Sulphur participation in cyclization reactions

(1)

Sulphur participation in cyclization reactions

Citation for published version (APA):

Loos, de, W. A. J. (1978). Sulphur participation in cyclization reactions. Technische Hogeschool Eindhoven. https://doi.org/10.6100/IR147991

DOI:

10.6100/IR147991

Document status and date: Published: 01/01/1978 Document Version:

Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers) Please check the document version of this publication:

• A submitted manuscript is the version of the article upon submission and before peer-review. There can be important differences between the submitted version and the official published version of record. People interested in the research are advised to contact the author for the final version of the publication, or visit the DOI to the publisher's website.

• The final author version and the galley proof are versions of the publication after peer review.

• The final published version features the final layout of the paper including the volume, issue and page numbers.

Link to publication

General rights

Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain

• You may freely distribute the URL identifying the publication in the public portal.

If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, please follow below link for the End User Agreement:

www.tue.nl/taverne

Take down policy

If you believe that this document breaches copyright please contact us at:

openaccess@tue.nl

(2)

(3)

SULPHUR PARTICIPATION IN

CYCLIZATION REACTIONS

PROEFSCHRIFT

TER VERKRIJGING VAN DE GRAAD VAN DOCTOR IN DE TECHNISCHE WETENSCHAPPEN AAN DE TECHNISCHE HOGESCHOOL EINDHOVEN,OP GEZAG VAN DE RECTOR MAGNIFICUS, PROF.DR. P.VAN DER LEEDEN, VOOR EEN COMMISSIE AANGEWEZEN DOOR HET COLLEGE VAN DEKANEN IN. HET OPENBAAR TE VERDEDIGEN OP

DINSDAG 12 DECEMBER 1978TE 16.00 UUR

DOOR

WAL THERUS ADRIANUS JOSEPH DE LOOS

(4)

DIT PROEFSCHRIFT IS GOEDGEKEURD DOOR DE PROMOTOREN

PROF. DR. H.M. BUCK EN

(5)

Aan mijn ouders Aan Marlies

(6)

Chapter I

Chapter II

CHAPTER I

General introduetion

With the increased knowledge of the biosynthesis of natural products the mechanistic aspects of biochemica! re-actions have received considerable interest of organic chemists. For a part this is also the result of a changed view about enzyme-substrate interactions. It is now recog-nized that the reaction catalyzed by an enzyme occurs at definite and relatively small parts of the generally large proteine molecule, resulting in specific interactions which are responsible for the stereochemical control of the re-action. In this way the action of the enzyme shows some resemblance with the action of an intramolecularly parti-cipating group.

An important class of natura! products are the steroids. They are built up in an early stag.e from acetic acid in a pathway which proceeds amongst others, via squalene

Cl)

and lanosterol

Ci),

in this sequence1

• The biochemica! route

from

l

to

i

is of interest, because here the typical steroid skeleton is formed, as elucidated by Corey and van Tamelen. First a terminal double bond of

l

is oxidized to yield the stable intermediate (38)-2,3-epoxysqualene (~). which con-tains one asymmetrie center. The opening of the epoxide ring and ensuing cyclization lead to the free or transient-lY stabilized carbenium ion ~. which undergoes a number of

[1,2] hydrogen and [1,2] methyl shifts and ultimately proton loss to yield lanosterol.

To what extent these reactions are regulated by enzymes can be deduced from the various simulation experiments. The

(9)

....

2

H

HO HO

4 3

selective oxidation of a terminal double bond can be

achieved in a polar solvent. Furthermore, the hydragen and methyl shifts praeeed suprafacially and with inversion of contiguration at the carbons of the steroid skeleton as in non-enzymic processes. Control is only needed at the

beginning and at the end. The cyclization could only be simulated in part. Cyclization of (E)-(E)-farnesylacetate

terminal epoxide (~) gives the bicyclic diol monoacetates

&

and

z,

the latter features the trans~syn relationship (hydrogen, methyl, hydrogen) characteristic for the

inter-mediate

l·

In this process the enzyme controls one of the

two alternatives. However, extension to larger systems gives rise to the formation of compounds which have a five membered C ring. Obviously enzymic participation is import-ant in this part of the process.

(10)

...

+ HO 5 6 /OAc ' I

HO~

H 7

The enzymatic conversion of 2 into 4 is an example of a polyene cyclization. This kind of reaction is increasing-ly applied in the synthesis of steraids and other poincreasing-ly- poly-cyclic compounds2 • The general pattern of this process is that somewhere in the molecule a carbenium ion is generated or an electron deficiency at carbon is induced leading to a ring formation by the intramolecular reaction with a double bond, resulting in a new carbenium ion. If a secend olefinic group is located in a position which allows further cyclization the process may go on. A point of discussion has remained whether the cyclization proceeds step by step with intermediacy of carbenium ions or in a concerted fashion. Both mechanisms have found experimental support 3

• 4•

It seems likely that the kind of mechanism depends on the relative stability of the ions befere and after the ring closure. For example, a cyclization initiated by a cyclo-pentenyl ion which should lead to a secondary carbenium ion has been shown to preeeed concertedly with the formation of a second ring. Gorvers et aZ. calculated the heats of form-ation of the two ions and showed that the cyclopentenyl ion occurring befare cyclization has a lower energy than the ring closed secondary ion5_• _{Furthermore it has been found} that in the solvolyses of a series of the chlorides

!

the alkene group participates in the ionization with aryl groups of less stahilizing ability (X is p-Br or m-Br), while

better stahilizing groups give rise to a reaction without participation6_• _{For camparing these results with the} me-chanistic features of a polyene cyclization, it must be born in mind that formation of ion-pairs may influence the

(11)

x

8 x p-OCH3 p-CH3 H p -Br m-Br percentage ossisted rout.e 9 55-70 94 95 98

reactions of carbenium ions. Hence ions formed by bond breaking and those occurring in a sequence of ring

format-ions may show different behaviour7_•

Several studies have been undertaken to influence the stereochemical outcome of a cyclization reaction. Substi-tuents introduced on the alkenyl chain may cause a high degree of asymmetrie induction. In one case it has been observed that the stereochemistry of the ring junction can

be affected by the combined action of two substituents8

•

Another tool to control the course of a cyclization has received only little attention. This is the inter-ruption of a cyclization process by nucleophilic capture of an intermediate carbenium ion, initiating a new reaction. Such a mechanism may be also operative in the enzymatic cyclizations. In this way influence on the ring junction can be expected and orientation effects may distinguish between five and six membered ring formation. Likely candi-dates for such interceptions are imidazole and sulphur containing groups. Both participate in biologica! reactions. Imidazole groups are essential for the working of many

enzymes as e.g. esterases and proteinases. The ability of sulphur to stabilize adjacent negative and positive charge is well established. lts biologica! importance is reflect-ed in its presence at key sites in several enzymes and co-enzymes. Presumably it plays an important role in the bio-synthesis of squalene by the coupling of two molecules farnesylpyrophosphate. Relevant to cyclization reactions is the repeated intermolecular methyl transfer in the

(12)

8-adenosyl-methionine dependent biologica! alkylation of the sterol side chain9 • Cyclization may be regarded as an intra-molecular alkyl transfer.

numerous routes for the

decomposition of the carbenium ion

In this thesis the involvement of participating groups in cyclization reactions is investigated. In Chapter II the generation and the reactions with nucleophiles of thiiranium ions are investigated. It is demonstrated that the reactions praeeed with inversion of configuration even when a tertiary carbon is involved. In Chapter III the effect upon ring junction is investigated. It is shown that only aia-fused products can be obtained. The precursor which should lead to trans fusion exhibits no cyclization tendency. In a non-nucleophilic, polar solvent a stable sulphonium ion is formed. The reaction of this ion with a strong nucleophile differs from those of comparable com-pounds. In Chapter IV precursors which contain an alkyl side chain similar to the part of squalene leading to C/D ring ciosure are investigated. The cyclization reactions of these compounds do not differ essentially from those initiated via a carbenium ion. In addition a precursor is synthesized which should cyclize via an intermediate having a cyclohexane ring in a boat configuration. Although this intermediate is formed, no cyclization takes place. In Chapter V a model for the biosynthesis of lanosterol and related triterpenes is described which is based on enzymic participation after the B ring formation.

(13)

Referenaes and Notes 1. Por reviews see:

a. E.E. van Tamelen, Acc. Chem. Res.,

!,

111 (1968). b. L.J. Mulheirn and P.J. Ramm, Chem. Soc. Rev.,

!,

259 (1972).

c. E.E. van Tamelen, XXIII International Congress IUPAC,

~. 85 (1975). 2. See for example:

a. E.E. van Tamelen, Acc. Chem. Res., ~. 152 (1975). b. R.E. Ireland, P. Bey, K. Cheng, R.J. Czarny, J. Moser

and R.I. Trust, J. Org. Chem., 40, 1000 (1975). c. W.S. Johnson, Angew. Chem., ~. 33 (1976).

3. P.A. Bartlett, J.L. Brauman, W.S. Johnson and R.A. Volk-man, J. Amer. Chem. Soc., 95, 7501 (1973).

4. E.E. van Tamelen and D.R. James, J. Amer. Chem. Soc., 99, 950 (1977).

5. A. Corvers, P.C.H. Scheers, W.A.M. Castenmilier and H.M. Buck, Tetrahedron, 457 (1978).

6. E. Palla, S. Borcic and D.E. Sunka, Tetrahedron Letters, 799 (1975).

7. P. Schipper, Thesis, Eindhoven (1977).

8. A.A. Macco, J.M.G. Driessen-Engels, M.L.M. Pennings, J.W. de Haan and H.M. Buck, Chem. Comm., in press. 9. M. Akhtar and C. Jones, Tetrahedron, 813 (1978).

(14)

CHAPTER 11

Thl

iranlu:~n

lons

11.1 IntPoduation

Intramolecular participation to ionization is reflect-ed in an enhancreflect-ed rate and leads to a cation of greater stability than the alternative carbenium ion. As a conse-quence side reactions are suppressed in the product form-ation. The overall process consists of two Walden inversions, resulting in a net retention of configuration. This tool to handle stereochemistry is very useful in substitution re-actions and can even be applied in systems which normally react via an SN1 mechanism.

Neighbouring group participation by sulphur substi-tuents is well known1 _{and has been first postulated in the}

reactions of ~-halo thioethers, like mustard gas. For example, Fueon et al. found that the ~-hydroxythioethers

l

and ~ gave the same S-chlorothioether

i

which is believed to arise from the thiiranium (episulphonium) ion

1.

occurring in both reactions2_•

yH3

C 2H5 -S-CH2-CHOH 1 CH 3 I HOCH 2 CH-S-C 2 H5 2 3 4 13

(15)

Nowadays, a large number of thiiranium ions has been investigated and several of them could be even isolated3

•

In most cases, however, the ions studied were generated in situ by the addition of arene- or alkylsulphenyl halides to alkenes4_• _{It is generally accepted that these reactions}

proceed via a symmetrically bridged intermediate.

8-8+ XSR

The direction of ring opening of these ions is influenced by the substituents. In general, the initial product com-position is controlled by steric factors, resulting mainly in a nucleophilic attack at the less substituted carbon. At room temperature the composition of the mixture changes in favour of the isomerie product, which is more stable. In this chapter the relative stability and the reactions of the ions Sa and Sb generated from the appropriate alcohols 6a and 6b, respectively, are investigated. In view of the use of related ions as initiators in cyclization reactions it was of interest to determine to what extent the react-ions of these react-ions are controlled by the sulphur group.

OH

(16)

II.Z The reaations of 4-tert-butyl-1-(phenylthiomethyl)-ayalohexanols ~ith acids

The two alcohols 6a and 6b were prepared from

4-tert-butylcyclohexanone and phenylthiomethyllithium5

, in a

ratio of 4:1, respectively6

• Pure samples

obtained by chromatographic means.

~

UCH2

sc

6H5

of each were

6a + 6b

Both alcohols could be converted to the corresponding

chlorides by HCl and SOC1₂without any detectable loss of

stereochemistry. When the reactions with SOC1 2 were

follow-ed with 1H NMR it was observed that each alcohol, tagether

with the tertiary chloride, gave rise to the formation of

another compound7

• In accordance with the products

ob-tained from addition of sulphenylhalides to alkenes, it was expected that these products were Sa and Sb.

Cl + HCI 6a S-

c

6H5 HCI _Cl ₊ 6b SOCI 2 7 b S-C 6H5 Sa or 6b HCOOH 9 15

(17)

The primary chloride, however, was converted simultaneously into the tertiary compound. After completion of the react-ion the trans (alkyl-alkyl) alcohol was entirely converted to the tertiary chloride, while the cis alcohol resulted in a mixture containing 10% of the primary chloride.

The reaction of and 6b with formic acid gave the alkene ~. containing less than 10% formates. When this alkene was treated with HCl a mixture of chlorides was ob-tained. After raarrangement to the thermodynamic equili-brium, 7a and 7b were formed in a ratio 1:4, together with a small amount of Sb.

The structure of the alcohols and chlorides was esta-blished with 1H and 13

c

NMR (see table I and II).

In the 1H NMR spectra the cis and trans isomers of each compound were distinguishable on the basis of their methylene (adjacent to the hetero atom) and/or tert-butyl signals. The methylene signal of the primary chloride was located downfield from the corresponding tertiary com-pounds in agreement with former observations~.

The 13

c

signals were assigned on the basis of their relative intensity and the resemblance to related compounds. In all compounds the tert-butyl is predominantly in the same position (the thermodynamically more favoured equa-torial position) as may be deduced from the chemica! shifts of the carbons of this group. The orientation of the other substituents is reflected in the chemica! shift of the carbon of the thioanisyl group and the ones in the 3 and 5 position of the ring by gauche interaction. The chemica! shift difference between the methylene carbon in an axial position and in an equatorial position is comparable with the chemica! shift difference between the methyls in the related 4-tert-butyl-1-methylcyclohexanols8_• _{A hydroxy} group or a chlorine induces a larger upfield effect than an alkyl substituant. As a ~onsequence,

c

3 and

c

5 appear at higher field in the trans isomers than in the cis isomers. It therefore appears that all the compounds ob-served are in a chair conformation with the tert-butyl group

(18)

in an equatorlal position.

II.3 Direat observation of the intermediate thiiranium ions The thiiranium ions Sa and Sb were stabie enough to be detected in non-nucleophilic solvents like

so

_{2. They were} prepared from the appropriate alcohols with an excess FS0_3H at -78

°e.

At -SO 0

e

fairly resolved 1H NMR spectra were obtained. At higher temperatures isomerization occurred to an equilibrium composition of 4:1 in favour of Sb (at

-30 °C). The 13

e

NMR spectra also showed two

se~

of signals in a 4:1 ratio. When this mixture was poured into a salution of LiCl in THF a mixture of chlorides and alkene was ob-tained, in which 8a and ~ were the major components. After rearrangement to the equilibrium composition 7a, 7b and Q

were present in a ratio 1:2:2.

The 1H NMR spectra are consistent with a sulphonium structure. The phenyl hydrogens are located at 7.67 ppm for Sb and at 7.64 ppm for Sa, O.S ppm downfield from the signals of the uncharged sulphides. The signals of the methylene group were at 4.33 ppm (singlet) and at 4.30 ppm

(AB quartet), respectively, comparable with the value 4.23 ppm9_, _{observed in a similar methyl thiiranium ion. The} tert-butyl methyls appeared at 0.97 ppm and at 0.90 ppm.

The 13

c

chemica! shifts (see table II) are assigned on the basis of intensity, line width in partially de-coupled spectra, and the correspondence to related com-pounds10. The values of the thiiranium ring are in good agreement with other ones. The asymmetry of the compounds is reflected in the differences between

c

₂and

c

₆as well as

c

_{3 and}

Cs

of both compounds. These differences are

greater in Sb in which the sulphur adopts an axial position. This is expected for

c

_{3 and}

Cs ,

but one would assume a comparable effect on

c

_{2 and}

c

_{6 in both ions. Another} re-markable feature is the reversed order (axial to equatorial) of the methylene carbons of the thiiranium ring with res-pect to the corresponding alcohols and chlorides. Possible

(19)

explanations are a hindered rotation of the phenyl group, which may cause steric interactions in both ions, or an in-creased population of the boat conformation of the less stable isoroer Sa.

II.4 Disaussion

The observed reactions as outlined in scheme 1, clear-ly demonstrate the participation of sulphur to the ioni-zation and the stereochemical control of the product format-ion by the intermediate thiiranium format-ions.

9 --10?

6b_t!,p7b

~Bb

scheme 1

It is obvious that the alcohols give the thiiranium ions without intermediacy of alkene ~ or a carbenium ion

!Q.

Such an ion cannot be excluded in the reaction of alkene 9 with HCl. (It seems likely that the interconversion of the thiiranium ions proceeds by such a species, because of the rotatien around the carbon-carbon bond which has to be in-volved). However, protonation at the least hindered side tagether with nucleophilic attack from the opposite direct-ion would also result in the formatdirect-ion of Sb.

The difference in stability between the two ions Sa and Sb must arise from electronica! and/or steric effects of the cyclohexyl part on the thiiranium ring. In Sb hyper-conjugation may cause a charge transfer from the axial

(20)

carbon-hydragen bond into the thiiranium ring, resulting in a weakening of the carbon-sulphur bond and consequently in a decrease of ring strain. A steric effect can be ratio-nalized by the interaction of the pseudoaxial substituent with the axial hydrogens located at

c

_{3 and}

c

_{5 of the}

cyclohexane ring. The free energy difference between a cyclohexane ring with an axial and one with an equatorial alkyl substituent is 1.7 kcal/mol. Fora thiomethyl group this difference is 1.0 kcal/mol. Replacing the methyl by a phenyl will not substantially affect this value, while the sulphur substituent will be pointed out of the ring in the highest populated conformation. In the ions Sa and Sb the increased distance to the axial hydrogens at

c

_{3 and CS} of the ring will lower this effect, but on the other hand the methylene group is held in an unfavourable (eclipsed) conformation. The influence of the phenyl will presumably be small, while the rotational freedom allows conformations in which the unfavourable interactions are minimalized. The occurrence of a baat conformation may be possible, because the bond angle in the thiiranium ring is smaller and hence interaction between the substituents in the 1 and 4 positions in a baat conformation will also be smaller than in normal cyclohexanes.

The equilibrium compositions to 8a and 7b to 8b reflect the smaller 1,3 diaxial interaction of a sulphenyl group compared to an alkyl substituent. In general, the tertiary chlorides are the energetically more favourable compounds (95 to 100%). Here the equilibrium composition is influenced by the cyclohexane ring. The primary chloride with an axial alkyl group is completely converted into the tertiary compound, while the one with an axial thiophenyl group remains present in the equilibrated mixture.

The kinetically controlled product composition shows that nucleophilic attack at both positions occurs but that, at least in Sb, attack on the less substituted carbon is favoured. Recently, Smith et al. found a nearly exclusive attack at the more substituted carbon11 _{and concluded that}

(21)

the intermediates of the reactions of alkenes with sul-phonylhalogenides differ from the thiiranium ions generated from 8-halothioethers. This is in disagreement with the results presented here, it seems more likely that the choice of nucleophiles and reaction conditions is of in-fluence on the product composition.

!1.5 ExpePimentaZ

1H NMR spectra were obtained on Varian Model T-60A and EM 360A spectrometers. Chemica! shifts are reported in ppm relative to TMS as international standard. 13

c

NMR spectra were obtained on a Varian HA-100 spectrometer equipped with a Digilab FTS-NMR-3 pulsing and data system. Chemica! shifts are reported relative to TMS as external standard. Microanalyses were performed in our laboratories by Messrs. H. Eding and P. van den Bosch.

4-(tept-Butyl)-1-(phenylthiomethyl)cyclohexanols (6a and 6b).

4-(tept-Butyl)-cyclohexanone (1.85 g, 12 mmo!) was added toa stirred salution of phenylthiomethyllithium (20 mmol) in THF (120 ml) at 0° under nitrogen. The mixture was kept overnight at room temperature. Ether work-up afforded the crude product in which 6a and 6b were present in a ratio of 4:1, respectively. Crystallization from pentane at -30 °C gave 2.31 g of alcohols (71~). Pure samples were obtained by chromatography over silicagel with CHC1₃as eluent, foliowed by recrystallization from pentane (mp: trans 72.5-73.7; ais 117.3-118.4 °C). (Anal calc C 73.33, H 9.41; found 6a 73.40, H 9.63; C 73.50, H 9.59).

The 1H NMR experiments were carried out by passing through HCl or addition of SOC1 2 to a salution of 50 mg alcohol 6a or 6b in 0.5 ml cn 2c1 2 or CDC13•

13

_c

_{NMR samples of chlorides 7a and 7b were prepared by} passing a stream of HCl through a salution of 30 ml benzene

(22)

Compound _c1 6a 71.73 fr 72.87 1!. 76. 18 7b 73.80 tr-;;,;s-methylb 68.S6 cis-methylc i 71.23

I

9 133.70 Sa 89.45 Sb 95.18 Table I

Selected 1H NMR Chemica! Shiftsa,b of 4-tert-Butylcyclohexyl Systems Compound 6a

I

7a Sa Sa 6b 7b

I

Sb

I

Sb

I

9 CH2X ' · 17

i '·"

3.50 4.30 3.02 3.3813,50 I 4.33 i 3.47 tert- _{0.83 0.80 0.93 0.90 0.85 0.88 0.72}

I

0.97

I

0.85 butyl

aDownfield from internal TMS; bsa and Sb in CD₃No₂; other compounds in CDC1 3

Table II

13 c Chemica! Shiftsa,b of 4-tert~Butylcyclohexyl Systems

69.

7

2

5 3 4 9 cz c3 c4 Cs c6 c7 cs Cg Phenyl i o,m o,m 3S.24 23.55 48.83 23.55 38.24 49.89 33.44 28.72 138.51 129.91 130.40 38.99 25.36 48.35 25.36 3S.99 44. 51 33.26 28.63 137. 13 129.91 131.01 39.74 23.95 48.21 23.95 39.74 51.08 33.44 28.58 13S.20 130.00 130.S4 41. 73 25.94 47.82 25.94 41.37 45.47 33.26 28.54 137.98 129.95 131.59 39.37 22.69 47.74 22.69 39.77 31.42 32.47 27.76 40.69 2 5. 13 47.94 25.13 40.96 25.42 32.39 27.76 126.51 27.79 44.73 . 24.97 29.64 42.57 32.95 28.10 13S.03 129.38 130.70 34.21 29.85 _47.00

I

29.55 36.66 ss. 1 g 33.38 27.84 120.33 133.76 132.24 31. 17 24.75 45.53 26.86 35.44 52.05 33.38 27.84 120.33 133.76 132.24 .

~ _{ain ppm downfield from external TMS; bsa and Sb in CD 3No2; other compounds in CDC1 3; cFrom reference 8}

p 127.00 127.39 127.35 127.84 126.69 134.88 134.88

(23)

containing 0.25 g of alcohol 6a or 6b for 4 hours. The salution was kept overnight under a HCl atmosphere. After washing with saturated bicarbonate and water, drying and concentration, there was left 0.255 g of product.

4-(te~t-Butyl)-1-(phenylthiomethyl)cyclohexene (~).

0.5 g of a mixture of alcohols 6a and 6b (ratio 2 to 3) was stirred in 30 ml formic acid (971) overnight at room temperature. The cloudy mixture was poured into water overlaid with pentane. The organic layer was wasbed with bicarbonate and water. After drying and concentratien there was left 0.43 g of product which consist of over

90% of alkene 9. NMR: 7.53-6.97 (m, Ar); 5.50 (m, alefin

(24)

Referenaes and Notes 1. See for example:

W.H. Mueller, Angew. Chem.,

!!•

475 (1969). K.O. Gundermann, Angew. Chem.,

Zi•

1194 (1963).

2. R.C. Fuson, C.C. Price and O.M. Burness, J. Org. Chem.,

.u.

475 (1946).

3.

s.

Ikegami, T. Asai, K. Tsuneoka, S. Matsumura and S. Akaboihi, Tetrahedron, 2087 (1974).

P. Raynolds, S. Zonnebelt, S. Bakker and R.M. Kellogg,

J. Amer. Chem. Soc., 96, 3146 (1974).

4. W.H. Mueller and P.E. Butler, J. Amer. Chem. Soc.,

2.Q, 2075 (1968).

G.H. Schmid, C.L. Oean and O.G. Garrat, Can. J. Chem.,

ii.

1253 (1976).

5. E.J. Corey and 0. Seebach, J. Org. Chem.,

11•

4097 (1966).

6. Alternatively, these alcohols could be prepared from oxiranes with thiophenolate. Because this reaction proceeds with retentien of the configuration at the spiro carbon, the geometry of 6a and 6b could be assigned beyond doubt by the use of oxiranes of known stereochemistry.

7. The amount of SOC1 2 used influenced the initial pro-duct composition. With 2 equivalents 7b and Sb were formed at about equal rate. With 5 equivalents Sb was more rapidly formed. The ratio 7a to 8a was in both cases in favour of 7a.

S. Y. Senda, J. Ishiyama and S. Imaizumi, Tetrahedron,

11·

1601 (1975).

9. G. Capozzi, 0. de Lucchi, V. Lucchini and G. Modena, Tetrahedron Letters, 30, 2603 (1975).

10. As in the 13

c

spectra~f

other phenylsulphonium ions (vide supra) the 13

c

shift of the para carbon appeared at lower field than in sulphides. The ipso carbon, however, is located at higher field. Apparently, the

(25)

charge transfer from the phenyl ring to the sulphur is small, but the ring is strongly polarized.

11. W.A. Smith, A.S. Gybin, V.S. Bogdanov, M.Z. Krimer and E.A. Vorobieva, Tetrahedron Letters,

11.

1085 (1978).

(26)

CHAPTER 111

Stereochemical aspects o-1 cyclizations

initiated by thiiranium ions

III.1 Introduation

The thiiranium ions Sa and Sb in Chapter II which are formed selectively from the alcoholic precursors react with nucleophiles with inversion of configuration at the

re~ctive carbon. The idea that these ions can be used as initiators in cyclizations originates from the work of Lansbury1 _{who has shown that the stereochemical outcome}

of cation-alkyne cyclizations is controlled by sulphur participation. The formolysis of the carbinol

1

resulted

in a mixture of hydrindans

A•

containing an excess cf the isomer in which the six- and five-membered ring are trans-fused. On the other hand, compound

1

gave almost exclusive-ly the ais-fused isomer

i·

In an attempt to explain these results, Lansbury has suggested that the reactions proceed via the intermediate sulphonium ion with the sulphur being placed axially (~),

because of the expectation that this ion is more stabie than the isomerie one with the sulphur in an equatorial position (&)· The ais-fused products should then result from a nucleophilic attack on the sulphur foliowed by in-sertion into the carbon-sulphur bond. The results of Chapter II confirm the first assumption; however, it is also obvious that the difference in free energy between the two isomers is relatively small, so that reactions proceeding via

&

may not be precluded. Lansbury has not established the configuration of the precursor

l•

but it seems likely (in view of the synthesis from the ketone and

(27)

R

=

CH3 3 R:: CH3 2 R = _{CH2SCsH 5} 4 _{R: CH 2}

sc

_{6 H5}

thiophenylmethyllithium) that the thioanisyl group accupies an equatorial position. In accordance with the results of Chapter II, this alcohol should lead to ion~. which may give ais-fused products by attack on carbon.

In this chapter the reactions proceeding via the thiiranium ions

Z

and

!

are discussed, examining the re-activity of such ions towards alkenes and clarifying the stereochemical aspects of the cyclization reactions.

(28)

III.Z The eynthesis and formolysis of ais and trans (but-3-enyZ)-1-(phenyZthiomethyZ)ayaZohereanoZe

Both the title compounds were prepared from 2-(but-3-enyl)-cyclohexanone (~)2_• _{Direct conversion with} thio-phenylmethyllithium gave mainly the trans (butenyl-thio-anisyl) isoroer 1Q (ratio 9:1). Alternatively the two com-pounds could be prepared by the epoxidation of alkene

g,

obtained from ~ with methylenetriphenylphosphorane, foliow-ed by opening of the oxirane ring with thiophenolate; the resulting mixture consisted of 60% 11 and 40~ 10.

~-!

~-

/

₁₃

scheme 1

The assignment of configuration was based initially on the stereochemistry of the reaction of ketone ~ with the thiophenylmethyl derivate. As for 4-tert-butylcyclohexanone, equatorial approach to the carbonyl should be favoured, leading to alcohol 10. This has been established by 13

c

NMR spectroscopy. The chemica! shifts showed a fair over-all correspondence with those of the related 1,2-dimethyl-cyclohexanols4. The differences in chemica! shifts between the ais and trans compounds are, however, not quite the same as those in the model compounds. This may well be

(29)

lated to differences in conformational equilibria for 11 and ais-1,2-dimethylcyclohexanol. The latter compound exists predominantly (ca 75~) in a chair conformation with the OH group in an equatorial position. The near equivalence of chemica! shifts of the carbon of the thiomethylene sub-stituant in 10 and

ll

indicates

ll

to exist largely in the conformation with the OH group in an axial position. An-other feature by which the two isomers can be distinguish-ed is the dependency of the chemica! shift on the tempe-rature. Variation in temperature only slightly affected the chemica! shifts of

lQ,

while on the other hand sub-stantial temperature effects were observed for

ll·

This is in agreement with the structure assignment, because in

lQ

the free energy difference between the two chair conformat-ions is relatively large (more than 2.5 kcal/mol) and therefore the amount of 10 in the less stabie conformation will remain negligible, even when the temperature is raised

(about 40°). In

!1

the free energy gap between the two

conformations is considerably smaller than in

lQ

and the equi-librium composition is in a range which is susceptible to temperature. 13

c

chemica! shifts are linearly correlated to the conformational equilibrium composition and therefore show the same effects.

Formolyses were performed on both mixtures as by

Lansbury. After one hour at reflux temperature the products were treated directly with LiAlH 4 and subsequently separated into an alkenic and an alcoholic fraction.

The alcoholic fraction was desulphurized with Raney-nickel and oxidized with chromic acid. Both mixtures gave the 9-methyl-cis-decalone

lZ

(see scheme 2).

This compound was identical with a sample prepared accord-ing to literature5_•

(30)

H 14 H 16

-H--

scheme 2 H 15 H 17

The 1H NMR spectrum of the alkenic fraction showed the characteristics of the terminal vinyl group, and in addition a broad singlet at 3.5 ppm and a complex multi-plet near 3.0 ppm for CH 2

s.

Further exposure to the re-action conditions resulted in a decrease of the vinyl

18 19

20 21

(31)

signals and the signal at 3.5 ppm. These signals are assigned to the uncyclized alkenes ~ and ~· The ones near 3.0 ppm conceivably arise from the bicyclic compounds 20 and 21.

III.3 Cy~Zization of the cis and tPans

1-bPomo-2-(but-3-enyZJ-(phenyZthiomethyl)~ycZohexane

Alcohols 10 and 11 were converted with PBr 3 to the corresponding bromides 22 and ~. respectively.

22 23

The bromides reacted with AgC10 4 in nitromethane contain-ing Caco 3 to prevent the formation of perchloric acid during the reaction. The cis bromide gave exclusively a mixture of the alkenes ~ and ~. no cyclized products being observed. The tPans bromide gave, besides a mixture of cyclized and uncyclized alkenes (ratio 1:1 basedon 1H NMR signals at 3.0 ppm and 3.5 ppm) also a crystalline substance. This compound could also be prepared from the alcohols lQ and

11

with perchloric acid in nitromethane. The absence of an olefinic group indicated a cyclized compound and the 1H and 13c NMR signals of the phenyl group demonstrated a sulphonium structure. In view of the foregoing (cyclization to cis-decalenium-ion) the sulphon-ium salt was assigned the structure 24. This structure was further confirmed by the 1H NMR showing signals for CH and CH 2 next to the sulphur at 4.53 ppm (multiple!) and at 3.96 ppm (AB quartet with an extra coupling of 1 Hz at the downfield part). Furthermore, the 13

c

spectrum showed the correct number of singlets, doublets and tri-plets. Dreiding models show 24 not to be free of strain.

(32)

6

8

2

25

Hence no direct comparison with other substituted (cis-)

decalins could be made. The doublet at 67.8 ppm and the triplet at 56.2 ppm showed one bond 13

c-

1H spin couplings of 155 Hz and 150 Hz, respectively. Therefore these sig-nals were assigned to

c

_{2 and}

c

_{11 , in this order (adjacent} to the charged sulphur). Furthermore, the aliphatic singlet at 51.2 ppm and the doublet at 41.5 ppm belang to

c

_{9 and}

c

₁₀, the latter seems to be located bere in a number of 9-methyl-cis-decalins (e.g.

11).

The remaining signals were assigned tentatively with reference to 9-methyl-ais-decalin in the chair-chair conformation (~) at low tem-perature as publisbed by Grant6

• The signals of

c

5 ,

c

6 and

c

_{7 do not deviate by more than}aa 1.5 ppm. The extra branching at

c

_{2 results in a downfield effect of 45 ppm} at

c

_{2 and a small upfield effect on}

c

_{4 , as expected. The} relatively small downfield effect on

c

_{1 and}

c

₃as well as the upfield shift of

c

_{8 are within an acceptable range.}

Compound 24 was unreactive towards formic acid. Ring opening could, however, be achieved with thiophenolate resulting in one isomer. Depending on the mode of ring opening either trans-2-phenylthio-9-phenylthiomethyl-ais-decalin ~ or the corresponding ais-2 product

lr

may result. Both can exist in two conformations as outlined in scheme 3. However, the conformations with the substituent

(33)

at

c

2 in the equatorial position (26a and 27b) will be more favourable than the corresponding axial position.

26 a b

b

scheme 3

The 13

c

shifts of these compounds will differ from the parent methyl-decalin ~ with respect to the ring with the sulphur substituent. Difference in the other ring can be deduced by replacing methyl by thioanisyl. The small difference between the methyl-and the phenylthiomethyl-cyclohexanols reveals that the chemical shifts for the carbons in the 3, 4 and 5 position of the ring are only slightly affected (Chapter II). Hence in compound 27 the 13

_c

shifts for Cs,

c

_{6 and}

c

_{7 should resemble the values} for

c

_{2 ,}

c

_{3 and}

c

₄of~ (22.8, 21.8, 28.4 ppm), but if structure 26 is correct the values for Cs,

c

_{6 and}

c

_{7 of 25}

(34)

Below 27 ppm only one signal is found, thus ~ can be ex-cluded. Further assignment based on multiplicity and correspondence to ~ is also consistent with ~;

c

_{4 is}

un-affected by the substituents. The sulphur substituents cause a downfield shift of 20 ppm at the a carbon, com-parable to the effect of a methylthio substituent (ca 18 ppm).

c

3 and

c

9,

s

to sulphur, both appear at lower field than in~ (ca 7 ppm). The smaller effect on

c

_{1 is due to} an additional upfield shift caused by gauche interactions with the other thiophenyl group, as is also observed for

c

_{8 and}

c

_{10 •}

Hence the product results from a nucleophilic attack on the methine carbon. This is rather unusual. Eliel et al. have investigated the reactions of the five-membered sul-phaniurn compounds ~. 29 and 30 with strong nucleophiles such as methylthiolate7_•

Q

_{~ ~}

I

28 29 30

It was found that ring opening by nucleophilic attack on methylene is preferred to methyl displacement, which in turn occurs faster than ring opening by nucleophilic

attack on methine (see

2Q).

In order to explain his results

EZiel has suggested that the relief of strain during the reaction might lower the energy of the transition state in the ring opening. This means that the transition state is reached after substantial progress along the reaction coordinate. This may also explain the direction of ring opening of ion 24. Reactions proceeding via late transition state lead to the more stabie product. In case of attack at the methine side 26a is formed and attack at the methyl-ene side gives 27a. The farmer is favoured thermodynamical-ly, so that the transition state leading to 26a will also be of lower energy.

(35)

II I. 4 Discussion

The results presented here do not agree with an

in-sertion mechanism proposed by Lansbury. Such a mechanism

leads to ion

l±•

which is stable to formic acid and,

con-sequently, this can not be an intermediate in the

format-ion of the formates 14.

The cyclization of a number of butenyl-cyclohexyl

compounds has already been investigated2

•

31 32 33 34

It was found that methyl alcohol gives predominantly

ais-fused products via a protonation-deprotonation

equi-librium. In contrast 32 and are conver.ted mainly into

trans-fused products without observable elimination. Fur-thermore, 34 leads to ais-fused products also

with-out elimination. 10 18 + 19

~

kc2 11 7 Scheme 4

The reactions involved in the formolyses of 10 and

11 are outlined in scheme 4 • As may be concluded from

the reactions of and 23 with AgC10₄, only

11

leads to

cyclization. The product composition will be determined by the relative rates of cyclization and elimination. It seems likely that in the transition states for cyclization the carbon-sulphur bond will be essentially broken and are,

(36)

therefore, comparable with those occurring in the cycli-zations without sulphur participation. As a consequence the transition state via which oia-fused products are formed will be of lower energy than the one leading to trana-fused products. In contrast to the reactions of

l1

the transition states arising from

2

and

!

have to be reached from precursors of unequal energy. If the relative stability of these ions follows the same order as their 4-tert-butyl analogs, the one with the sulphur axial (!)

will be of lower energy. Consequently, the activation energy for ring formation from

2

will be larger than those from

!•

so kc1 > kc2. Furthermore, from the dehydration of cyclic tertiary alcohols it is known that such reactions proceed preferably via a trana-diaxial transition state.

(Endocyclic alkenes are also easier formed from compounds with an axial leaving group such as

2

than from the corresponding equatorial isomer). Hence keZ > ke1.

In genera!, the preferenee for oia ring junction suggests an early transition state in which the relative stability of the formed decalin is of minor importance. On the other hand, the cyclizations of 32 and ~ may proceed via late, product-like transition states also leading to trana-fused products. This seems to disagree with the reactions of 34. However, the steric requirements of the phenyl group will prevent intermediacy of a planar carbenium ion and in addition influence the stability of the decalins as suggested by Harding.

III.5 Ezperimental

2-(But-3-enyl)-1-(phenylthiomethyl)cyclohexanol

(!Q

and

11)

from 2-(but-3-enyl)-cyclohexanone (~).

Prepared as for 6 in Chapter II in 94\ yield. IR: 3470 cm- 1• NMR: 3.13 Cs, CH 2S); 4.7-51 (m, 2x olefin H); 5.3-6.2 (m, olefin H); 6.9-7.5 (m, Ar).

(37)

2-(But-3-enyl)-methylenecyclohexane (~).

Ketone ~ (3.5 g, 23 mmol) was added toa stirred salution of methylene triphenylphosphorane (29 mmol) in DMSO (45 ml) under nitrogen. After stirring for 2 hr at 55° the mixture was poured into water. Etheral work-up afforded 3.0 g (85%) of 13. NMR: 4.7-5.1 (m, 2x olefin H); 5.3-6.2 (m, olefin H); 4.3-4.7 (m, 2x olefin H).

4-(But-3-enyl)-1-oxaspiro[2.5]octane

Clll·

A salution of m-chloroperbenzoic acid (2.18 g, 0.013 mol) in CH₂c1₂ (50 ml) was added to a stirred salution of diene

(1.5 g, 0.01 mol) in CH_{2c1 2} (60 ml). After 1 hr at 0°, excess peracid was destroyed by the addition of 5 ml of

10% aqueous sodium sulphite. After washing and drying, bulb to bulb destillation afforded 1.05 g (63%) of

ll·

NMR: 4.7-5.2, 5.4-6.2 (see 10 and

ll);

2.3-2.8 (mixture of two AB quartets, epoxide); (m, 2x alefin H).

2-(But-3-enyl)-1-(phenylthiomethyl)cyclohexanol

(lQ

and 11 from 13.

Epoxide

ll

(1.05 g, 6.3 mmol) was added to an ice-cooled salution of thiophenol (0.77 g, 7 mmol) and KOH (0.35 g, 6.3 mmol) in 80% aqueous methanol (30 ml). Stirring was continued for 1 hr at 0° and at room temperature overnight. Etheral work-up foliowed by chromatography on silica gel gave 1.4 g (81%) of a mixture of lQ and

ll·

Isomerie pure samples were obtained by chromatography under pressure (silica gel/hexane-ethylacetate (9:1)).

Formolyses of lQ and

11

and product characterization. A mixture of alcohol (2 g, 7.2 mmol) and formic acid (30 ml) containing acetic acid anhydride (3 ml) was refluxed

for 1 hr. The cooled salution was poured into a bicarbonate solution. After etheral work-up the crude product was added toa stirred suspension of LiA1H 4 (1 g) in THF (40 ml).

(38)

Table I

13_c_{Chemical Shiftsa of. Alcohols}_{jQ and}₁₁_{and Related Dimethylcyclohexanols}

10

""""

"

I

Compound c1 Cz C3 c4 cs c6 c7 es Cg c,o c,1 Phenyl

i o,m o.m p 1Q 74.42 43.32 27.92 26.38 22.72 37.98 46.49 29.16 33.17 139.92 115.44 138.69 129.78 130.261126.87 11 74.51 42.48 27.92 25.01 24.09 37.72 46.49 28.85 32.73 139.79 115.44 138.51 129.78 130.57 i 126.87

I

::~::>"'

70.98 40.46 30.69 26.08 22.14 40.10 28.75 . 15.28 C?..S- _{72.98 42.28} 32.21 25.48 24.26 41.49 zo.so i 15.47 I dimethylb 8

In ppm downfield from external TMS; in CDC1 3; bFrom reference 4 Table II

' 13c Chemical Shiftsa of l!b, llc and l&c Compound c, c2 C3 c4 Cs c6 c7 Cs Cg c10

I

c11 Phenyl i o,m

_I

o,m p 24° 35.73 67.77 26.27 26.07 28.51 26.45 24.51 37.13 51.20 41.38 56.24 125.76 131.41 1129.71 134.02 -zse 30.36 22.83 21.81 28.41 28.05 27.78 _{22.S3142.32 33.05}

41.761

28.30 26d 35.64 43.67 28.85 28.67 2S.45 27.48 22.81 35.46 40.10 39.26 46.80 139.08 129.82

I

130.92 127.00 39.26 i 139.95 129.95 132.65 127.62

(39)

After hydrolyse, extraction with ether, drying and evaporation of the ether the mixture was separated by chromatography on silica gel (pentane foliowed by chloro-form) into an alkenic and an alcoholic fraction. The aleo-holie fraction was added to 16 ml of Raney-nickel in 100 ml of methanol and refluxed for 18 hr. After filtration the product was taken up in chloroform and wasbed with bicarbonate and water. Drying and concentratien afforded 0.5 g of a mixture of alcohols. NMR: 0.90 and 0.95 (s, CH

3); 3.5-4.2 (m,

ego).

A sample of these alcohols (0.15 g) in 7.5 ml of acetone was treated with an excessof Jones reagent. After 5 min isopropyl alcohol was added. Etheral work-up afforded after purification by TLC 0.11 g of ketone 17 which was identical (1H and 13

c

NMR) with a sample prepared by an alternate route5_• _No _tPans _ketone (NMR: 0.97, methyl signa!) was detected.

Cyclization of alcohol 10 and 11 with perchloric acid. Toa stirred salution of alcohol

lQ

and

11

(0.5 g, 1.8 mmol) in nitromethane under nitrogen was added 0.5 ml of perchloric acid (60%). Afterstanding overnight the

mixture was diluted with 50 ml of ether and wasbed with a saturated bicarbonate solution. After stripping off the ether the nitromethane salution was extracted twice with hexane. Concentratien of the hexane fraction gave 0.21 g of a mixture of alkenes. NMR: 3.0 (m, SCH

2), 5.5-5.3 (m, alefin H). The nitromethane fraction was concentrated and the residue was taken up in hot ethanol. The product

(0.22 g) crystallized on standing. Mp: 139.5-140.0 °C (after recrystallization); anal.

c

₁₇H₂₃

c1o

₄

s

(358.90): calc.

C 56.89, H 6.46, Cl 9.88, S 9.94; found C 56.95, H. 6.44, Cl 9.88, S 8.93.

1-Bromo-2-(but-3-enyl) (phenylthiomethyl)cyclohexane (~and~).

Toa stirred salution of alcohol

lQ

or

11

(2.15 g, 7.78 mmol) in ether (25 ml) was added an excess of PBr₃ (0.3 ml,

(40)

3.16 mmol) at room temperature. Afterstanding overnight the salution was paured onta ice and bicarbonate. After extraction with ether and washing with bicarbonate and brine the salution was dried and concentrated to afford 2.33 g of bromide, which was cantaminated with alkene (20%). The mixture was used directly or storedat -20°. NMR: ~: 3.78 (AB quartet, CH₂S), 4.7-5.2 (m, 2x olefin H), 5.3-6.2 (m, alefin H), 6.9-7.6 (m, Ar); 23: 3.40 (s, CH 2S). Reaction of~ and ~ with AgC10₄.

In a typical experiment 0.5 g of~ in nitramethane (15 ml) was added dropwise with exclusion of light to a salution of AgC10₄ (0.45 g) in nitramethane (25 ml) in which Caco₃

(100 mg) was suspended. After stirring for 15 min the mixture was poured into 15 ml of a saturated bicarbonate salution overlaid withether (25 ml). This mixture was stirred for 10 min and subsequently filtered over celite. The organic layer was washed with bicarbonate, dried and concentrated in vacuo to yield 0.45 g of product which consisted of two phases. The alkenic fraction was taken up in pentane. The residue was taken up in hot ethanol. The sulphonium salt (0.115 g) crystallized on standing. Concentratien of the alkenic fraction afforded 0.332 g

of a mixture of alkenes.

Reaction of sulphonium salt 24 with thiophenolate. A salution of

l±

(0.1 g, 0.28 mmol) in methanol (5 ml) was added to a stirred solution of thiaphenol (0.034 g, 0.31 mmol) and KOH (0.025 g, 0.45 mmol). After 2 hr water was added and etheral wark-up foliowed by TLC affarded 95 mg (88%) of 26. NMR: 3.10 (AB quartet, CH

2S); 3.6-2.9

(m, CHS).

(41)

RefePenaes and Notes

1. P.T. Lansbury, T.R. Demmin, G.E. Dubais and V.R. Haddon,

J. Amer. Chem. Soc.,

zz,

394 (1975).

2. K.E. Harding, R.C. Ligon, C. Tsench and T. Wu, J. Org.

Chem., ~. 3478 (1973).

3. Preparatien of epoxide

11

by the reaction of ketone 9

with dimethylsulphoniummethylide and subsequent ring opening with thiophenolate gave exclusively alcohol 10. 4. Y. Senda, J. Ishiyama and S. Imaizumi, Tetrahedron,

1601 (1975).

5. A.J. Birch and R. Robinson, J. Chem. Soc., 501 (1943).

6. D.K. Dalling, D.M. Grant and E.G. Paul, J~ Amer. Chem.

Soc., 95, 3178 (1973).

7. E.L. Elliel, R.O. Hutchins, R. Mebane and R.L. Willer,

(42)

CHAPTER IV

Five- versus si x - membe:red

:ring :fo:rma-tion

IV.1 Introduetion

The formation of polycyclic compounds by the acid-catalyzed cyclization of 1,5-dienes is a well-known react-ion which bas been observed frequently in the chemistry of terpenes. The interest of organic chemists for such react-ions is a result of the high stereochemical outcome. For example, lanosterol which bas seven asymmetrie centers is formed from the precursor (35)-2,3-epoxysqualene which bas only one asymmetrie center. Such selectivity is not res-tricted to enzymatic reactions, but can also be achieved in non-enzymatic cyclizations. Selective formation of six new asymmetrie centers bas been reported by Johnson~

Despite the considerable work in this field, the mechanistic details have remained veiled. Stork and Esahenmoser2

pro-posed that each alkene group undergoes a trans addition as result of a concerted process. More recently, van Tamelen3

concluded on the basis of kinetic evidence that at least a part of the cyclizations proceeds step by step. In order to rationalize the ~rans addition to each alkene he des-cribed the intermediates as "frozen" carbenium ions. It is mechanistically significant that in a sequence of ring formations trans fusion is preferred and in a single cyclization step, when the carbenium ion is generated by protonation of an alkene or dehydroxylation of an alcohol, cis fusion often occurs (vide infra). This indicates that the stereochemical course of reactions via such carbenium ions depends on the way they are generated. An explanation

(43)

can be found if one assumes that cyclization is possible from more than one rotational conformation and that the activation energy for ring ciosure is less than the energy harrier for interconversion of the different rotational conformations. It has been found that olefinic elimination from a carbenium ion in the gas phase needs only a small excess of energy in addition to the difference between the heats of formation of starting and formed substances~. An upper limit of 11 kcal/mol has been found, but it has been argued that the real value may be close to zero. The

situation in an intramolecular process is more complex, because the transition state may be strained. However, since cyclization can occur with elimination and migration of methyl and hydrogen, the activation energy has to be small. If one considers a single cyclization step, four different products can be expected apart from the ring junction (see scheme 1).

(44)

There are two types of conformations which can lead to cyclization, those which have a pro-boat form and those which have a pro-ehair form. A new bond can be formed with

each of the carbons of the alkene group, hence six- or five-membered rings may result. In non-enzymatic reactions cyclization to five-membered rings only occu~when _{R1 is a}

hydrogen and R

2 and R3 are alkyl groups and always in such cases by a route including a boat folding. All other substitution patterns lead to six-membered rings usually via a pro-ehair folding, but reactions proceeding via a pro-boat folding arealso observed (see Chapter I).

Enzymic reactions show the same characteristics, however, when R1 is a hydrogen and R2 and R3 are alkyl groups, cyclization to a six-membered ring is also ob-served, but in such cases always via a chair-like con-formation (see for example C ring of lanosterol).

In the context of the foregoing proposals one can imagine that there are three conformations which allow ring formation. Two boat-like conformations result in a six-membered and a five-membered ring, respectively, whereas chair-like foiding oniy generates a six-membered ring. If for some reason cyclization via a boat foiding is not possibie, then a six-membered ring wiii result. In this way enzymes can control a cyclization process just by disfavouring some conformations. It can be seen from the cyclization reactions of compounds ~ and

&

that the fold-ing by which cyclization takes place, may be of much more importance than the relativa stability of the re-sulting carbenium ions2_{a. Both ring formations proceed}

via a boat-like conformation leading to compounds of dif-ferent stereochemistry. It must be mentioned that when cyclization initialiy leads to the less stabie ion, yet the more stabie one can result, because raarrangement reactions may occur after cyclization. Such reactions are frequently observed in solvolyses. For exampie the products derived from the acetolysis of tosyiate

l

are formed via ion ~5_• _{In the same way}

_i

_{may be converted}

(45)

Tos 0

7 8

into ~· However, in a polyene cyclization such rearrange-ments demand presumably a higher activatien energy than ring closure. It seems plausible that, besides direct interaction with the chain, the conformations from which cyclizations are possible can be restricted by geometrical changes of the carbenium ion.

9

As indicated by Dreiding roodels the approach to the re-active carbon in a non-planar ion can be achieved more easily via a chair-like folding than via a boat-like folding.

In Chapter III it has been shown that cyclization reactions of sulphonium ions praeeed with inversion, which determines the stereochemistry of the ring fusion. In this chapter the reactions of the substrates 9 and 10 are investigated. In contrast to species with a butenyl

(46)

side-chain, cyclization via a boat-like conformation will lead to other products than cyclization via a chair-like fold-ing. The substitution pattern of the first alkene group usually results in compounds of structure ~ which will give no further cyclization. If, however, a six-membered ring is formed, a second ring formation can occur.

IV.Z Syntheaia and cyclization of 1-bromo-2-(4~8-dimethyZ

(E)-non-dienyl)-1-phenylthiomethyl-cycZohexane Ketone

11

was synthesized by alkylation of N-cyclo-hexylcyclohexylimine with homogeranyliodide, which was prepared in three steps from geraniol, according to a route developed by Corey6

• Treatment with

phenylthiomethyl-lithium gave alcohol

11

which could be converted to bromide

~ _{with PBr 3 in ether. By analogy with the butenyl}

com-pounds (Chapter III) it was expected that

11

had a trans configuration (alkyl-alkyl). This was confirmed by the

1_{H NMR spectrum of the bromide} ₁₃_{which showed the same}

AB quartet as the corresponding butenyl compound (the ais compound gave a singlet).

0

Attempts to cyclize alcohol

11

failed. Formic acid reacted faster with the trisubstituted alkenes than with the hydroxyl group. Weaker acids gave no reaction at all. The conversion to bromide ~ proceeded without affecting the double bonds. Treatment with AgC10₄ afforded a mixture of alkenic and sulphonium compounds. The alkenic fraction

(47)

9 - - - 1 ...

14

consisted of uncyclized substances as indicated by the 1H NMR spectrum. The sulphonium ion fraction was also a mixture of isomers containing two major components as could be shown with 13

c

NMR. In each of these components one alkene group was present, showing only one ring to be formed. The most likely candidates are ~ and ~ arising from a cyclization via a boat conformation of the side-chain to ion~ (of structure 1), in which the carbenium ion was captured intramolecularly. Cyclization via a chair conformation can be excluded, because in an ion of type

l

interaction of the carbenium ion with the sulphur is geometrically unfavourable. Also six-membered rings are unlikely, because of the lack of conformity with ion

of Chapter III. The 1H NMR spectrum agreed with this assignment. The methylene hydrogens flanking the positively charged sulphur appeared at 3.98 ppm as a singlet. In the aliphatic region methyl signals were observed at 1.35 and 1.72 ppm (in addition to the signal for the alkenic

methyls at 1.60 and 1.67 ppm, respectively). The phenyl hydrogens appeared at 7.77 ppm as was demonstrated in other phenylsulphonium compounds. An attempt to elucidate

(48)

the structure in the same way as in Chapter III was un-successful: reaction with thiophenolate resulted in a mixture of alkenes. Apparently, eliminatien at the more substituted site is favoured over a nucleophilic displace-ment at the methylene carbon. It is wel! known that in a tertiary compound eliminatien is preferred to displacement, hence this result does not rule out the structures 15 and 16.

IV.3 Generation and reactions of a boat-like cyelohexyl sulphonium compound

Ion 10 has the characteristics for an intermediate which may occur in the epoxysqualene-lanosterol conversion: boat conformation and structure of the side-chain. The thiophenyl group simulates an enzymic site which partici-pates in the reaction. Generation of such an intermediate must proceed via a carbenium ion with a thioanisyl substi-tuent located at

c

_{4 of the cyclohexane ring. Synthesis of} substrates of this type must proceed via ketone

12

having a eis configuration.

12

was prepared according to scheme 2.

1 \

Q-

_Br 18 19 20 21 0 22 23 17 scheme 2 47

(49)

The known ester

1!

was prepared by a metbod developed by

Jung1_• _{After ketalization of the ketone function the}

ester~ _{was reduced with LiAlH4• Treatment with P(C6H5) 3}

and CBr 4 afforded bromide

ll•

which was converted to sul-phide

11·

After hydralysis ketone ~ was alkylated with homogeranyliodide via the cyclohexylimine similar to the preparatien of ketone

l!·

The two stereoisomers of

lL•

distinguishable by means of the doublets for the CH 2

s

protons in the 1H NMR spectra, were formed in about equal amounts. Isomerizations in tert-BuOH/tert-BuOK changed the ratio to 1:5. Pure samples were obtained by chromatography. It was expected that the major component has a ais confi-guration, because in this isomer the two substituents are both in an equatorial position. This assignment was con-firmed by 13

c

NMR (see table I). The differences between the cis isomer and ketone

11

are caused by substituent effects for the thioanisyl group which are comparable with those of other alkyl groups in the 4 position of the ring8

•

The situation encountered in the trans isomer is more complex, because the equilibrium composition of the two chair-like conformations is not known. However, the signals of the ring carbons of this isomer (except the carbonyl carbon) all appeared at higher field than those of the eis isomer. This might be expected, because all these car-bons are involved in gauche interactions caused by the axial substituents in the 2 or 4 position of the ring.

The formation of a bridged sulphonium ion (~) from an open carbenium ion could be demonstrated by the reaction of a mixture of alcohols (~) prepared from ketone 23 with CH 3Mgi, with perchloric acid in CD3No2.

Sulphur participation in cyclization reactions

Sulphur participation in cyclization reactions

SULPHUR PARTICIPATION IN

CYCLIZATION REACTIONS

CONTENTS

CHAPTER I

Cl)

Ci),

l

i

l

....

&

z,

l·

...

HO~

!

x

!,

!,

Thl

lons

l

i

1.

yH3

~

sc

c

c

c

c

c

so

°e.

e

e

se~

c

c

c

c

Cs

c

Cs ,

c

c

6b_t!,p7b

~Bb

!Q.

c

c

c

c

c

I

I

I

I

I

i '·"

I

I

69.

7

2

I

!!•

Zi•

.u.

s.

ii.

11•

11·

c

spectra~f

c

11.

CHAPTER 111

_c

-H--

_c

_{~ ~}