, it follows that the deuterium in cis[4-Dl]isopentenol will be retained in the cis-4-bromoisopentenol and absent from the trans; the converse will
be true with trans-[4-Dl]isopentenol. By separation
of the geometrically isomeric bromoisopentenols (by
gas-liquid chromatography) and examination of each
isomer for deuterium (by mass spectrometry) it could
be shown that deuterium was retained in the cis-4bromoisopentenol (28) derived, ultimately, from
(3 R)[ZR-Z-Dl]mevalonate (24) and in the trans-4bromoisopentenol (29) derived from (3 R)[ZS-Z-D11mevalonate (25). From the other two bromoisopentenols, the deuterium was largely lost. Thus the enzymic formation of isopentenyl pyrophosphate was
established as a trans elimination, and the stereochemical correlation between C-2 labeled mevalonates and
C-4 labeled isopentenyl pyrophosphates was secured.
(+ Enantiomer)
ates; the 4R hydrogen was completely retained and
all of the 4 s hydrogen was lost.
When mevalonates labeled at the 2- and 5-positions
were the precursors, only deuterium could be used as
the labeling isotope. Farnesol or squalene derived enzymically from these labeled precursors was degraded
by ozonolysis; the levulinic acid, which contained the
potentially asymmetric centers to be investigated, was
oxidized by alkaline sodium hypoiodite to give as main
products iodoform and succinic acid.
We had already prepared, for a different purpose, asymmetrically labeled monodeuteriosuccinic acid of known absolute configuration, and had shown that its optical activity
in the ultraviolet is easily measurable by a modern spectropolarimeter; [or]250nm is -18 O, so that a few milligrams of
pure succinic acid suffice to determine the direction and extent of asymmetric labeling. Squalene made biosynthetically
3. Formation of Polyisoprenoids from C, Units
The stereochemistry of the alkylation step leading to
polyisoprenoids was defined by tracing the fate of
deuterium or tritium when these stereospecifically
labeled mevalonates were used as substrates for cellfree preparations of rat liver. Sometimes farnesyl
pyrophosphate was chosen as the end-product for
examination; sometimes squalene.
When farnesyl pyrophosphate is synthesized enzyrnically from three mevalonate units, three hydrogen
atoms are eliminated which were originally attached to
C-4 of mevalonate. By mass-spectrometric examination of farnesol derived by enzymic hydrolysis from
biosynthetic farnesyl pyrophosphate, it was found 191
that when (3 R)[4R-4-Dl]mevalonate was the precursor
all three deuterium atoms were retained, but that
deuterium-free farnesol was formed when (3 R)[4S-4Dllmevalonate was the precursor. This result was confirmed by examination of squalene, biosynthesized
from either C-4 deuteriated or C-4 tritiated mevalonAngew. Chem. internat. Edit.
1 Vol. 7 (1968)1 No. 12
from (3R)[5R-5-Dl]mevalonate[91, and farnesol from
(3R)[2R-2-Dl]mevalonate [lo], were degraded to succinic
acid as described. Each of these succinic acids was found to
consist essentially of [2R-2-Dl]succinic acid (301, the optical
activity corresponding to the deuterium content of the
specimen, as measured by mass spectrometry of the methyl
ester or the anhydrideclll.
Assume that a molecule of isopentenyl pyrophosphate,
which is to be alkylated by a molecule of a primary
allylic pyrophosphate (SI), is oriented on the enzyme
as represented in (32). The two carbon atoms of the
double bond and the four atoms attached to them a r e
in the plane of the paper. Then our results show (i)
that complete inversion of configuration occurs a t
the primary allylic carbon atom (C-2) when pyrophosphate ion is displaced and the new C-C bond is
formed, (ii) that this bond is formed from below the
plane of the double bond in isopentenyl pyrophosphate, and (iii) that the 2R hydrogen atom of the isopentenyl pyrophosphate is eliminated to form the new,
trans, double bond.
[ l l ] G . Popjak and J . W. Cornforth,Biochem. J. 101, 553 (1966).
From (i) the conclusion may be drawn that the condensation cannot be initiated by formation of a carbonium ion from the allylic pyrophosphate, unless
(a) the carbonium ion is spatially so restricted that
rotation about the C-CH; bond is forbidden, or
(b) formation of the new C-C bond occurs so rapidly
that there is insufficient time for this rotation to occur. The simplest explanation, requiring no such
special hypotheses, is that the new bond is formed as
the old bond breaks: a reaction of S N type
~ where inversion of configuration is normal. This makes the
C-C bond formation a concerted process.
Is it possible to extend this concept and to represent
the whole condensation as a concerted process, a
continuous drift of electrons starting from the C-HR
bond at position 2 of isopentenyl pyrophosphate and
ending with the departure of pyrophosphate anion?
If the stereochemical requirements (ii) and (iii) are
considered together, this appears unlikely. For since a
trurzs double bond is being generated in the product,
the entire isopentenyl pyrophosphate molecule must be
oriented for a concerted reaction, as shown, with HR
on C-2 (that is, Ha) below the plane of the paper.
Since the electrons of the C-HR bond are being used
to form the new double bond, C-3 of the isopentenyl
pyrophosphate is having electrons supplied to it (to
form the new double bond) and withdrawn from it
(to form the C-C bond) on the same side; effectively,
a nucleophilic substitution with retention of configuration. This seems energetically far less favorable than a
concerted mechanism in which electrons are supplied
to, and withdrawn from, opposite sides of C-3; but in
the present case this would lead either to a cis double
bond or to the elimination of Hs: both contrary to experiment. Indeed, when rubber (which has cis double
bonds) is biosynthesized, it is Hs which is eliminated,
not HR.
These considerations have led us to postulate a twostage mechanism for the alkylation, in which an
electron-donating group X is supposed to participate.
This group might be part of the enzyme, but it might
also be an enzyme-bound water molecule, or an oxygen atom of the pyrophosphate group in isopentenyl
pyrophosphate (a hypothesis put forward by Johnson
before the stereochemical details were known).
The first stage is then a trans-addition of the allylic
group and of X to the double bond of isopentenyl
pyrophosphate; the second stage is a trans-elimination of X and HR from the intermediate (33) to
generate the new double bond. If the group X has a
formal negative charge, its binding to C-3 might not at
any stage be more than that in a close ion-pair.
squalene could lead without formation of stable intermediates t o all known types of tetracyclic and pentacyclic triterpenes by a series of stereospecific steps. The
cyclizations, which are all classifiable as olefin alkylations, were assumed to proceed by "antiparallel"
pf CH3
(trans) addition, to a double bond, of a carbonium ion
and an electron-donating group (in this case the xelectrons of another double bond).
In order t o derive the observed stereochemical arrangement
of the final product, it was always necessary t o postulate
one or more stereospecific intramolecular rearrangements of
non-classical carbonium ions. Sometimes, two successive
stereospecific rearrangements of the same complex cation
had to be assumed.
An analogous mechanism could account for the observed
stereochemistry of alkylation of isopentenyl pyrophosphate.
According t o this, the first step would be the formation of the
non-classical carbonium ion ( 3 4 ) ; next, this would undergo
intramolecular rearrangement to another non-classical ion
(35) which would allow the now stereochemically favorable
elimination of HR t o give the final product (36). It is a logical disadvantage of this mechanism that formation of the
non-classical carbonium ion from the other side of the double
This mechanism accounts satisfactorily for the observed stereochemistry of the condensation. However,
an alternative explanation is possible.
In a classical publication 1131, Eschenmoser, Ruzicka,
Jeger, and Arigoni showed how the cyclization of trans[12] W. S. Johnson and R. A . Bell, Tetrahedron Letters No. 12,
27 (1960).
[13] A . Esclienmoser, L. Ruzicka, 0. Jeger, and D . Arigoni, Helv.
chim. Acta 38, 1890 (1955).
Angew. Chem. internat. Edit. J Vol. 7 (1968) J NO. I2
bond [as in (37)] would allow a mechanistically similar formation of ( 3 6 ) without need to postulate an intermediate rearrangement.
One might regard the mechanism (34) + (36) as
equivalent to the “X-group” mechanism in which the
electrons of a C-H bond act as an X-group. The essential difference between this and the true “X-group”
mechanism is spatial. The rearrangement of (34) to
(35) entails a large displacement in space of the allylic
group R (or alternatively of the whole isopentenyl
pyrophosphate molecule). This displacement must occur (since no stable intermediate is permitted) at the
“active center” of the condensing enzyme, to which
the substrates have already been bound for the initial
stage of the reaction. When it is also considered that
the group R may be large, and that its allylic double
bond leaves it little flexibility at the point of attachment to isopentenyl pyrophosphate, the difficulty of
postulating intramolecular rearrangement of (34) to
(35) becomes obvious. This can be regarded as an
argument in favor of the X-group mechanism, since
the capacity of the active center of an enzyme to bind
a specific substrate in a specific conformation favorable for reaction should be greater, the less movement of the substrate is necessary during the enzymic
reaction. In the X-group mechanism, little movement
is necessary.
4.1. Formation of Lanosterol from Squalene
More evidence is available on the transformation Qf
squalene to lanosterol than on the formation of any
other cyclic terpenoid. This is not a “non-stop” reaction, since squalene epoxide (38) appears to be an
obligatory intermediate r14-151. The classical theme of
cyclization, adapted to start from this intermediate, is
as shown (Scheme 1).
Scheme 1.
4. Cyclization of Squalene to Triterpenoids
It is instructive to examine whether the cyclization of
squalene to tetracyclic and pentacyclic triterpenoids
can be formulated with X-group mechanisms in place
of the original concept of stereospecific intramolecular
rearrangements of carbonium ions. This turns out to
be possible. The first three postulates of the original
scheme can be retained: that all-trans squalene is the
starting-point, that the substrate molecules assume definite conformations for cyclization, and that (to
paraphrase the third requirement) whenever electrons
are abstracted from a particular substituted carbon
atom, the deficiency is always made good from the
opposite side. The fourth postulate, that no stable
intermediate intervenes between squalene and the
final product, is not acceptable for X-group mechanisms. It was a n intellectual t o u r de force to show that
every known type of triterpenoid could arise from
squalene by a “non-stop” mechanism, but this view
tends to require incompatible properties in the enzyme mediating a cyclization.
Since no stable intermediate is permitted, the whole process
has to be supposed to occur on a single enzyme. This enzyme
must be able, at the same active center, to impose a particular
conformation on the squalene chain and t o allow considerable freedom of movement in this chain during rearrangements; and to control stereochemically the intramolecular,
space-demanding rearrangements of carbonium ions without
participating in them chemically. If, in contrast, the formation of stable intermediates allows the participation of more
than one enzyme. the spatial requirements of the active
center of each enzyme can be less exacting.
Angew. Chern. internnt. Edit. 1 VoI. 7 (1968) No. 12
We can consider whether it is useful to postulate a n
X-group mechanism for any of these ring-closures and
subsequent rearrangements.
Formation of rings A and B: Here, the only way to
involve an external X-group is to postulate that the
initial cyclization leads to a cyclopentane ring [(39),
exemplified for ring A], and that rearrangement to a
cyclohexyl cation precedes further cyclization.
There seems to be no reason here why rings A and B
should not be closed by a concerted process, the xelectrons of the next double bond playing the part of
the X-group, and six-membered rings being formed ab
initio. Moreover, the alternative is an anti-Markovnikov addition to the double bond and is to this
extent energetically unfavorable.
Formation of ring C: Here, the intervention of an Xgroup does lead to a cyclization in the Markovnikov
sense, and formation of the intermediate (40) would
be favored (though rearrangement to the energetically
1141 E . J . Corey, W. E. Russey, and P. R. 0. de Montellano, J.
Amer. chem. Sac. 88, 4750 (1966).
[151 E. E. van Tamelen, J . D. Willetr, R . B. Clayton, and K. E.
Lord, J. Arner. chem. Sac. 88, 4752 (1966).
less favored cyclohexyl cation must still occur if the
cyclization is to continue). Secondly, it has been shown
that non-enzymic cyclization o f squalene epoxide does
yield as one product a cyclopentanonaphthalene
derivative (41), an indication that this mode of cyclization is energetically favorable [161. Thirdly, the dihydrosqualene epoxide (42) can be partially cyclized
enzymically t o a mixture of products, a major component of which is probably the alcohol (44)[171.
This of course could arise from an intermediate (40)
by simple elimination of HX.
possible, is not on the evidence a necessary postulate.
Formation of ring D: This is a cyclization in the
Markovnikov sense and there is no incentive to suppose that a six-membered ring might be formed first
and rearrange to a five-membered ring. (The objections to the sequential rearrangement of nonclassical cations, as formulated in Scheme 1, have
already been stated.) The alternative is to postulate an
electron-donating group X which completes the addition. Accepting that the other cyclizations are concerted, one then has a multiplex electron-shift leading
from squalene epoxide (38) to the intermediate (46).
Little displacement of any part of the molecule is
required for this cyclization.
T o decide whether this finding is evidence for an intermediate such as (40), we can consider the “normal”
mechanism in which a six-membered ring C is formed
directly by a continuation of the concerted cyclization that closed rings A and B. This is an anti-Markovnikov cyclization, but on the enzyme it might be
promoted by the stereochemical arrangement of the
substrate, the x-electrons of the next double bond
being available to complete the addition only in the
anti-Markovnikov sense. If these x-electrons were not
available, as in the dihydrosqualene epoxide (42), the
cyclization may be assumed to proceed as far as the
cation (43). In the absence of an electron-donating
group this cation would tend to rearrange to the more
stable tertiary form (45). Loss of a proton from this
cation could well give mainly the olefin (44) since one
hydrogen atom in the methyl group would always be
in a stereochemically favorable position for the final
elimination. One has to conclude that the intervention
of an X-group in the closure of ring C , though quite
[I 61 E. E. van Tamelen, J. D. Willett, M. Schwarzz, and R. Nadeau,
J. Amer. chem. SOC.88, 5937 (1966).
[I71 E. E. van Tamelen, K. B. Sharpless, R . Hanzlik, R . B. Clay!on, A . L. Burlingame, and P. C . Wszolek, J. Amer. chem. SOC.
89, 7151 (1967).
The rearrangement to lanosterol can proceed directly
from the final non-classical cation pictured in Schemel,
the four migrating groups and the proton finally to be
eliminated already having the favorable trans-antitrans-anti-trans arrangement. For analogous rearrangement of the intermediate (46), a preliminary
rotation through 120” is necessary about the single
bond joining the side-chain to ring D. This gives the
conformation (48) and might occur in one of three
1) Rotation is confined to the side-chain and occurs
without detachment of (46) from the enzyme on which
it was formed. For example, the group X might be a
water molecule attached to an anionic site on the
enzyme; after cyclization, the resulting -OH group
might swing to a cationic site where its separation,
initiating rearrangement, would be facilitated. This
mechanism would also require 120 movement of the
rest of the side-chain; but this is a part of the molecule
which does not need to be bound at the active center,
Angew. Chem. internat. Edit. Vol. 7 (1968)
No. I 2
and which apparently can be altered considerably
without affecting the cyclization 1181.
2) The intermediate (46) is held to the enzyme by
its X-group while the rest of the molecule is reoriented,
perhaps on an adjacent part of the enzyme, in the
configuration favorable to rearrangement.
3) The intermediate (46) dissociates from the enzyme,
and rearrangement with elimination of HX and formation of lanosterol (47) occurs on another enzyme.
Also in the case of pentacyclic triterpenoids, X-group
mechanisms can explain the various structures and
stereochemistry, but it is necessary t o assume the intervention of two X-groups, and also to postulate
double 1,2-rearrangements of the following type:
Several examples of this type of rearrangement are
already known; e.g., the rearrangement of Sq6P-dibromocholestane and of other vicinally disubstituted
steroids 1191.
The postulated course of the cyclization can be represented as in Scheme 2 for the case of p-amyrin
(49). It is clear that this process could not be considered to take place on a single enzyme, and therefore
that if the X-group mechanism is correct at least one
intermediate between squalene epoxide and P-amyrin
remains to be found. Recognition of such an intermediate will also decide between the “non-stop”
I181 E. E . van Tamelen, K . B. Sharpless, J. D . Willett, R. B. Clayton, and A. L. Burlingame, J. Amer. chem. SOC. 89, 3920 (1967).
I191 D . H. R . Barton and J. F. King, J. chem. SOC.(London)
Angew. Chem. internat. Edit. 1 Vol. 7 (1968) / No. I 2
Scheme 2.
mechanism originally proposed and the stepwise
mechanism to which our stereochemical studies have
led us.
Received: May 20, 1968
[A 669 IE]
German version: Angew. Chern. 80, 917 (1968)
91 1
Без категории
Размер файла
801 Кб
alkylation, olefin, biosynthesis
Пожаловаться на содержимое документа