INFORMATION to USERS the Most Advanced Technology Has Been Used to Photo Graph and Reproduce This Manuscript from the Microfilm Master

Home , Dodecahedrane, Palladium on carbon

INFORMATION TO USERS The most advanced technology has been used to photo graph and reproduce this manuscript from the microfilm master. UMI films the original text directly from the copy submitted. Thus, some dissertation copies are in typewriter face, while others may be from a computer printer. In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyrighted material had to be removed, a note will indicate the deletion. Oversize materials (e.g., maps, drawings, charts) are re produced by sectioning the original, beginning at the upper left-hand comer and continuing from left to right in equal sections with small overlaps. Each oversize page is available as one exposure on a standard 35 mm slide or as a 17" x 23" black and white photographic print for an additional charge. Photographs included in the original manuscript have been reproduced xerographically in this copy. 35 mm slides or 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.

■UMIAccessing the Worldfc Information since 1938 300 North Zeeb Road. Ann Arbor, Ml 48106-1346 USA

Order Number 8804121

Synthetic approaches to structurally Interesting, spherical hydrocarbons and synthesis of monosubstituted dodecahedrane derivatives

Weber, Jeffrey Charles, Ph.D.

The Ohio State University, 1887

UMI 300 N. Zeeb Rd, Ann Aibor, Ml 48106

SYNTHETIC APPROACHES TO STRUCTURALLY INTERESTING, SPHERICAL HYDROCARBONS AND SYNTHESIS OF MONOSUBSTITUTED DODECAHEDRANE DERIVATIVES

DISSERTATION

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

Jeffrey Charles Weber, B.A.

The Ohio State University 1987

Reading Committee Approved By Dr. Anthony W. Czarnik Dr. Harold Shechter Adviser'5 F Dr. Leo A. Paquette Department of Chemistry To My Parents

ii ACKNOWLEDGEMENTS

I extend my sincere thanks to Professor Leo A. Paquette for providing the opportunity to complete this research. His encouragement and enthusiasm during my graduate studies are appreciated. I want to thank Anthony Schaefer for his friendship and support during our years at The Ohio State University. I owe special thanks to my family and fianc&e, Tina Kravetz, who have offered understanding and unending support throughout this work. Finally, I want to thank Kay Kampsen for her efficient and accurate preparation of this typewritten manuscript.

iii VITA

March 29, 1961 ...... Born - Evansville, Indiana 1983 B.A., DePauw University, Greencastle, Indiana 1983-1984 ...... Teaching Associate, Depart ment of Chemistry, The Ohio State University, Columbus, Ohio

1984-1987 ...... Research Associate, Depart ment of chemistry, The Ohio State University, Columbus, Ohio

PUBLICATIONS "Intramolecular Anionic Cyclization Route to Capped [3]Peristylanes," Garratt, P.J.; Doecke, C.W.; Weber, J.C.; Paquette, L.A. J. Org. Chem. 1986, 51, 449. "The Crystal and Molecular Structure of endo.endo-2.6- Diphenacyl[4]peristylane," Engel, P.; Weber, J.C.; Paquette, L.A. Zeit. Kristallogr. 1987. "Selective Functionalization of Dodecahedrane," Weber, J.C .t Kobayashl, T.; Paquette, L.A.; presented at the 19th Central Regional Meeting of the American Chemical Society, June 25, 1987, Columbus, Ohio.

FIELD OF STUDY Major Field: Organic Chemistry TABLE OF CONTENTS

Page DEDICATION ...... ii ACKNOWLEDGEMENTS ...... iii VITA...... iv LIST OF TABLES ...... vil LIST OF FIGURES ...... viii LIST OF SCHEMES ...... ix

CHAPTER I. INTRODUCTION ...... 1 II. SYNTHETIC ROUTES TO OCTAHEDRANE ...... 4 II-l. Introduction ...... 4 II-2. Synthesis of Cyclopropyl Ether 28 . 8 II-3. Alternative Approaches to Octahedrane ...... 18 II-4. Possible Synthetic Schemes for the F u t u r e ...... 31 III. SYNTHETIC ROUTES TO DECAHEDRANE...... 33

III-l. Introduction ...... 33 III-2. Approaches to Diketone 5 ...... 34 III-3. Review of the Initial Approaches to 2 ...... 37 III-4. Photochemical Approaches to Capping the [4]Peristylane Framework .... 45 III-5. Synthesis of a-Diketone 99 ...... 57 III-6. Outlook for the Future ...... 63

v Table of Contents (continued)

Page IV. ROUTES TO SUBSTITUTED DODECAHEDRANES ..... 67

IV-1. Introduction ...... 67 IV-2. Review of the Route to Aldehyde 1 3 5 ...... 69 IV-3. Continuing Toward Dodecahedrane ... 75 IV-4. Dehydrogenation of Secododeca- hedrane Derivatives ...... 79 IV-5. Direct Functionalization of Dodecahedrane. Formation of Monosubstituted Derivatives .... 91 IV-6. Synthesis of Amino Derivatives .... 119 IV-7. Nuclear Magnetic Resonance Observations ...... 130 IV-8. Paving the Way to New Molecules ... 141 IV-9. Summary ...... 147 EXPERIMENTAL SECTION ...... 148 REFERENCES AND NOTES ...... 211

vi LIST OF TABLES

Table Page 1. Proton NMR Chemical Shifts for Selected Dodecahedrane Derivatives (DDH-X) ...... 130 2. Carbon NMR Chemical Shifts for Selected Dodecahedrane Derivatives ...... 133 3. Substituent Effect for Selected Dodecahedrane Derivatives ...... 134

vii LIST OF FIGURES

Figure Page 1. side and Top Views of endo,endo-2#6- diphenacyl[4]peristylane (81) ...... 52 2. *H (300 MHz) and X*C (75 MHz) NMR Spectra of 163 ...... 94 3. Carbon-Hydrogen CorrelationSpectrum of 163 .. 95 4. *H (300 MHz) and 1SC (75 MHz) NMR Spectra of 175 ...... 106 5. Carbon-Hydrogen CorrelationSpectrum of 175 .. 107 6. X-Ray Structure of Ester 175 ...... 108 7. XH (300 MHz) and 13C (75 MHz) NMR Spectra of 178 ...... 114 8. a-Effect. Chemical Shifts (ppm) of a-Carbons, DDH-X vs. t-Bu-X ...... 136 9. {J-Effect. Chemical Shifts (ppm) of 8-Carbons, DDH-X V S. t-BU-X ...... 136 10. f-Effect. Shifts of f-Carbons, Substituent Effects vs. oj Values ...... 138 11. 6-Effect. Shifts of 6-Carbons, Substituent Effects vs. oj Values ...... 138 12. e-Effect. Shifts of e-Carbons, Substituent Effects vs. ag Values ...... 140 13. c*“E£fect. Shifts of c~Carbons, Substituent Effects vs. oj Values ...... 140

viii LIST OF SCHEMES

Scheme Page 1 32 2 35 3 36 4 39 5 46 6 57 7 64 8 72 9 77 10 78 11 85 12 98 13 Ill

14 117 15 121

ix CHAPTER I

Introduction

The Paquette group has been Involved in the synthesis of a series of structurally interesting, spherical hydro carbons. This family of polycyclic saturated systems is related by having a central n-membered ring connected by alternate carbons to two (n/2)-membered rings. The sim plest of these molecules are dodecahedrane (3) and its two lower homologues £-[3s.5*]octahedrane (1) and £-[42.5®J- decahedrane (2). The successful synthesis of 3 has caused interest to become focused on the smaller polycyclic molecules.

Dramatic alterations in structural topology are expected to accompany changes in n. Octahedrane and 2 decahedrane are of synthetic interest in order to define the precise nature of these geometric perturbations and to correlate the structural features with chemical react! vities. The topology of the molecules often gives rise to unexpected results in normally predictable chemical trans formations as will be exemplified in the sequel. In developing pathways to these highly symmetrical molecules, a synthetic design which takes maximum advantage of symmetry characteristics is desired. Elegant use of this concept greatly facilitated the synthesis of dodecahe drane (Chapter IV). The advantages to this strategy are obvious: first the number of steps to the target molecule is reduced by a factor of two; and second, the NMR spectra of intermediates quickly provide definition of whether the reaction has proceeded according to design. The routes to each molecule are discussed individually in the upcoming chapters. Chapter II deals with the various synthetic approaches to octahedrane from diketone

4 Although 4 is highly symmetrical and possesses manipulative functionality, conversion to 1 has not been accomplished to date. Chapter III discusses the many approaches toward decahedrane from its precursor dione 5 using strategy which follows a symmetrical approach. The synthesis of 2 has 0

5 also eluded us to this time, but as with 1 much invaluable chemical knowledge of the systems has been gathered. Special interest in dodecahedrane has been shown by both synthetic and theoretical chemists. The synthesis of the hydrocarbon in 1982 did not satiate these interests because the amount of dodecahedrane available for further study was greatly restricted. Improvements in the syn thetic scheme have allowed preparation of relatively large amounts of the parent hydrocarbon for these studies to commence. Initial investigations of the system are discussed in this dissertation (Chapter IV). These advances have resulted in the synthesis of a number of functionalized dodecahedrane derivatives and set the stage for the continued study of the system and the subsequent synthesis of second generation molecules. CHAPTER II

Synthetic Routes to Octahedrane

II-l INTRODUCTION

Octahedrane (1) has been of continued synthetic interest to a number of research groups for several years.1'2 ^o date, only the dimethyl derivative 9 has been prepared by Hirao and co-workers,2 whose synthetic strategy is shown below in abbreviated format.

6 i. h 2n n h c o n h 2 l KOH,A

3. O ji M e g S 0

Z.KOiBu, d iglym e, a

4 Treatment of dione 6 with etheral boron trifluoride gave 7 in 47% yield. Wolff-Kishner reduction of the carbonyl groups in 7 followed by ozonolysis of the exo-methylene groups gave 8 (overall 59%) following reductive workup. Conversion of 8 into its bis-tosylhydrazone and treatment with potassium tert-butoxide in refluxing diglyme furnished hydrocarbon 9. An awareness of this successful synthesis reenforced our desire to achieve the ultimate goal of arriving at the parent hydrocarbon (1), and this chapter deals with our synthetic approaches to date. Examination of the carbocyclic framework of 1 in a retrosynthetic sense has led us to consider diketone 4 as a potentially suitable starting material. Two symmetrically disposed carbon atoms are seen to be lacking and these

1 4

must ultimately serve as the linchpins for construction of the two cyclopropane rings. Diketone 4 was available from cyclopentadiene in three steps. The dimer of cyclopentadienone oxime.10 was

,0H / N 0

i f ) ^ U NaOEt /r^ IN HC1 )]T N 0

prepared as reported in the literature,3 and hydrolysis of the dioxime with levulinic acid gave dicyclopentadienedione ll.4 The photochemical conversion of 11 into 4 was achie ved, although not as efficiently as claimed in the litera ture. ^ Direct photoactivation of 11 in benzene solution at wavelengths > 305 nm led to cage diketone 12 via an 0

4 intramolecular [2+2] cycloaddition. Under these condi tions, where absorption of light by 12 is appreciable, the desired rearrangement product 4 can be obtained only after longer reaction times at the expense of 12, thus showing the reversible nature of the [2+2] cyclbaddition. Yields for the production of 4 are reported to be 82% on small scale (55 mg of 11, 16 h), but in attempts to prepare the i i amounts needed for a multistep synthesis[ the yields drop i to approximately 40% (2 g of 11) and reaction times must be I greatly increased (5 days) to achieve complete consumption | of starting materials. In this fashion, we were able to j prepare the necessary diketone in amounts sufficient for further elaboration. Ganter and co-workers have demonstrated that the car- | bonyl groups in 4 are sterically compressed,6 a fact which markedly lowers the reactivity at these centers (particu larly following nucleophilic attack at one of them). Indeed, we found in all initial attempts to achieve nucleo philic addition to the carbonyl groups that no adducts were formed. At extended reaction times, the diketone simply decomposed in the basic media (the only carbon nucleophiles

i known to attack the ketone groups are I methyllithium and methylmagnesium bromide prior to our studies6^). This approach to the preparation of 1 was subsequently reinves- j tigated and will be discussed later in this chapter. 8

0 (EtOLPCHCN Decomposition

We found that bridging across the molecule as in 13 reduced the steric compression at the remaining carbonyl

group, thereby allowing the implementations of those chemical manipulations necessary for proceeding toward 1.

II-2 SYNTHESIS OF CYCLOPROPYL ETHER 288

Keto ether 13 was prepared from 4 by reduction with lithium tri-tert-butoxyaluminum hydride and cyclization with 2N sodium hydroxide in methanol.7•9 9

4 14 13

Wadsworth-Emmons condensation of 13 with the anion of diethyl cyanomethylphosphonate generated with n-butyllithium proceeded to give 15 as an inseparable 4:1 mixture of isomers (81%). Reduction of the a,p-unsaturated nitrile

NC 15 16

m-CPBA

KH. THF

OH NC CN 18 17 functionality with magnesium in methanol10 occurred with complete regiocontrol and with delivery of hydrogen to the |3 carbon from its convex face to give 16 (94%). Con sideration of the structural geometry of 16, i.e. the presence of the acetonitrile residue in the cavity under the double bond, made possible the prediction that epoxida- tion should occur from the outer face of the olefin center. Indeed, 17 was formed in 96% yield on treatment with m- chloroperbenzoic acid. The stereochemical questions relevant to the conversion of 15 to 17 could not be unequivocally ascertained by spectral analysis; however, the correctness of the assump- N tions was proven in ring closure to IB. Treatment of 17 with a large excess of potassium hydride in refluxing tetrahydrofuran for 30 min gave alcohol 18 in 78% yield. Although two possible cyclization pathways are open to 17, only one is followed using the conditions stated.11 Obviously our goal of attaining 1 can be achieved with either isomer (18 or 19), but it is of some interest to ponder the underlying reasons for this selectivity. MM2 calculations on structures related to 18 and 19 in which the hydroxy and cyano groups have been excised to simplify matters showed the heterocylic ring system of 18 to be 7.7 kcal/mol more stable than that of 1 9 . Apparently, the added antagonist twist that the molecule must accommodate 11

“OH NC 19

during carbon-carbon bond formation in the transition state leading to 19 is enough to deter the reaction pathway. Treatment of alcohol 18 with tosyl chloride in pyridine yielded tosylate 20 (85%), which in turn readily underwent cyclization to give 21 (61%) when refluxed with a large excess of potassium hydride in tetrahydrofuran solvent for 30 minutes. 12

At this point, we had envisioned opening the ether linkage and performing the necessary chemical manipulations to arrive at keto olefin 23.

OH Me-.SU - - * - - i or equiv. NC 22

re p e a t _ sequence' 13 Repetition of the chemistry just completed could be reap plied to 23 to give dicyanooctahedrane 24 and ultimately 1. In order to obtain 23, the ether linkage needed to be cleaved. A number of reagents were studied with only 13 trimethylsilyl iodide resulting in a chemical transforma tion. Unfortunately, treatment of 21 with trimethylsilyl iodide at room temperature did not result in ether clea vage, but in opening of the cyclopropane ring to give 25 (87%).

TMSI

CN 21

The endo orientation of the cyano group was assumed on the basis of kinetically controlled protonation during the workup. It was also found that 25 could be reconverted to 21 upon treatment with potassium hydride. Thus,, with the ability to reform the cyclopropyl ring came the hope that more forcing conditions would result in cleavage of both the cyclopropane and ether linkages to deliver 26. After oxidation of the hydroxyl functionality in 26 to give ketone 27 and subsequent treatment with base to affect ring 14 formation and elimination of hydrogen iodide, keto olefin

23 would be in hand.

i o o

21 - - * - - *

NC r NC NC 26 27 23

Unfortunately, no evidence for the formation of 26 was gained after prolonged exposure of 21 to trimethylsilyl iodide in refluxing carbon tetrachloride; only lower yields of 25 (41%) were realized. It was thought that removal of the cyano group may lessen the likelihood of opening the cyclopropane ring, and effect instead the desired chemical transformation at the ether linkage. Attempted decyanation of 21 directly to hydrocarbon 28 14 using sodium in ammonia resulted in recovery of starting material. Consequently, a stepwise procedure for obtaining

28 was sought. The nitrile was reduced with diisobutylaluminum hydride*** to give aldehyde 29 after hydrolytic workup (58%). Treatment of 29 with Wilkinson's reagent16 under various reaction conditions including refluxing benzene only returned starting material. Attempted photodecarbonylation 15

H HOOC 30 21 29 /

Sr 28 31

of 29 using a 450 watt mercury lamp also yielded no products resembling the desired heterocycle. Oxidation of a solution of the aldehyde allowed to stand open to the atmosphere resulted in quantitative transformation to the acid 30. Unfortunately, the acid failed to undergo Huns- diecker degradation when subjected to several variants of this reaction (HgO, Br2, CH2Br217a* AgNO,, Br}, CH2Br217b). In light of these failures, it was quite surprising and fortuitous that in the treatment of tosylate 20 with excess potassium hydride for reaction times of one hour that not 16 only cyclopropyl ring formation but also decyanation occurred. The highly volatile heterocycle 28 could be

KH, THF

KH, THF A, l h > OTs CN 21 28

isolated in 11.1% yield. Nitrile 21 could in turn also be decyanated to give 28 in yields which varied from 8 to 12%. One possible mechanistic rationale for this transformation involves hydride attack at the carbon carrying the cyano group with the expulsion of hydrogen cyanide leaving the unstable tertiary carbanion which is protonated upon workup. The inefficiency of this reaction could be traced in part to the high volatility of 28, but degradation side- products most certainly formed competitively. Despite the complications, the desired hydrocarbon was now available to attempt the ether cleavage. Treatment of 28 with trimethylsilyl iodide at room temperature for three days resulted in the formation of two major products resulting from cyclopropyl ring cleavage (lH NMR analysis) with no companion products derived from the desired ether cleavage. 17

cyclopropyl TMSI ^ ring cleavage

The more notable features of the'300 MHz XH NMR spectrum of 28 Include three distinctive cyclopropyl signals at 6 2.09- 1.72 and a pair of absorptions due to the H-C-0 ether pro tons at 6 4.46 (t, J = 5.3 Hz) and 4.34-4.31 (m). It was obvious from the spectra of the reaction products that the ether linkages are still intact and that the resonances due to the cyclopropane ring no longer exist. No attempt was made to characterize the products further nor to optimize conditions, since it was abundantly clear that the neces sary chemoselectivity was not being followed. At this point, it was obvious that if our goal to arrive at 1 was to be realized, a new synthetic approach had to be developed. II-3 ALTERNATIVE APPROACHES TO OCTAHEDRANE

Due to our inability to cleave the ether linkage in either 21 or 28, a synthetic strategy to facilitate the cleavage was sought. The incorporation of a handle to mediate this cleavage reaction while leaving the cyclopro pane ring unaffected might not preclude use of the pre viously described chemistry. Thus the target molecule could be elaborated more efficiently using known chemistry and the difficulties expected in trying to manipulate the 19 ketonic functionality of 7 might be avoided. Bromide 32 was expected to fulfill this need. Its conversion to 33 and subsequent treatment with a metal such as magnesium held the prospect of effectively opening the ether linkage to give the desired hydroxy olefin functionality present in 34. Treatment of keto alcohol 14 with N-bromosuccinimide led to 32 in 91% yield. Upon examination of molecular models, it became apparent that the attached bromide might hinder the approach of nucleophilic reagents to the carbonyl center. Experimentally, this prediction reared its ugly head. Attempts at condensing the anion of diethyl cyanomethylphosphonate with 32 resulted in the recovery of starting material. Also, attempted Peterson olefination with trimethylsilylacetonitrile18 and addition of lithium acetonitrile*8 only returned 32. In light of these failures, efforts were channeled toward the direct manipulation of the carbonyl functionali ties within diketone 4. Preferably, both carbonyl groups might be manipulated at once in order to shorten the synthetic scheme. However, manipulation of one group at a time would most certainly have been acceptable.

,Ar° - > x ..... 35 1 20 Initial attempts at functionalizing diketone 4 involved Wadsworth-Emmons condensation with the lithium anion of diethyl cyanomethylphosphonate. After extended reaction periods, the starting material was seen to decompose in the basic media with no appearance of the desired products 36 or 37. Due to this sensitivity of 4 to basic

4 36 37

solution, it was deemed necessary to use a very reactive nucleophile if any success in functionalizing the carbonyl groups was to compete with the decomposition pathways. Whereas the Wadsworth-Emmons olefin synthesis involving phosphonate substituted carbanions has been shown to be ineffective in olefination with highly hindered ketones, the Peterson olefination reaction involving silicon substituted carbanions has proven successful.20 The reagent most studied was trimethylsilylacetonitrile IQ (TMSAN). Experimentally, lithio trimethylsilylacetoni trile could react with 4 to give many products. 21

CN OH TMS

0 CN 0

38 36 (CH3)3SiCH2CN 0 LDA >

CN 4 N NC

TMS CN 39 37

The mono-adduct 38 would result from addition of one equivalent of the TMSAN anion without subsequent elimina tion. Elimination of the (J-hydroxy silane would give a, (3- unsaturated nitrile 36. Products 39 and 37 would result from further exposure to excess reagent. Any of these products would be welcome due to the belief that they all could be converted into 37. This bis-unsaturated nitrile possesses the functionality necessary for conversion into 1 using the anionic cyclization route developed above. Initially, treatment of 4 with excess TMSAN anion at low temperature resulted in no reaction. On warming to room temperature, the two products formed in varying amounts were isolated and identified as 38 and 39. In a standard procedure, the diketone was stirred with six equi valents of lithio TMSAN and 38 and 39 resulted in 15.6% and 30.5% yield, respectively. Interestingly, no 36 or 37 was observed or isolated in the many reaction variants attemp ted. Shorter reaction times yielded more 38, whereas longer reaction times gave only 39 in very low yields (< 9.5%), presumably due to decomposition of products in the basic media. In explaining the product distribution and the fact that no elimination products were isolated, one must assume that elimination within the 8-hydroxy silane, which is usually facile in Peterson olefinations, is slow relative to the addition of the anion to the carbonyl group. Thus after elimination to 36 occurs (or the initial addition occurs on 4) a TMSAN anion equivalent rapidly attacks giving 39. Before we could proceed, it was felt that due to the number of steps required to transform these compounds into octahedrane that the yields needed to be improved if possible. In order to accomplish this goal, changes in the counterion and solvent were investigated. A switch from sodium to potassium did not improve the yields;

4 rather, no reaction occurred at all. Addition of 23 hexamethylphosphoramide and 18-crown-6 to increase the nucleophilicity of the carbanion only resulted in faster decomposition of the starting material and products. The amounts of products obtained using the standard procedure (6 equivalents of TMSAN anion) were also found to vary significantly from run to run. The stated yields were by far the best seen during the investigation. Before additional time was spent trying to improve the yields, the chemistry of these adducts was studied for its intrinsic value. Treatment of mono-adduct 38 with bases (KH, NaH) in various solvents (THF, DME) gave no elimina tion. Instead, the trimethylsilylacetonitrile group was removed returning the diketone 4. This unfortunate result

4 38 36

may possibly shed light on the actual mechanism by which 38 is formed from 4. In basic media (anion of TMSAN), it appears that an equilibrium exists between 4 and 38 and this equilibrium 24

{CH,),SiCHCN TMS 0

4 38

can be driven to the side of the diketone. Use of an excess of TMSAN anion, which was needed for the reation to proceed at a reasonable rate, also shifts the equilibrium toward 4 by speeding up the retrograde reaction. And since the diketone is known to decompose under basic conditions, low yields of products resulted. Acidic eliminations21 were therefore studied briefly. Treatment of 38 with acids (H2so,,, BFs*Etao) led only to decomposition. Thus, it appeared that 38 was of no syn thetic value in our route to 1. The chemistry of 39 was next investigated. Elimina tions using bases were not studied due to the small amount of material available and the prospect that base would lead to the retrograde product 36. This obvious backtracking was not acceptable. Attempts to achieve acidic elimination were examined, but to no avail. Treatment of 39 with etheral boron trifluoride returned starting material, 25

CN CN NC,

TMS 39 36

whereas reaction with sulfuric acid resulted in protodesi- lylation and etherification giving 40.

39 40

At this point, it was decided that functionalization with TMSAN should be abandoned in favor of a more promising and efficient route which was being studied concurrently. In looking for very reactive nucleophiles, lithium aceto- nitrile19 was envisioned as a direct source of 41. Elimi nation of the tertiary alcohol functionality would give the unsaturated nitrile 36 (Peterson olefin product) which could be resubmitted to the conditions and ultimately result in the formation of 37 (double Peterson olefin product).

OH CN

0 0 -45 C~> RT 4 41 36

Quite promisingly, the addition of lithium acetonitrile did proceed as desired to give 41 in high yield (86%). The heterogeneous reaction was efficient and rapid (3 h) at -78 °C, but at higher temperatures (-45 °C ■+) the alcohol 41 rapidly regenerated the diketone 4. On larger scale, the reation proved very sluggish (slow reaction and inconsis tent yields) due to the insolubility of the lithium acetonitrile even under high dilution conditions. Unfor tunately, all attempts to dehydrate 41 (I,; HMPA, 4; S0C12, Pyr; CFjSOjCl, EtsN) were unsuccessful. To facilitate the elimination pathway, preparation of a suitable derivative from 41 that would contain an efficient leaving group was sought. Many attempts to prepare various derivatives using standard procedures resulted only in recovery of the alcohol. However, quenching the alkoxy anion with acetic 27

36 CH-CN 0 TPTCtf CN ! -HX THF 4 41

0 CN

anhydride at -78 °C resulted in formation of acetate 42 in low (35%) and unreproducible yields. Attempted elimination

NC OCCH

4 42 36 28 using standard procedures (DBU; DBN; DABCO, DMAP) only returned the acetate, with no' elimination products being observed. Interestingly, treating 42 with sodium hydride in dimethoxyethane for 15 min at ambient temperature gave diketone 4. Presumably the sodium hydride deprotonates the methyl group- of the acetate, which subsequently degrades

O ^ H

42 4

with expulsion of ketene and the stable acetonitrile anion. Attempts to prepare the trifluoroacetate and trifluoromethanesulfonate derivatives in similar fashion failed. Encouragingly, the phosphonate derivative 43 could be syn thesized using chlorodiethylphosphonate as the trapping reagent.

O it OP(OEtl2 ° 1. LICHEN. -7 8 ° C 0 CN u 2. (EtO)2PCl 4 43 The phosphate was found to be very polar on silica gel and difficult to purify for accurate yields to be determined. Nevertheless, elimination of the slightly impure material was studied. It was believed that due to the bulkiness of this group tucked under the carbon-carbon double bond that elimination would proceed readily. To our amazement, treatment of 43 with bases (Et,N, NaH, KH, KOt-Bu, LDA) invariably returned starting material. One can only assume that the bases cannot reach the acidic hydrogens a to the nitrile group which would lead to the expulsion of the leaving group.

Not wanting to forego the ability to functionalize the diketone with such ease, we investigated another approach. This ploy involved treating ether 40 with strong bases to achieve opening of the ether linkage and arrival at bis- unsaturated nitrile 45. The ether linkage was present to eliminate extrusion of the side chain as observed pre-

4 viously with hydroxy silane 38, alcohol 41, and acetate 30

OH _ N a O H CN CN H20,CH3 0f?

41 44 0 (EtO)2PCH2CN J, n-BuLi

CN CN 45 40

42. The experimental sequence started with the treatment of 41 with sodium hydroxide in methanol to give 44 in high yield (93%). Wadsworth-Emmons condensation on the remain ing carbonyl group delivered the unsaturated nitrile 40 (91%). Disappointingly, treatment of 40 with various strong bases (KH, KOt-Bu, NaNH,, KNH3, rv-BuLi) only returned the starting material with no appearance of 45.

4 31 II-4 POSSIBLE SYNTHETIC SCHEMES FOR THE FUTURE

At this date, the research attempts toward octahedrane have ceased due to the endless failures In the necessary chemical manipulations that we deemed hopeful. Avenues for future research In arriving at our ultimate goal ard shown In Scheme 1. Wittig or Tebbe reaction of 4 should give tetraolefln 46. Selective oxidation using borane chemistry may prove to be difficult, but conditions possibly could be found to give diol 47. Oxidation of 47 would give dialdehyde 48 which could perhaps be prepared in a more direct way using 22 chemistry developed by Magnus. Peterson olefination using (methoxy(trimethylsilyl)methyl)lithium may prove to be reactive enough (as seen in TMSAN) to give sufficient bis-enol ether quantities for conversion into 1. Acidic hydrolysis of 49 would give dialdehyde 48. As with similiar compounds used to prepare [3Jperistylane by Garratt,1 48 may be sensitive to epimerization and give rise to 50 where the carbonyl functions exist in the exo positions out of the cavity. Formation of the tosylhydrazone should be straightforward, along with subsequent carbene generation and insertion into the olefin would give octahedrane in as few as four laboratory steps from 4. 32

Scheme 1 OH

,0 Wittig H or 1. borane h Tebbe 2. oxidative reaction workup 46 OH 47

(CHjJjSiCHgOCH, PCC sec-BuLi OMe

CHO MeO x xepimerization 49 48 *

OHC CHO

= N N H T s

TsHNN=CH V 51

I I 'k

H H -> CH 1 CHAPTER III

Synthetic Routes to Decahedrane

III-l INTRODUCTION

Synthetic approaches to j>-[4J.5B]decahedrane (2) have been underway in the Paquette group for several years. The pathways start from [4]peristylane-2,6-dione (5).

2 5 52

Adoption of 5 as the precursor molecule of choice was obvious due to its symmetry and ketonic functionality. Also, 5 was the key intermediate to [4]peristylane (52), which was synthesized at The Ohio State University in 1983. 23 A successful conversion of 5 into 2 was believed to rest on the development of a strategy to cap the

33 34 [4]peristylane framework. As progress toward 2 began, it became evident that interconnective bonding between pairs of methylene carbons along the periphery of 5 introduced enhanced levels of strain. Under these circumstances, normally predictable chemical transformations sometimes proceeded in unexpected ways. The initial attempts at

*5 A belting the [4]peristylane framework have been published and will be briefly discussed (Section III-3). First, the preparation of dione 5 will be reviewed (Section III-2).

Ill-2 APPROACHES TO DIKETONE 5

Two pathways have been developed to arrive at dione 5 starting with tricyclo[5.2.1.02 •6]deca-2,5,8-triene (53). The first of these involves a seven-step procedure which is 23 24 amenable to large scale (Scheme 2). ' The synthesis was initiated by Diels-Alder addition of p-toluenesulfonylacetylene to triene 53 in refluxing ben zene solution under an inert atmosphere. Because of the tendency of adduct 54 to undergo air oxidation, no attempt was made to isolate the adduct; rather, the substance was directly subjected to peracid oxidation to give 55. While on small scale (1.7 g of 53), near quantitative yields of 55 were achieved, recourse to preparatively useful quan tities (17-28 g of 53) caused yields to drop to 50-62% 35

Scheme 2

T i - = - H MCPBA CsHg.A

Ts T s 53 54 55

I hV 1 3 5 0 nm ' a c e t o n e

CH3OH-

T t 57 56

after recrystallization. Molecular models of 55 reveal that installation of the oxirane ring from the exo surface has the anticipated effect of compressing the ti systems in a face-to-face arrangement. 25 Irradiation of acetone solu tions of 55 through pyrex in a Rayonet apparatus with 350- nm light gave the [2+2] cycloaddition product 56 in quantitative yield. 26 Exposure of 56 to periodic acid in reflux ing 10% aqueous methanol solution afforded the desired diketo sulfone 57 in 95% yield. 27 Of the various methods 36 examined to arrive at 5, the most expedient involved sequential ketalization, reductive desulfonylation with 28 lithium in ethylamine, and acidic hydrolysis (69% for 3 steps). The overall yield from 53 of 33-41% made pos sible the preparation of significant quantities of dione 5. 2 9 The second procedure, outlined in Scheme 3, is ini tiated by Diels-Alder cycloaddition of triene 53 with [Z]- l,2-bis(phenylsulfonyl)ethylene in methylene chloride at room temperature for 2 days. Subsequent peracid oxidation afforded 58 in 85% yield on large scale. The difficulties in the sequence unfortunately arise at this point. To

Scheme 3

■S0*Ph cS 0 2Ph MCPBA S02Ph SO,Ph S02Ph SOjPh 53 58

I-2V. No/Hg CHjOH

HIO.

CHjOH 3 5 0 0 h HjO acetone 60 59 arrive at diene 59, reductive elimination of the disulfone was necessary. The desulfonylation to 59 could be achieved by using 1-2% sodium amalgam with yields upward of 85%, but the reaction can only be carried out on small scale due to the need for large amounts of mercury. To eliminate this scale limitation, the desulfonylation was attempted using 30 magnesium in methanol. The epoxy diene 59 was prepared albeit in low yields using this procedure but large amounts of side-products resulted, presumably from overreduction of the olefin centers. The photolytic cycloaddition of 59 and subsequent periodic acid treatment afforded 5 in yields comparable to the previous protocol. Thus, although this procedure involved only five steps, it did not allow for large scale manipulation.

III-3 REVIEW OF THE INITIAL APPROACHES TO 2

The successful transformation of 5 into 2 could rest on the development of a successful strategy for "capping” of the [4]peristylane framework. Several possible dissections of 2 to 61 are conceptually possible, though the task of experimental reconstruction has proven to be an extremely difficult one. Nevertheless, considerable useful infor mation has been gained by examining methods for "belting" 38

n = 2,3,4

5 61

[4]peristylanesy i.e., bridging across the cavity with chains of varied atomic length and functionaltiy. The initial retrosynthetic analysis is shown in Scheme 4. Hydrocarbon 2 might result from Norrish Type II photolysis of 62 followed by removal of the hydroxyl groups. This dione was envisioned to arise from cleavage of the vicinal diol 63 using standard methods. Again, photolytic cyclization of a-diketone 64 would connect the belted chain to the methano bridges, giving rise to 63. Acyloin reaction of diester 65 and subsequent oxidation of the a-hydroxy ketone would give 64, with 65 being available in two steps from dione 5. Experimentally, the sequence began as desired under the aegis of Dr. John W. Fischer. When 5 was subjected to Wadsworth-Emmons condensation with excess trimethyl phos- phonoacetate, bis-a,(3-unsaturated diester 66 was obtained as a mixture of geometric isomers, as indicated by 19C NMR. 39

Scheme 4

2 6 2 63U o <

5 65 64

Direct hydrogenation of this material afforded 65 as a colorless, viscous oil. The stereochemical homogeneity of this product was evident from its simplified and 13C NMR spectra, which reasonably describe only a structure having two planes of symmetry. Concordant with steric factors prevailing within 66, delivery of hydrogen should have occurred from the exo surface to force the acetic ester residues into the interior of the molecular cavity as in 65. 40

,C00CH3 ,cooch 3 0 0 3 \ h CH ^

Despite the apparent proximity of the functional groups 31 in 65, all attempts to effect Dieckmann or acyloin cycli- 32 ration of the diester were unsuccessful. Various bases and reaction conditions were examined but all to no avail.

O S iM a3 COOCH-, CKjOOC COOCH3

• HI H l|!eckmann Acyloin ^ condensation condensation

68 65

This inability to construct a three- or four-carbon belt in the manners indicated in 68 and 69 is taken to be a reflec tion of the severe nonbonded steric interactions that necessarily come into play when aligning the acetic ester residues into the- necessary geometry for condensation.

This interaction will be elaborated further ' in this chapter. These experimental difficulties required development of a new pathway to arrive at the desired a-diketone 64. In three laboratory steps 65 was transformed into dibromide 70. The functionality of this compound offered options for

Cf^OOC .COOCH3 in ■» f

65 70 71 further manipulations. Initially, coupling of the dibro mide to give the hydrocarbon 71 was attempted following a 33 procedure reported by Whitesides. Thus, 70 was allowed to react with excess magnesium metal and exposed to silver triflate with the appropriate control of dilution. Disap pointingly, no 71 was obtained as indicated by GC-MS analysis. 42

SCH,

Stevens 4----- rearr.

74 72 Ramberg H Backlund m-ipBA reaction

Although this coupling to give a four-atom belted [4]peristylane had failed, 70 did react with sodium sulfide to give sulfide 72. This five-atom belted [4]peristylane was the first compound in which two adjacent carbons of the peristylane network had been connected. Efforts turned to possible ring contraction of 72 via the Ramberg-BScklund 34 35 reaction or Stevens rearrangement. Under various Ramberg-BScklund conditions, no ring contraction to the olefin was observed. Also, attempts at Stevens rearrange ment by generation of the sulfonium salt with 43 gg dimethoxycarbenium tetrafluoroborate and treatment with strong bases did not give the expected product, instead, 3-elimination occurred to give 74. Sulfone 75 acted similarly under Ramberg-BHcklund conditions. At this point, bridging the [4]peristylane framework with a four- carbon chain was obviously to prove more difficult than initially anticipated. A possible means of overcoming the above complications was to install temporarily a central bond across the [4]peristylane system in such a way as to alter profoundly the molecular geometry. The requisite bond was introduced by treatment of bis-unsaturated ester 66 with zinc in ether saturated with gaseous hydrochloric acid. The benefits derived from the presence of the intraannular bond can be seen in the ease with which 76 undergoes the acyloin con densation. The intraannular bond forces the methano brid ges outward from the [4]peristylane cavity, bringing the acetic ester residues into close proximity and allowing for ready carbon-carbon bond formation. Direct oxidation of the acyloin product gave a-diketone 77 (32%), which existed as its monoenolized tautomer. Conversion of 77 to the a-bromo derivative 78 was straightforward, but attempts at achieving a 1,4-dehydrobromination to give 79 were unsuc cessful. The strain in 79 does not apear to be so extreme as to prevent the dehydrobromination from occurring, 44

HiimQ COOCH, CHjOOC^ ^ CH^)OCv ^C00CH3 1 .N o -K , M a 3SlCI HCI ather alhar 2.FaCt3|HCI 66 76 alhar 77

Py *H B r3 C H C I3

79 78

but realization that the CHBrCO proton is the most acidic within 79 and would be removed first may explain this failure. Fragmentation of this resulting enolate is unlikely and experimentally, 79 was not recovered following exposure to base. 45 III-4 PHOTOCHEMICAL APPROACHES TO CAPPING THE [4]PERISTYLANE FRAMEWORK

With the completion of Dr. John Fischer's work and still lacking an intermediate which would allow for further elaboration to 2, this investigator's routes toward the target molecule began. The nature of our approach was focused on the study of photochemical methods. The extra energy at one's disposal in photochemical transformations is well known, and it was hoped that this added energy would allow us to overcome the forementioned synthetic difficulties. The photochemistry of carbonyl compounds has received a great deal of attention as an efficient means of func- 37 tionalizing unactivated positions on alkyl chains. Photoexcited carbonyl groups have been found capable of Y-hydrogen abstraction in a process known as the Norrish II 37 38 reaction. # Commonly, the diradical intermediate has two choices: radical collapse to form cyclic alcohols or 39 bond scission (see Scheme 5). 46

Scheme 5 "tlOH o o 'H r

< A > +

Aldehydes as well as ketones have been observed to undergo these processes.In addition, many examples of cyclopentane ring formation (6 -hydrogen abstraction) have appeared in cases where Y“hydrogens are lacking or inacces sible for abstraction by the oxygen of the carbonyl.3^ a '40 With photochemistry in mind, the dicarbonyl compounds 80 and 81 were considered attractive molecules. The photo chemical products possess functionality which could be manipulated into 2. The Norrish products (82 and 83) could be doubly dehydrated, and following subsequent photolysis to affect the [2+ 2 ]cycloaddition, the hydrocarbon or the diphenyl derivative would be in hand. 47

R RC. CR HO

-by.

R-H 82 Ph B1 Ph 83 i i 2H20

* J » v .

R=H 2 R=H 84 Ph 86 Ph 85

Dialdehyde 80 was obtained in 54% yield by PCC oxida tion of 87.

OH OH

PCC, CH2C12 * NaOAc, Celite

87 80 82 48

Two-fold 6 -hydrogen abstraction and subsequent ring closure of the 1,5-diradical was necessary to give 82. Unfortunately, all photochemical attempts to realize these chemical events at ambient temperature resulted In decomposition of 80. Since photolytlc reactions of aldehydes often result In decarbonylatlon, 40 the reaction was repeated at low temperature. Down as low as -60 °C, the only product Isolated was Identified as 88 by NMR analysis.

HC CH HC. CH

-60 C

80 88

Despite the fact that decarbonylatlon results In formation of a relatively unstable primary radical, the volatile 88 was Isolated in 47% yield after chromatography. As indi cated in the literature,this photodecarbonylation process can be greatly reduced by changing the aldehydic hydrogen into a phenyl group. Thus, 81 was sought. The diphenacyl derivative 81 was surprisingly difficult to prepare in an expedient manner. Initially, the synthesis of 81 was attempted from diacid 89, which in turn was prepared by saponification of diester 65. 49

XOOCHj HOOC COOH ,COO Li

c H K "5% h2° / H ^ 0H ) M / 2. HC1 \H

65 89 90

Addition of phenyllithium to a tetrahydrofuran solution of diacid 89 resulted in formation of the dilithio dicarboxy- 42 late 90, which routinely precipitated from solution. Due to the insolubility of this salt, addition of another equivalent of phenyllithium did not occur, and 89 was recovered upon workup. The use of high dilution or HMPA and TMEDA as solvents did not facilitate the reaction. At this point, we returned to prepare 81 from dione 5. Attempted Wadsworth-Emmons condensation using diethyl 43 benzoylmethylphosphonate according to Borowitz under a wide variety of conditions (solvents, metal hydrides, temperatures) surprisingly only returned starting material. 50

C6l%CO .coc6HS

In light of these difficulties to prepare 81 in an expedient manner, the decision was made to use a less efficient procedure to obtain a sufficient amount of 81. The route starts with dialdehyde 80. Addition of

.CHOHPh PhOHCH ,COC6 H5

phenyllithium afforded 92 and subsequent oxidation with manganese oxide gave 81 in an unoptimized yield of 43%. The high crystallinity of this compound allowed for growth of an x-ray crystal whose analysis helped to ascertain 44 insight into the subsequent photochemical transformation. 51 As one can see from the x-ray structure (Figure 1) and

through close examination of models, the phenyl groups would prefer to be up away from the [4]peristylane cavity with the adjacent carbonyl groups positioned to abstract a hydrogen. The question is which hydrogen would be prefer red? Close examination of models show the hydrogens on the

0 - and y-carbons are not in the preferred proximity to the carbonyl oxygen but possibly close enough in the excited state to be a viable process. However, abstraction of a 0-

Ph Ph p- or jf- Hydrogen Abstraction —------> Fragmentation

or y-hydrogen might well result in bond fragmentation, since the diradical would not be able to undergo ring formation. Conversely, abstraction of the 6 -hydrogen would give the cyclized product 83.’ For this

O Bond Rotation ------» 83 52

25 26

27 20 24 23 28

FIGURE 1.

Side and Top Views of endo. endo-2,6 -diphenacyl[4]peristy lane (81). 53 abstraction to occur, however, the sidechain must rotate inwardly into the cavity, Once this occurs, overlap of the carbonyl oxygen and 6 -hydrogen would be extensive. Presumably, the excited 4tate of the carbonyl group would possess sufficient energy for this rotation to occur, overcoming as it does tha severe nonbonded steric interac- i tions present. From model examination, this photochemical

! closure was viewed positively if abstration of the (5- and

i y -hydrogens was a slow process relative to the needed bond rotation. Experimentally, all attempts to achieve this transformation proved disappointing. Only slow degradation of 81 was observed. In light of this failure, efforts were channeled to ! another photochemical pathway, one dealing with the intra- l molecular [2+2] photocyclization of diene 94. The volatile

i , 45 diene was prepared using chemistry reported by Grieco.

OH OH H 1. o-N02PhSeCN I n-Bu,P J > + 2 . H202

87 04

Diol 87 was treated witjh o-nitrophenylselenocyanate and tri-n-butyl phosphine to give the diselenide which was 54 directly oxidized and eliminated to give 93 and 94. Diene 94 could be obtained as the sole isolated product using longer reaction times (53%). Would this diene undergo intramolecular [2+2] photocyclization to give cyclobutane 95?^®

[2+3 ■2 H,

94 95 2

The resulting disecodecahedrane was envisioned to be capable of two-fold dehydrogenatioji to give 2 under the conditions now utilized to prepare dodecahedrane from ji secododecahedrane (Chapter 4). The proximity of the hydrogens to be removed for carbon-carbon bond formation is i very similar to that encountered in secododecahedrane. Unfortunately, the prerequisite photochemical transforma tion was unsuccessful under many different conditions of sensitization. Often, recovered starting material was the experimental result. On the assumption that the lack of a chromophoric group in 94 was partly responsible for the unreactivity towards cycloaddition, functionality was introduced to 55 absorb the light energy and facilitate the [2+ 2 ] transfor mation. To this end, 94 was photolyzed in the presence of phenyl benzeneselenosulfonate to give the addition product which was directly oxidized to furnish the solid bis-

A *7 phenylsulfone 96. The terminal positioning of the

H 1. PhSO-SePh, hv Iffii-) Decomposition

phenylsulfonyl groups was established by lH NMR analysis. Unfortunately, the photolysis of 96 gave no new product; only fragmentation appeared to be operational. The negative results recorded so far both in capping the [4]peristylane framework and in forming a bond to the methano bridge may arise from nonbonded steric interations. In reactions such as the acyloin condensation of 65, coupling of 70, and photocyclization of 94 and 96, decom position of the starting material was frequently observed. 56

Bond Rotation. I

If a carbon-carbon bond is to be formed, the reactive functionalities contained within the sidechains must approach each other. However, this often requires bond rotation in the manner illustrated above and development of steric interaction between X and the endo hydrogen of the methano bridge of the [4]peristylane system. Whereas model studies may predict this rotation to proceed because of the flexibility of the methano bridges to swing outward (il lustrated below), all experimental results indicate this rotation not to be feasible. 57 In response to these experimental failures, it was deemed necessary to develop a new synthetic strategy if arrival at decahedrane was to be successful.

III-5 SYNTHESIS OF a-DIKETONE 99

The new synthetic pathway is shown in Scheme 6 .

5 99 98 Arrival at 2 from 62 was outlined earlier, but arrival at 62 was now to result from ring expansion of 97 presumably via diazo chemistry. Bis-dione 97 was believed to be accessible via cleavage of vicinal diol 98 using standard procedures. Photolysis of a-diketone 99 to give 98 was considered promising due to the close proximity of the carbonyl oxygens to those hydrogens that were to be abstracted. The synthesis of 99 originated with dione 5. Wittig 48 reaction of 5 using the salt-free method of Still gave bis-methylene hydrocarbon 100 (90%). Treatment of 100 with the borane-tetrahydrofuran complex and oxidative workup

OH HO

100 101

Jones /reagent c h 3 o 2 c c o 2c h 3 H OzC c o 2 h

CH-1 f- OBU benzene 103 102 59

gave the polar dlol 1 0 1 as a viscous oil which was directly oxidized with Jones reagent to the Insoluble dlacld 102 (51%). Again, the hydrogen has been delivered from the exo face forcing the functional group Into the cavity of the [4]perlstylane framework. The dlacld was esterlfled using 49 the procedure of Ono. Obviously, to arrive at a-diketone 99 requires that an acyloin condensation be successful. The difference between dlesters 103 and 65 Is the lack of the methylene groups between the [4]perlstylane framework and the ester moieties In 103. The Importance of this apparently slight change Is shown In studying the models. In 103, the ester groups have no ability to exist outside of the cavity, but Instead are held in very close proximity to each other. This arrangement of the endo ester moieties was expected to eliminate the difficulties caused by the nonbonded steric

p M e Q S' interactions seen so many times before, thus allowing the capping sequence to proceed. Indeed, this slight structural change proved substan tial in the ease of acyloin condensation. Treatment of 103 under classical conditions and subsequent oxidation gave the yellow, UV active a-diketone 99 in yields as high as

67%.

CO 2CH 3 ,0 CH 3 O2C 1. TMSCl, Na toluene,

103 99

This transformation gave the first and only carbon capped

[4 ]peristylane to date and excitement for possible success ful arrival at 2 was building. Photolysis of 99 now required Y-hydrogen abstraction and ring closure to give diol 98. Removal of the hydroxyl 61 groups would afford another structurally Interesting hydrocarbon 104. Molecular models of these compounds do not show an excessive amount of strain. Thus, the syn thesis of 104 was believed to be eminently feasible. Also, further chemical manipulation of 98 might well lead to decahedrane as shown. Cleavage of vicinal diol 98 would give bis-cyclobutanone 97. Ring expansion would most likely result in the formation of the .two symmetric regioisomers 62 and 105, both of which could be carried onward to 2 .

HO OH

Pb(OAc)

98 97 J ring ^expansion

2 62 105 62 Experimentally, the photolysis of 99 has not to date given rise to the desired diol 98, although a-diketones are known to undergo Norrish II ring closure to furnish a- hydroxy cyclobutanones 50

Photolysis of 99 using a 450 W mercury lamp in benzene and benzene-acetone-tert-butyl alcohol solvent systems resulted in photoreduction as analyzed by NMR.

99 106

Irradiation in a pentane-methylene chloride solution using a sunlamp has afforded a product of unknown structure. The mass spectrum indicates incorporation of a pentyl unit into the compound. IK and 13C NMR spectra show no carbonyl group to be present and off-resonance 1SC NMR studies 63 reveal the presence of three methyl groups and two quater nary oleflnic carbons. Obviously, extensive structural rearrangement has occurred. The product is a solid but no x-ray quality crystals have been grown to date. Conse quently, the structure remains unknown. Although photolysis of 99 has not given any products resembling 98, much additional experimentation is certainly warranted. If this route for entry to both 2 and 104 remains unworkable, other schemes may prove applicable, a few of which are presented in section III-6 .

III-6 OUTLOOK FOR THE FUTURE

There exists, obviously, a wide variety of methods to arrive at decahedrane. It is impractical to discuss all the possibilities, but a route using intermediates from previous work is shown in Scheme 7. 64

Scheme 7

HO'X ?°2H C H a C 9CH3 O H C 9H0 H

102 107 108

Arrival at diketone 107 may provide the opportunity for quite exciting chemistry. This compound should be acces sible from a number of routes, two of which are illustra ted. Treatment of diacid 102 with excess methyllithium and appropriate quench should give 107 if addition of a second equivalent of methyllithium can be accomplished. Should this route prove unsuccessful, the second alternative is assumed to work. Treatment of diol 101 with PCC should give dialdehyde 108. Photolysis of 108 would very likely result in decarbonylation as seen previously, but addition of methyllithium and oxidation would afford 107. Photoly- tic ring formation as attempted before may end with favor able results in this case. Due to the steric bulk of the methyl groups, the compound is assumed to exist preferably in conformation A where the carbonyl groups are thrust into the cavity and the methyl groups out away from the 65

[4]peristylane framework. Thus, photolysis of 107 will likely result in successful Norrish cyclization giving 109.

Dehydration would give rise to diene 110 and ozonolysis with reductive workup would afford the desired bis-cyclobutanone 97. Continuation of Scheme 6 could eventuate in a synthesis of 2 but without a doubt some modifications will be necessary at the bench. Two additional approaches can also be devised to arrive at 104 using common intermediates. Formation of a carbon- carbon bond connecting the carbonyl carbons would give 98 66

HO OH

i'H Mg(Hg)

97 98

NNHTs TsHNN 104 ••‘"H NH-NHTs mm jfc m W ^

99 111

and removal of the hydroxyl groups would afford 104. a- Dlketone 99 Is also a potential precursor. Formation of bis-tosylhydrazone 1 1 1 , subsequent carbene generation, and insertion may give the hydrocarbon. Advances toward 2 and 104 all but ended when the dodecahedrane project was taken up by this investigator. The vast amount of time required did not allow for concurrent progress on decahedrane, but the advances that have been made to this date provide a valuable background on the chemistry, leaving potentially successful routes for another synthetic chemist to attempt. CHAPTER IV

Routes to Substituted Dodecahedranes

IV-1 INTRODUCTION

The Platonic solid dodecahedrane has been and remains of great aesthetic and synthetic interest. The Paquette group succeeded in synthesizing the hydrocarbon in 3 7 a 51 1982. ' Recent improvements now allow amounts of dodecahedrane to be prepared adequate for considerable further chemical investigation. With these synthetic improvements, many theoretical and synthetic questions have been and continue to be answered and reporting of these data constitutes this section of the thesis. Many approaches toward dodecahedrane have been studied 52 in the past and are summarized in relevant dissertations. Until 1987, only the Ohio State workers had developed a synthesis of this much sought-after molecule. Recently, Prinzbach and co-workers reported a rather inefficient 53 method for preparing dodecahedrane (3). Their route involves a gas-phase isomerization of [l.l.l.l]pagodane

67 68

along(1 1 2 ) to give dodecahedrane in a maximum yield of 8%. Despite the low yield, this result is impressive.

It was the goal of this researcher to study the dodecahedryl system in the hopes of preparing functionalized derivatives. The development of a new dehydrogenation method has provided a more efficient means for converting secododecahedrane to dodecahedrane. This protocol together with additional advancements will be detailed in Sections IV-4 along with attempts at converting functionally adaptable secododecahedranes into the corresponding dodecahedrane derivatives. The surpringly simple direct functionalization of the parent hydrocarbon will be discussed in Sections IV-5 and IV-6 , with the latter outlining the synthesis of a number of amino derivatives prepared to evaluate biological activity. Section IV-7 will present and discuss NMR data of the various deriva tives and IV-8 will review initial findings in new chemis try and routes to new "second generation" molecules. The 69 route to the hydrocarbon and building of the framework will be summarized in Section IV-2. Section IV-3 will continue the synthesis toward 3 while reviewing both the original route and an alternative sequence developed and used by this author.

IV-2 REVIEW OF THE ROUTE TO ALDEHYDE 135

The strategy for construction of the dodecahedryl framework starts from cyclopentadienide anion (113), an inexpensive source of cyclopentane units. Oxidative

MeOOC-CsC- COOMt

113 114

COOCHj 115 116 70 coupling of 113 with iodine to 9,10-dihydrofulvalene, followed by a domino Diels-Alder reaction with dimethyl 54 acetylenedicarboxylate yielded diesters 115 and 116. The desired adduct 115, readily separated by selective saponi fication of the more rapidly hydrolyzed isomer 116, was obtained in yields of 10 to 15%. Although the isolated yields are low, the molecular complexity achieved in this single operation outweighs the disadvantages. Diester 115 contains fourteen of the required twenty carbons in the form of four cis, syn-fused five-membered rings resulting in six cis-locked methine hydrogens.

COOCHj 0 115 117 118

N oOCH3, c h 3o h

OOCH. C00CH C00CH COOCHJCOO 0+5 121 120 118 71 In continuation of the synthesis, hydrolysis of 115 gave diacid 117, which when subjected to iodolactonization

conditions, afforded 118 clearly. The lactone rings in 118 were easily opened in methanol containing a catalytic amount of sodium methoxide to give 119. Jones oxidation

produced 1 2 0 , and reductive elimination of the iodine atoms furnished diketo diester 121 (65% overall). Introduction of the six remaining carbon atoms is shown

in Scheme 8 and starts by bis-spiroannulation of 121 with diphenylcyclopropylsulfonium ylide. Baeyer-Villiger oxi dation of 122 followed by treatment of bis-lactone 123 with phosphorous pentoxide in methanesulfonic acid provided bis- enone 124. The overal yield of 124 from 121 was found to range from 20 to 33%. Subsequently, it was found that a change to the ethyl ester series causes yields to increase

to 50-55%.56 The dodecahedrane sphere begins to develop upon catalytic hydrogenation of 124. Cyanoborohydride reduction of 125 affords bis-lactone 126, which is transformed into dichloro diester 127 upon treatment with methanolic hydro gen chloride in the presence of an additional chloride ion source. This symmetrical approach benefits from the obvious advantage that the number of steps required is reduced by a 72

Scheme 8 0^ 0 C H S

c«30_o 9 NaCNBHi s

126 125 124 Ha , CHjOH R T . u a CHjOv^O ci

3 J 6 %

127 factor of two. Thus the procedure has allowed for fast and efficient building of a molecule which contains the required carbons and functionality adaptable to progressing toward dodecahedrane. Unfortunately, deviation from this symmetrical approach becomes necessary at this point. Dissolving metal reduction of dichloro diester 127 using six equivalents of lithium in liquid ammonia at -78 °C results in the formation of dianion 1285?, which upon addition of one equivalent of chloromethyl phenyl ether gives tetraseco keto ester 129 and the transannular hydroxy 73 ester 130. Separation of the products using MPLC reprodu- cibly afforded 129 in 50% yield.

CHjO^O

128

PhOCHjCl 1 ch3°vT .O

PhOCH? 120

Homo-Norrish photocyclization of 129 using a 450 W mercury lamp in a benzene-tert-butyl alcohol (4:1) solvent system furnished alcohol 131. Dehydration to 132 was achieved by refluxing a benzene solution of 131 in the presence of a catalytic amount of p-toluenesulfonic acid. Subsequent diimide reduction afforded diseco ester 133 (overall yield of 85%). 74

PhOCHj/j^O A I > PhOCHjJ^'' PhOCHjjJjfv^P r hv (TaOH) C.H., a '

1 2 9 131

i RsNNH., B.Oi PhOCH2<1 CH2PH P h O C H ^ rcc (lBu),A1H CIljClj C.H.“

135

With construction of the front portion of the dodecahedryl framework complete, we now focused on stitching up the top. Alteration in the oxidation level of the carbomethoxy group was now necessary. To this end, 133 was reduced with excess diisobutylaluminum hydride and the resulting alcohol (134) was subjected to buffered PCC oxidation. Aldehyde 135 was isolated in 90% yield for the two steps. The chemistry deployed to arrive at 135 has followed the elegant work of previous group members. The sequence proceeded quickly and uneventfully, thereby setting the 75 stage for those subsequent bond-forming reactions which will be discussed in the following section.

IV-3 CONTINUING TOWARD DODECAHEDRANE

With triseco aldehyde 135 in hand, we were now ready to proceed toward the desired dodecahedrane derivatives. Photolytic pathways were available for advancing to 136,

PhOCH2

hv v jJjcj + decarbonylation -78*C \ vt / / products

135 136

but as cited in chapter 3, the photolysis of aldehydes often results in decarbonylation. Indeed, photofrag mentation was the major pathway. However, decarbonylation could be controlled using low temperatures. Irradiation of the aldehyde in toluene-ethanol (9:1) solution with circu lation of cold methanol (-78 °C) through a jacketed pyrex photolysis vessel permitted yields as high as 43% to be achieved. The endo and exo isomers 136 could be separated by chromatography; however, for practical purposes the combined isomers were directly utilized in the subsequent 76 reaction. At this point, branching from the original

synthesis was necessary in attempts to adopt functionalized seco derivatives for the construction of substituted dodecahedranes. In the historic synthesis (Scheme 9), the phenoxymethyl sidechain was removed at this stage. Birch reduction of phenyl ethers 136 and acidic hydrolysis furnished the diol mixture 137. This diol mixture underwent PCC oxidation to keto aldehyde 138 and excision of the aldehyde moiety was subsequently accomplished by treatment with potassium hydroxide in ethanol. Yields were found to be very low (20-40%), but sufficient amounts were obtained by Ternansky to complete the synthesis. Photocyclization of 139 pro ceeded without complication and gave seco alcohol 140. To increase the efficiency in the sidechain removal step, direct photolysis of 138 was attempted. As expected, irra diation not only resulted in decarbonylation but also in homo-Norrish cyclization to furnish directly alcohol 140. Acid-catalyzed dehydration gave 141 which was reduced to 142 with diimide. The crucial final step to give dodeca hedrane (3) was accomplished in 34% yield by dehydrogenating 142 with 10% palladium on carbon. Improvement of this step will be commented upon in Section IV-4. 77

Scheme 9

CHzOPh

137 138

KOH ICgHaOH (TaOH) C«H«, A

139

10* Pd-C * 250 C

142 3 78 For our purposes, it was decided to leave the sidechain intact and use it as a building block to the desired deri vatives. The pathway adopted is shown in Scheme 10. Buffered PCC oxidation of epimeric alcohols 136 delivered diseco ketone 143 whose subsequent photocyclization gave alcohol 144 in chromatographically purified yields of 72% for the two steps. Elimination of water from 144 under

Scheme 10

PCC 4

136 143 144

TsOH benzene

TsOH ♦ NalO benzene

147 145 146 basic conditions using phosphorous oxychloride in pyridine- benzene gave seco olefin 145. Attempted acidic dehydration of 144 resulted in the formation of cyclic ether 146 via generation of the carbocation and subsequent intramolecular CO Friedel-Crafts electrophilic acylation. Treatment of olefin 145 with tosic acid also provided 146. Diimide redaction of 145, generated from hydrazine hydrate and sodium periodate, gave phenoxymethyl secododecahedrane 147 in 91% yield from 144 after recrystallization from acetone. We were now ready to investigate the dehydro-genation step in order to arrive at a monosubstituted dodecahedrane. But first, a review of and improvements in the dehydrogenation procedure need to be presented and discussed.

IV-4 DEHYDROGENATION OF SECODECAHEDRANE DERIVATIVES

The first dodecahedrane ever isolated was 1,16- dimethyldodecahedrane (149). By means of an acid-catalyzed isomerization of olefin 148 involving methyl migration, the 59 most highly symmetrical dimethyl derivative was obtained.

CF3 SO3 H CH2CI2 -4 25* C

ch3 148 149

This procedure was found not to be suitable for the syn thesis of methyldodecahedrane. The alternative successful method discovered involved dehydrogenation (10% Pd/C ).60 80 This method was utilized in the synthesis of the parent

spherical hydrocarbon. The pivotal step in the original synthesis consisted of heating an intimate mixture of secododecahedrane (142) and a fifty-fold amount of 10% palladium on carbon presaturated with hydrogen at 250 °C for several hours.

142 3

The undesirable features of this otherwise remarkable transannular bond formation are quite obvious and unsatis factory if large amounts of dodecahedrane are to be made. The efficiency of this step is very low (30% after purifi cation) , but, more importantly, a severe limitation on throughput (maximum of 2 mg of 142 per run) and intolerance of pendant groups were experienced. Obviously, an increase in yield would be welcome, but the scale limitation that restricted the amount of dodecahedrane available for further study had to remedied. Recently, the Paquette group developed a superior, preparatively useful 81 dehydrogenation system which increases both the yield and scale of the reaction.

The removal of pendant groups encountered with the original procedure stems in large part from the presence of excess hydrogen, which was necessary to curtail the pro pensity of Pd(0) for 1,2-dehydrogenative elimination. The Pd/C mixture accomplished the cyclization reaction, but without excess hydrogen present the palladium continued to dehydrogenate the initial product. Many products resulted, some of which appeared to contain sites of unsaturation. In the presence of hydrogen, this destructive process was suppressed (GC analysis), delivering dodecahedrane as the only recognizable product. Unfortunately, this excess hydrogen also facilitated the hydrogenolysis of various sidechains. The optimal alternative conditions were realized to be those wherein one transition metal would effect smooth transannular carbon-carbon bond formation in the presence of a second metal capable of absorbing the liberated hydrogen. Zerovalent titanium is well recognized to be amenable to conversion to a dihydride alloy stable to above 600 °C, thus fulfilling this latter requirement.6* Treatment of an intimate mixture of 142, 5% platinum on alumina, and titanium metal at 250 °C for 10 h furnished dodecahedrane in a crude yield of 80%. After removal of 82 by-products through trituration with ether, pure 3 was isolated in yields ranging from 45-50%. The increase in yield was very encouraging but, more importantly, the scale problem encountered with the Pd/C method was greatly ameliorated. Simple recourse to large tubes allowed for scale-up. Experimental details follow: the seco derivative was mixed with the Pt/Al2Oa catalyst and added to a glass tube containing titanium metal, which had to be kept under argon at all times. The vessel was evacuated and sealed under high vacuum. The tube was shaken to mix the components and immersed in a hot bath (210 °C) for the appropriate time. The tubes were opened and the product was eluted with benzene. One point needs to be stressed. Although the scale had been increased in this procedure, the necessity to keep the titanium metal covered with an inert atmosphere did set limits on the scale. Also, more care was necessary to guarantee against inactivation of the metal by exposing it to air. Titanium metal rapidly acquires an oxide coating when exposed to the atmosphere. In similar fashion, the dimethyl derivative 151 was 58 prepared in an isolated yield of 75%. Importantly, dealkylation was observed concurrently giving methyldodecahedrane (152) to the extent of 10-15%. This finding was of particular concern to us because this dealkylation 83 process was certain to jeopardize our plans for crafting substituted dodecahedranes from functionalized seco derivatives. Obviously, this reductive cleavage had to be controlled if this was to be a viable method for the preparation of monosubstituted dodecahedranes.

5* Pt-AljjOj T1(0) 200*C

152

Application of the newly developed methodology to phenoxymethyl secododecahedrane 147 gave methyldodecahedrane (152) in low yield. Although no attempt was made to

CHjOPh

5% Pt-Al202 T1(0)

147 152

optimize this reaction, it was already apparent that this reductive cleavage pathway was a preferred route. The need 84 for a suitable sidechain that would not undergo this

cleavage was sought. The phenyl group was removed by submission of 147 to Birch reduction and subsequent acidic hydrolysis of the dienol ether. When the hydroxymethyl derivative 153 (92%) was submitted to the dehydrogenation conditions (7 h),

CH 2OPh c h 2o h

1. Na.NH, St Pt-Al-0, EtOH, THf TU°? 3, 2. 3H HC1 210 C

147 153

removal of the entire sidechain occurred to give dodecahe drane (3). Stopping the reaction after 4 h gave mixtures of 3 and 142, indicating that the sidechain is removed prior to carbon-carbon bond formation. This reaction has now become the preferred method for preparing the parent hydrocarbon. Yields after purification average from 50-55% to as high as 65%. The overall yield from the epimeric alcohols 136 to dodecahedrane (3) was 35%, dramatically increased over the historic procedure. Altering of the sidechain oxidation state and subse quent dehydrogenation were attempted in search of a suitable derivative (Scheme 11). 153 154

Jones \ Jones 5V P t-A l 2 03 reag en t reagent \ T i( 0 ) -20*C 4 -4 5 "C 210*C*

155

Buffered PCC oxidation of 153 proceeded uneventfully to give aldehyde 154 in 92% yield. Keating in the presence of the Pt and Ti metals expectantly gave the parent hydrocarbon. Acid 155 was also sought. This seco derivative was found to be surprisingly difficult to synthesize. Jones oxidation of 153 at room temperature furnish a solid (79%), insoluble in most organic solvents, which was assumed to be 155. After the acquisition of mass spectral data, this assumption was proven incorrect. MS analysis showed a base peak of 320 (M+ of 155 - 306) along with higher molecular weight mass peaks. Apparently oxidation of the methine 86 carbon-hydrogen bonds had occurred in addition to the sidechain. The identity of the molecule has not been determined, although this result may shed light on additional properties of the system and prove to be quite interesting in its own right. Returning to the synthesis of acid 155, Corey has shown PDC oxidation of alcohols in dimethylformamide as solvent 62 results in the isolation of acids. This method, when applied to 153, furnished solely the over-oxidation pro duct. Thus oxidation of aldehyde 154 to 155 was studied.

Treatment of 154 using standard procedures (NaCIO,,63 Ca(OCl)j64) only returned starting material. In light of these failures, we returned to using Jones reagent. It was thought that cooling the reaction mixture may result in slowing or possibly eliminating this excessive oxidation. Complicating the procedure experimentally is the fact that at lower temperatures both 153 and 154 are guite insoluble in acetone, the typical solvent for Jones reac tions. Since the solubility of the aldehyde was greater than that of the alcohol, it was studied first. The solubility difficulty was overcome by suspending 154 in acetone and cooling to -45 eC followed by the addition of tetrahydrofuran until complete dissolution occurred. Treatment for 1 h with excess Jones reagent and quenching at low temperature gave acid 155 in high yield (85%) after 87 workup. Ho over-oxidation product was obtained under these conditions. Attempts at converting alcohol 153 to 155 were not as successful. For complete dissolution of 153 in the co-solvent system, temperatures of -20 °C were necessary. Oxidation at this temperature often resulted in obtaining mixtures of all the possible products. Thus the two-step procedure from 153 to 155 was preferred. As seen before, the carboxy group was removed under the dehydrogenation conditions. With these intermediates available, we next sought slightly more sophisticated com pounds which might survive conditions. Since alkyl groups remain largely unaffected in the dehydrogenation step, adoption of a functionalized carbon substituent may give rise to the corresponding dodecahe drane derivatives. The molecule of choice was vinylsecododecahedrane 156. Wittig reaction of 154 readily gave 156 in high yield (85%) after recrystalization from acetone.

KN(TMS)

104 150 157 88 It was known from initial experiments that the platinum solely acts as the dehydrogenating agent with no participa tion by the titanium. Hydrogen, of course, is liberated in this process with the titanium being present to absorb the gas in order to minimize the hydrogenolytic cleavage of the pendant group. In the derivatives mentioned before (153, 154, 155), it is likely that the platinum preferentially removes the hydrogens from the sidechain functionality liberating carbon monoxide (153, 154) and carbon dioxide (155). Removal of the sidechain hydrogens in 156 was expected to be a slower process, allowing for possible success. Submission of 156 to the now standard conditions resulted in little dodecahedrane production (<10% as determined by GC). The primary product was not vinyl- dodecahedrane (157), but ethyldodecahedrane (158) [300 MHz lH NMR, CDClj, 6 3.37 (s, 16 H), 3.00-2.92 (m, 3 H), 1.38 (q, J = 7.28 HZ, 2 H), 0.84 (t, J = 7.28 Hz, 3 H)].

C H -C H j

h 2c h 3

156 3 158 89 Dehydrogenation to form the dodecahedrane framework occurred, but the evolving hydrogen was donated to the vinyl group. Separation of products from this disap pointing reaction and further characterization of 158 was not attempted. Through this result, a greater understanding of this remarkable process was gained. Alkyl groups have proven capable of withstanding the drastic conditions, while sidechains containing a heteroatom cannot. Most impor tantly, this experiment shows that the titanium metal cannot absorb the evolving hydrogen quickly enough to prevent destructive side reactions. Dehydrogenation of 156 was reinvestigated, but this time without the Ti(0) pre sent. As indicated by capillary gas chromatography, similiar ratios of 158 and 3 resulted. Due to these results, which show the titanium to be ineffective in its stated role and unneeded for the reaction to proceed, it was removed from all subsequent dehydrogenation procedures. This removal of Ti(0) allows for further scale-up in these reactions to prepare dodecahedrane. Also, the physical implementation of the apparatus is faster, since the need for an inert atmosphere was no longer important. Various attempts to dehydrogenate the many seco deri vatives were unsuccessful (solvents, pressure tubes, temperatures, hydrogen absorbers). The last-ditch attempt 90 involved the allylic ester 159 which was prepared from 155 using DBU and allyl bromide (88%).49

P02H

DBU ch2»chch2bI;

155 159 160

It was hoped that the allyl group in 159 would absorb the hydrogen liberated. Thus the unsaturation would act as its own "hydrogen sponge" giving the n-propyl ester 160. There is one obvious question with the compound; will the oxygen functionality survive the drastic conditions? The answer is no. Submission to standard conditions succeeded in fur nishing dodecahedrane (3) as the sole observable product. In light of the failures to this approach, efforts turned to functionalizing the parent hydrocarbon directly with the hope of selectively producing a monosubstituted derivative. This approach is covered in the next section. 91 IV-5 DIRECT FUNCTIONALIZATION OF DODECAHEDRANE. FORMATION OF MONOSUBSTITUTED DERIVATIVES.

As Illustrated in Section IV-4, the inroads to sub stituted dodecahedranes from the corresponding secododecahedranes proved to be unworkable. Thus we turned to the direct functionalization of the parent hydrocarbon. An obvious model for this work was the extensively studied chemistry of the adamantane system. As we proceeded to investigate the chemistry of the dodecahedryl system, observations were frequently made that reactions successful with adamantane and its derivatives failed time and time again. In retrospect, one can conclude that the chemistry of the adamantyl and dodecahedryl systems only parallel each other in the loosest sense. One aspect of adamantane chemistry that did hold true for our system was bromination of the parent hydrocarbon. There are two major reaction pathways followed in direct functionalization of adamantane: radical and ionic reactions. Whereas radical chemistry often affords mixtures of multifunctionalized products, ionic processes have shown impressive selectivity, with control of

functionalized products being easily achieved.65 Treatment of adamantane (161) with neat bromine at the reflux temperature produces the monobromide 162 in quantitative Br,, reflux — fc— ■---- ►

161 162

yield.This reaction is catalyzed by Lewis acids, thereby indicating its ionic nature. With the proper choice of catalyst and conditions, one to four bromines can be introduced sequentially into the molecule. Substitution results only at the tertiary carbon atom and each bromine is in turn more difficult to introduce than the last. One major reason attributable to this remarkable selectivity is the inductive effect that the bromine atom exerts on the other positions. With the second reactive site only two carbons away, the electron-withdrawing power of the bromine atom lessens the likelihood of electrophilic attack and attendant formation of a carbocationic center. This effect is obviously very pronounced in the adamantane case but can it be expected to control the substitution of dodecahe drane? One might imagine that, after initial bromination to give 163, secondary reactions could be possible. An inductive effect would undoubtedly be exhibited at all of 93 the remaining unsubstituted centers, but the position

farthest from the substituted carbon is five atoms removed. The effect would be proportionately lessened. Obviously,

3 163

nightmarish mixtures would present insurmountable difficul ties in product isolation. Nevertheless, we proceeded forward. To our delight, stirring of 3 overnight in neat bromine at room temperature furnished only the product 163 as indicated by modern analytical techniques. The yield in this preparation of the monobromide was quantitative. Repetition of this procedure has never resulted in the observance of other products. Apparently, the electron- withdrawing ability of the bromine atom is sufficient to restrict further electrophilic attack on another center in 163. The proton and carbon NMR spectra are shown (Figure

2 ) along with the carbon-hydrogen correlation spectrum (Figure 3). Interpretion and discussion of the NMR data for 163 and other derivatives is deferred to Section IV-7. 94

iir n n r rs Y ¥

3. SO 3.00 t r I \ T 100 90 80 79 60 PPM

FIGURE 2. *H (300 MHz) and 1SC (75 MHz) NMR Spectra of 163 95

• i

mVl

0vk

S j

SJ LA

#»m m01 m*

FIGURE 3. Carbon-Hydrogen Correlation Spectrum of 163 96

B r OH

163 164

Following the precedence set by bromoadamantane, we sought to replace the bromine atom in 163 with a hydroxyl group to give 164. This was not realized. Refluxing 163 in water-tetrahydrofuran solution in the presence of ge potassium carbonate only returned the bromide. Also, the addition of silver nitrate to accomplish a silver ion- 66 assisted hydrolysis also resulted in no change. Recourse to refluxing 163 in dimethylformamide with concentrated hydrochloric acid only furnished complicated mixtures, as indicated by capillary gas chromatography. These experi ments begin to show the difference between the adamantyl and dodecahedryl systems. With this inability to prepare 164 from 163, we turned again to the direct functionalization of the parent hydro carbon. Treatment of 3 with excess lead tetraacetate in methylene chloride and trifluoroacetic acid furnished 67 68 trifluoroacetate 165. ' Direct hydrolysis with aqueous 97 sodium hydroxide gave 164 in 76% yield. This radical reaction (3 -* 165) often proceeded erratically, giving products assumed to arise from polyfunctionalization.

Pb(OAc) CF,C0,H benzene r e f lu x

Direct oxidation of the unactivated tertiary carbon- hydrogen bonds in the parent was also achieved with limited success using m-CP&A 69 as indicated by GC analysis. Unfortunately, this reaction proceeded very slowly and, as expected from radical reactions, many products were present after the long reaction times. With sufficient amounts of 163 and 164 in hand, we again returned to precedented adamantane chemistry to make additional monosubstituted derivatives. Treatment of the adamantyl derivatives with protic acids in the presence of a nucleophile resulted in the preparation of many new derivatives.®®'^0 Adaptation of this protocol to the dodecahedrane system proved unsatisfactory (Scheme 12). Treatment of 163 with sulfuric acid in acetonitrile in an effort to induce a Ritter reaction and deliver acetamide 98

Scheme 12

O Br N H C C H 3 CR,CR w.J— i 2. H20

163 166

OH N3 a NaN,

164 167

,c o 2 h

3 O R 163 KochyHaaf,

1 68 166 only returned starting material. An attempt to convert alcohol 164 to azide 167 produced the same negative results. Also, attempted Koch-Haaf carboxylation on both the bromide 163 and hydrocarbon 3 resulted only in recovery 66 71 of starting material. ' Obviously, these standard conditions were not sufficiently forcing to regenerate the carbocation. Presumably, the 1-adamantyl cation is stabilized through hyperconjugation with the carbon-carbon bond which is positioned antiperiplanar to the electron deficient orbital. The dodecahedryl cation has no carbon- carbon bond properly positioned for this stabiliza-tion effect to operate. In light of these experimental failures, we turned to more severe conditions to generate the cation using Lewis acids. Dissolving bromide 163 in benzene and treatment

Br Cl Ph-H FeCI,

163 169 170

with a large excess of ferric trichloride for 15 min afforded a mixture of 169 and 170 as indicated by GC. To ensure complete electrophilic aromatic substitution, the 100 reaction mixture was slightly warmed and stirred for 3 h; 170 was obtained quantitatively. This positive result encouraged further use of Lewis acids (initially all chloride-containing). Unfortunately, under circumstances where unreactive nucleophiles were involved or where the nucleophile could not be used as solvent, chloride 169 was observed as the sole product. The chloride could be isolated in high yield from a halogen exchange process involving exposure of 163 to ferric trichloride in an inert solvent. Surprisingly, the use of moist ferric trichloride gave no alcohol 164; only chloride 169 was recovered. At longer reaction times and with stronger Lewis acids (AlBra, AlCla), decomposition was the chosen pathway and no trap ping of various nucleophiles was observed. This result is exemplified by the reaction of 163 with trimethylsilyl cyanide under many different reaction variants. Literature reports show that bridgehead halides can be converted to the corresponding cyanides upon treatment with tin tetra-

v TMSCN

SnCl 4

163 171 101 72 chloride as the Lewis acid. Application of these conditions to 163 resulted only in conversion to chloride

169. The alternative use of aluminum tribromide decomposed the starting material even in the presence of added potassium cyanide as a supplemental cyanide ion source. Another approach was to generate the dodecahedryl carbocation with a strong Lewis acid followed by addition to a solution containing an exces of some nucleophile. The obvious choice of trapping agent was methanol to give 172.

Br 1. AlBr3 2. CH30H Decomposition

163

Due to the relative instability of the cation, short reation time and quick trapping were deemed necessary. Experimentally, treatment of bromide 163 with aluminum tribromide at room temperature for 2 min followed by pouring into a suspension of potassium carbonate in methanol afforded no products consistent with 172. Only decomposition was observed. This attempt illustrates the instability of the dodecahedryl cation at ambient tempera tures in the presence of strong Lewis acids. 102 This brings up an important point. For years,workers have attempted to isomerize Ca0H 20 hydrocarbons to dodeca hedrane using Lewis acids. However, all attempts have been

i i o m e r l z e

173

73 unsuccessful. Can the cation be formed at low tempera ture and trapped successfully? Olah has reported an elegant procedure for the generation of carbonium ions from alkanes (for NMR data) at low temperatures where they are stable, with subsequent quenching to prepare methoxy 74 derivatives. The procedure involves treating the hydrocarbon with magic acid and pouring the resulting

1. Super Acid > r 3c-OCH3 3 2. CH30H, K2C03 103 cation solution onto cold methanol. The cations have also been generated from derivatives Including halogen com pounds. Thus we wanted to adapt this chemistry to the dodecahedrane system. Bromododecahedrane was chosen as precursor to the cation instead of the hydrocarbon because of its better solubility characteristics. Dissolution of 163 in chloroform and sulfuryl chloride fluoride at -78 °C containing fluorosulfonic acid and antimony pentafluoride

Br 1. CH2C12, FS03H, $bF5, OCH- S02C1F, -78'C 2. CH3OH, NaOCH3 *

163 172

resulted in the immediate formation of a yellow solution. This mixture was stirred for 30 min and poured into an equally coldsolution of methanol containing sodium methoxide. Methoxydodecahedrane was isolated in 61% yield. As one would expect, the absence of antimony pentafluoride returned 163; also, sulfuryl chloride fluoride could be removed without affecting the transformation. This success provided great hope for some interesting chemistry. This magic acid cation generation protocol held promise because of the experimental difficulties seen in preparing 104 a carbon functional!zed dodecahedrane by conventional procedures. Carboxylic acids have been prepared by reaction of carbon monoxide and magic acid solutions of 75 saturated hydrocarbons followed by appropriate workup. In an adaptation of this precedent, we prepared carbomethoxydodecahedrane. Cation formation was accomplished as previously described. Bubbling carbon monoxide through

H ,C+

163 174

| h 2o

ch2n2

175 168

the solution trapped the cation as the acylium ion 174. Pouring this solution into water afforded the acid 168. For purification purposes, this acid was converted to ester 175 with excess diazomethane. Yields of 175 approach 60% 105 after chromatography. The NMR data are Included on the following pages (Figures 4 and 5).

With the successful preparation of this long-sought carbon-functionali2ed dodecahedrane, its conversion to other derivatives was initiated. Selected transformations will be discussed in Section IV-6 where pursuit of the amino derivatives is described. In our continuing study of the dodecahedrane system, we wanted to obtain x-ray structure analyses of various deri vatives to determine if any structural deformations result upon functionalization. The growth of suitable x-ray quality crystals has been quite difficult. A property of dodecahedrane and its derivatives is the fact that twinning of crystals often results. The x-ray structure determina- tion of the hydrocarbon resulted from a twinned crystal 76 but analyzing twinned crystals cannot normally be achieved due to their complex diffraction patterns. Crystals of bromide 163 and phenyl derivative 170 have been grown but were found to be badly twinned and to date analyses have not been accomplished. Fortunately, a single crystal of ester 175 was grown and its crystallographic analysis has been successfully completed (Figure 6). Attempts to grow suitable crystals of various derivatives continue and the results will be reported at a later date. 106 V

■T,-TTIjTM,|TTf|jMT,|1TTI|M- 1 I -r T i 1 i ' i 1 r 4 « 00 1 50 5 03 LBB *■ BP 70 60 5«

FIGURE 4. XH (300 MHz) and iaC (75 MHz) NMR Spectra of 175. 107

U u I* iM II# w

FIGURE 5 Carbon-Hydrogen Correlation Spectrum of 175. FIGURE 6. X-Ray Structure of Ester 175. 109 At this point, a return to silver ion-promoted chemis try was made in an attempt to uncover appropriate condi tions. As stated previously, attempted silver ion-induced hydrolysis of bromide 163 to 164 was unsuccessful. This result was originally surprising, but considerable experience has been gained since these initial attempts. In order to obtain trifluoroacetate 165 in a more reproducible manner, silver ion-promoted solvolysis was studied. Treatment of 163 dissolved in trifluoroacetic

AgOCOCF, cf3c o 2h

163 165 acid with silver trifluoroacetate afforded 165 in 80% yield. The key to this transformation appears to be the fact that both the silver salt and 163 are soluble in the solvent. Attempts at converting 163 into cyanide 171 using silver cyanide in various solvents were unsuccessful under all conditions, presumably due to this salts' low solu bility in the various media. If the silver salt is not available in solution, the reaction will not proceed. 110

Br .focV CH2Cf2 or ISOCYANIDE 163 171

To overcome this problem, a silver salt which is soluble in organic solvents was needed. The one that immediately came to mind was silver trifluoromethane- 77 sulfonate. The solubility of this salt in many organic solvents is extensive and this effect is shown by the successful transformation of bromide 163 into methoxydodecahedrane (172) and acetylaminododecahedrane (166) as shown in Scheme 13. To arrive at 172, bromide 163 was taken up in methanol and dichloromethane and treated with silver triflate. An unoptimized yield of 71% was obtained on small scale. The acetamide 166 which we had great hopes of converting into aminododecahedrane (Section IV-6) was prepared equally readily. The bromide was suspended in acetonitrile and treated with silver triflate. The mixture required refluxing, which allowed for the dissolution of 163 and Ill

Scheme 13

172

Br AgOTf

163

NHCCH

166 subsequent conversion to 166 after hydrolysis. Yields of this transformation were quantitative. Another product obtained in studying these silver ion- promoted reactions was fluoride 176. In this case, a

AgBF4

CH2 C12 1

163 176 112 halogen exchange resulted upon treatment of 163 with silver 78 tetrafluoroborate. Silver salts have been very useful in easily preparing compounds which have proven to be experimentally troublesome to obtain (165, 172) and others which were not possible to prepare by other means (166, 176). Bromododecahedrane was also converted into the pre viously known methyldodecahedrane (152).60 Treatment of

Br CH3 (CH3)3A1, hexane

163 152

163 with trimethylaluminum in hexane gave 152 in 90% yield.79 A very interesting transformation resulted in attempts to prepare azidododecahedrane 167 using literature proce- 80 dures. We had envisioned 167 to be a possible precursor to the amine. Treatment of 163 with tin tetrachloride in the presence of azidotrimethylsilane resulted in rapid conversion to a sole product isolated in 78% yield after purification. The IR spectrum contained a strong absorp tion at 2100 cm“* indicative of an azide. Hopes that the 113 conversion was a success to give 167 grew until further spectral data dashed those hopes. The proton NMR spectrum (Figure 7) was obviously not consistent with a monosubsti tuted dodecahedrane. Acquisition of the carbon spectrum

Br .N,

SnCl 163 167I

TMSN,

1 7 8 177

showed twelve signals indicating loss of the symmetry axis with only a plane of symmetry remaining. These spectra and mass spectral data showed 178 to be the structure. In explaining this unusual transformation, insight can be 81 gathered from work of H.C. Brown. Brown has shown that trialkylboranes coordinate to the azide followed by loss of nitrogen from the intermediate with subsequent or concurrent migration of the alkyl group from boron to 114

S B?

SNw

V k A w i l illii kdri l i k IkJIt

■ I i r i \ i i i i ) 'i i i t | i « i T 1" ~T~ T“ -i r- 3 *.0 5.5 3.a ?.5 90 00 ’0 ca 00

FIGURE 7. XH (300 MHz) and l3C (75 MHz) NMR Spectra of 178. 115

BFfc R-N\ ♦ R-N-R N£N fO i a BRi

nitrogen. Hydrolysis gives secondary amines. Since in the reaction of 163 an excess of tin tetrachloride is present,a similar analogy can be drawn. The tin reagent can

r \ ^SnCI 3 SnCI^ « ( ^ iN -Cl

167

177 116 coordinate to the azide with lose of chloride ion. Because the Lewis acid has no alkyl group to transfer, ring expansion to inline 177 occurs with loss of nitrogen as the driving force. This imine is assumed to be a highly 82 strained species and excess azidotrimethylsilane rapidly adds across the carbon-nitrogen double bond. Not surpri singly, attempts to convert 178 under basic conditions to a derivative such as 179, which could be converted into the parent homoazadodecahedrane (180), have failed, presumably 82 83 owing to this excessive strain. '

KOH LfAlH

1 7 8 179 180

Transformation of 178 to prepare a potentially useful compound is currently under study. The diamine 181 may

178 181 117 give additional insight in our attempts to determine pharmacological activity of the system.

In other developments (Scheme 14), dichlorocarbene, generated under phase transfer catalysis conditions, has

Scheme 14

CHC13 .NaOH CHCI b e n z e n e ^ B z M e .fa l

o n o 2 n o 2 * CH2 C12

184 183

been shown to insert into one of the carbon-hydrogen bonds to give dichloromethyl derivative 182, albeit in modest yield (42%).84 Interestingly, carbene insertion using 85 ethyl diazoacetate only returned starting material. Also, attempts to nitrate the hydrocarbon using nitronium QC tetrafluoroborate did not give rise to nitrododecahedrane (183); instead, the nitrate ester 184 was isolated in 84% 118 yield. Commercial nitronium tetrafluoroborate is known to contain significant amounts of nitric acid which will facilitate the synthesis of nitrate over nitro. However, the purified reagent furnished the same result. Obviously, on such small scale the total removal of nitric acid is not easily accomplished. As illustrated in this section, the direct functionalization of the hydrocarbon and acquisition of certain deri vatives has been successful. Since the synthesis of dodecahedrane in 1982, it has taken 5 years and considerable effort to perfect the procedure before this researcher was able to carry the chemistry one step further and prepare monofunctionalized derivatives. Ironically, adamantane chemistry was used as the model for this study but, in retrospect, the majority of reactions attempted on the dodecahedryl system failed even though they were successful with adamantane, its deriva tives, and other hydrocarbons which possess bridgehead positions. It is the opinion of this investigator that the chemistry exhibited by dodecahedrane only remotely follows the precedence set by admantane. 119 IV-6 SYNTHESIS OF AMINO DERIVATIVES

The reasons for preparing amino derivatives of dodecahedrane are two-fold. The first is for pharmacological interest, since similar spherical molecules are known to possess biological activity. Amantidine (l-aminoadamantane) has been shown to exhibit powerful antiviral activity 87 against certain strains of influenza and is also used in 88 the treatment of Parkinson's disease. Therapeutic results evolve from the high degree of lipid solubility, low extent of ionization, and lack of plasma-protein binding which ensures facile and rapid passage of the drug 89 into brain and cerebro-spinal fluids. It may be the capability of amantidine to penetrate tissues and fluids that underlies its efficacy; however, its symmetric shape also is of immense relevance. From the large number of saturated polycyclic amino compounds that have been prepared for antiviral evaluation, only a very few select 90 (highly symmetric) compounds exhibit useful activity. Owing to the symmetry that monosubstituted dodecahedrane derivatives possess, they may be more potent antiviral agents than 1-aminoadamantane itself. The second and equally important reason is the fact that the National Institutes of Health have funded this project since its inception in order to gain access to 120 these data. The millions of dollars this project has brought into the Paquette group over the years have not only allowed this work to proceed, but have also spawned many other projects. Without the preparation of an amino derivative, funding will surely be curtailed. Unfor tunately, as is well recognized in synthetic chemistry, a large quantity of material was needed to prepare the various compounds. These amounts could alternately have been used to study more chemically interesting phenomena associated with the system. Without further discussion, we envisioned the synthesis of four compounds (185 - 188) to evaluate the biological activity.

185 R= CH2NH2-HC1 188 186 nh2*hci 187 We wanted to start with derivatives already available in the synthesis of the seco derivatives, with the realiza tion that preparation of dodecahedrane derivatives would require new intermediates to be synthesized. Aldehyde 154 and acid 155 prove to be viable precursors to 185 and 186, respectively. The pathways are shown in Scheme 15.

Scheme 15 122 Addition of azidotrimethylsilane^1 to 154 furnished siloxy azide 189 which was reduced using lithium aluminum 92 hydride to give the free amine. Treatment of an etheral solution of the amine with ethanol saturated with hydrogen chloride afforded hydrochloride 185 in 65% overall yield after extensive purification. The amine hydrochloride 186 was obtained in a four-step procedure starting with acid 155. Formation of the mixed anhydride and addition of excess sodium azide gave the acyl azide. This compound was not isolated due to the propensity of such compounds to explode. Instead, it was refluxed overnight in toluene in order to give rise to the isocyanate (IR 2260 cm"1). Heating the latter compound in the presence of 6 H hydro chloric acid gave 186 in an overall yield of 70%. Biological testing on these compounds has been com pleted. They were found to be five to ten times less active than aminoadamantane. However, they were also found 93 to be quite toxic with no usable therapeutic window. Since antiviral drugs 185 and 186 are of little value, it was hoped that the dodecahedrane derivatives would have increased antiviral activity and significantly decreased toxicity. The synthetic routes to 187 and 188 follow. There are many imaginable avenues available to prepare the dodecahedrane derivatives. Obviously, arrival as directly as possible from known compounds was desirable. Attempts at preparing 187 from the hydrocarbon (NCI,, AlClj) using chemistry applicable to the adamantyl system 94 failed. Also the application of solvolysis reactions to bromide 163 (HCONH2,A95; NH,, AgOTf; MeaAlNHj ; NH(TMS)a,A) under a wide variety of conditions was unsuccessful. In light of these failures, direct conversion of bromide 163 to amine 187 was put aside for a more promising route. Conversion of 163 to acetamide 166 proceeded rapidly and in quantitative yield as stated previously. This delivered the first derivative where nitrogen was attached directly to the dodecahedryl framework. Unfortunately, all attempts to transform 166 into 187 using a number of basic and acidic hydrolyses conditions only returned starting material.

>*\ > n h c c h 3

Y v Hydrolysis*

Under more forcing conditions (ethylene glycol, potassium hydroxide, a70,96), a chemical change did occur, but no products corresponding to 187 were observed. All attempts 97 98 to remove the acetyl group failed (Li, NHS ; NH2NH2, A ; 99 DIBAL ; CHjLi). Next; attempts were made to prepare and hydrolyze imino ester 190. The literature shows this method to be very useful In removing acetyl groups under

O NHCCH3 n=cch3 E t ^ 187 CHjjCI 2

168 190

mild conditions.100 Standard procedures exist which involve treating the acetamide with triethyloxonium tetrafluoroborate101 to form the imino ester and addition of dilute acid or base to afford the amine. No success along these lines was seen in our system; only 166 was returned. Whether 190 was formed or hydrolysis returned 166 is not certain. The isolation of 190 was not attempted. A related procedure used extensively in 13- lactam chemistry was applied to 166. This involves preparation of the imino chloride and subsequent hydrolysis. Work was proceeding on this front when a successful method was found, but if more quantities of the amine are needed for future work this chemistry should be reinvestigated. 125 Nevertheless, since only a small amount of 187 was necessary for testing and the sequence evolved from an intermediate in allied work aimed at the preparation of methylamine 188, the procedure was easily pursued. The route starts with ester 175. In our initial studies, the ester was saponified to give acid 168 (87%). This deriva-

NaOH * 187 -I- 188

175 168

tive was surprisingly unreactive when subjected to proto cols found to be successful when applied to synthesis of seco derivative 186. Also, a two-step procedure aimed at directly converting acid 168 to 187 with diphenylphos- phorylazide^-0^ failed. In light of these complications, amide 191 was deemed to be important as an intermediate in arriving at 187 through Hoffmann rearrangement and at 188 by reduction. As often seen in these seemingly easy transformations, difficulty was experienced. Conversion of acid 168 to 191 through the acid chloride failed. A recently published procedure to arrive at a primary amide from the 126 corresponding ester using ammonia in methanol with cyanide 103 catalysis also failed. Weinreb's aluminum amide reagent

A i proved to be the method of choice. Conversion to 191 in near quantitative yield required treatment of 175 with a large excess of dimethylaluminum amide for an extended period of time (2-3 days) at ambient temperature. Gentle warming vastly increases the rate of reaction, with completion occurring in 24 h.

CNH

175 191

With precursor 191 in hand, both 187 and 188 were as close as one step away. First the synthesis of amine 187 will be discussed. As evidenced in the recent literature yields of what are considered Hoffmann rearrangements have been increased over the classical methods. The treatment of a primary amide with iodobenzene derivatives as oxidants give the 105 amine in high yields. A standard technique involves addition of the amide to bis(trifluoroacetoxy)iodobenzene 127 in acetonltrile and water, thereby affecting the rearrange ment and hydrolysis to the amine in one pot. Application of this procedure to our system gave a product initially identified as the isocyanate. Confusingly, its further manipulation failed to produce identifiable amounts of the amine. Thus, a modification in this procedure was attempted, viz. the solvent was changed to tert-butyl alcohol. To our delight, carbamate 192 was isolated

c n h 2 NHC02l“Bu PhI(0C0CF3)?) EtOH . t-BuOH 150*0

through trapping of the isocyanate with solvent. Subse quent treatment of 192 with hydrogen chloride affected hydrolysis and decarboxylation, and furnished the amine hydrochloride salt. With one derivative (187) synthesized, efforts turned to the methylamine 188. From work of H.C. Brown, primary amides are known to undergo ready reduction to give high yields of the methylamine derivative.106 In these reports, comment is made that the benefit of borane over lithium 128

0 N 1. BHj’THFj r -c -n h 2 r -c h 2n h 2«h c i 2. HC1 1 9 3 1 9 4

aluminum hydride in the reduction is that no carbon- nitrogen bond heterolysis has ever been observed. Unfortunately, amide 191 did not conform to the precedence and reacted quickly at room temperature (< 1 h) to give

1 9 1 195

only the hydroxymethyl derivative 195. The preparation of 195 was proven by reducing 175 with DIBAL and obtaining the same product. Repetition of this reduction at -78 °C only returned starting material and slow warming to room tem perature gave 195 as the only observable product. Treat ment of 191 with LAH interestingly gave dodecahedrane as 129 the sole product (GC analysis). Other reduction methods applied to 191 failed.

Because of the unlikely pathway followed during amide reduction, another suitable derivative was sought. Nitrile 171 was believed to fulfill this need. Dehydration of amide 191 with thionyl chloride in pyridine indeed gave the nitrile. Quantities of 171 have been limited, but pre-

1._ Reductiop p ( r 2. HC1 \ \ .

liminary reduction attempts have not resulted in isolation of 188. However, work along these lines are currently continuing in this laboratory. Whether or not a synthesis for the methylamine deriva tive 188 can be found, this project's longest term goal of synthesizing aminododecahedrane has been realized. With this compound, insight into the pharmacological features of this system will be investigated and reporting of the data will materialize at the appropriate time. 130 IV-7 NUCLEAR MAGNETIC RESONANCE OBSERVATIONS

The preparation of monosubstituted dodecahedranes has allowed for the examination of both proton and carbon NMR spectra and general trends will be discussed in this section. Table I lists the chemical shifts for the proton signals of a number of substituted derivatives. Shown previously in Figures 2 and 4 are spectra of typical compounds, bromo-(163) and carbornethoxydodecahedrane (175).

Table 1. Proton NMR Chemical Shifts for Selected Dodecahedrane Derivatives (DDH-X). (b ) Substituent Position of the Proton' ' X P Y 6,e,;

F 3.37 3.55 3.37 Cl 3.72 3.61 3.38 Br 3.94 m and t, J = 11.6 Hz 3.60 3.38 OH 3.23 m and t, J = 10.9 Hz 3.55 3.35 OMe 3.49 3.37 3.37 o n o 2 3.56 3.56 3.40

c h 3 2.92 3.36 3.36 co2 Me 3.73 m and t, J = 10.7 HZ 3.50 3.41 CONH2 3.74 m and t, J = 10.5 HZ 3.53 3.42 CN 3.80 m and t, J = 11.0 Hz 3.59 3.42 NHCOj-t-Bu 3.23 3.53 3.36 NHCOCH3 3.32 3.55 3.37

H 3.38 --

| B j 47 All spectral data obtained in CDC13; chemical shifts from CHClg used as internal standard. 131 The resonances of the p- and ^'hydrogens, as well as those of the 6,e, and c grouping are usually well separated, and assignments can readily be made on the basis of chemical shift and integrated intensities. The remote 6,e, and c protons are the least displaced from the value (6 3.38) of dodecahedrane itself and always exist as a singlet with varying peak widths. The Y-hydrogens are featureless.

broad singlets slightly downfield from the signal of the remote protons. As would be expected, the p-proton shifts vary largely with changes in the substituent. Another feature of these p-protons for many derivatives is the presence of a triplet reaching out from the broad multiplet (see Figures 2 and 4 for examples). The triplet results from coupling with the two equivalent Y-protons adjacent to the p-position. The broadening seen is assumed to result » from long-range coupling. Examination of a model demon strates that the p-protons lie in "W"-plan arrangement with both the y - and 6-protons. Long-range coupling over four 132 107 bonds has been Implicated in saturated systems. Since most long-range couplings are small in magnitude, line broadening could result from this phenomenon. Interpretation of the 1,C NMR (Tables 2 and 3) proved to be more difficult. The peak heights were of great help in correlating chemical shifts to the various carbons. The heights of a given peak roughly correlated with the number of equivalent carbons associated with that signal. The quaternary carbon is easily identified owing to the fact that it is always the peak shortest in height. Also, the resonances of the substituted carbon atoms are shifted farthest downfield from the other signals. The (3-carbons were likewise readily identified as the next downfield signal. The difficulty was encountered in assigning peaks to the ?- and 5-carbons. The two heighest signals in the spectrum of the substituted dodecahedranes were always close together, usually one signal existing slightly downfield relative to the parent signal (66.93 ppm) and the other showing a slight upfield shift. The question was which signal belonged to which carbon? Intuitively, one would expect the carbon closer to the substituent (?) to be shifted downfield. To answer this question definitively, carbon-hydrogen correlation data were obtained on a select group of derivatives in the expectation that trends would exist between substituent 133 groups (i.e. all halogenated derivatives would exhibit the same trend; all oxygen, carbon, and nitrogen bonded com pounds would likewise be respectively distinctive).

Table 2. Carbon NMR Chemical Shifts for Selected Dodecahedrane Derivatives. Is) Substituent Position of the carbon' '

X a PY 5 e S

Br 96.98 79.98 65.67 66.53 66.09 66.87 Cl 104.39 78.40 65.81 66.66 66.00 66.89 F 137.64 71.26 65.16 66.76 65.56 N.O. OH 115.99 74.95 65.71 66.99 65.71 66.89 OMe N.O. 68.90 65.47 66.98 65.77 66.79 ONOj 129.89 70.33 65.19 66.47 65.94 66.67 OCOCFj N.O. 71.65 65.34 66.51 65.90 66.69

CHS 75.55 74.67 67.17 66.74 66.39 67.06 CH.OH 81.86 68.91 67.10 66.54 66.56 67.02 CHO 90.00 67.23 66.96 66.86 66.91 67.11 CO 2 Me 84.54 70.97 66.92 66.86 66.86 66.99 c o n h 2 85.61 71.19 66.98 66.98 66.75 N.O. CN N.O. 73.91 67.03 66.67 66.87 N.O. Ph 84.42 75.15 67.19 67.06 66.71 N.O. NHCOCHj 95.77 74.21 66.06 66.68 66.16 66.80 H 66.93

All spectral data obtained in CDC1,; chemical shifts from CHC1, used as internal standard. N.O. = signal not observed.

In the halogen substituted derivatives, it was found via the C-H correlation spectrum of bromide 163 (Figure 3) and through observed C-F coupling in the fluoro derivative that the y-carbon was the source of the signal which 134 appeared more upfield with the 6-carbon signal downfield. The same shift pattern was observed with alcohol 164 and amide 166.

Table 3. Substituent Effect for Selected Dodecahedrane Derivatives. f a \ Substituent Position of the carbonv '

X a P 7 6 e C

Br 30.05 12.99 -1.26 -0.40 -0.84 -0.06 cl 37.46 11.47 -1.12 -0.27 -0.93 -0.04 F 70.71 4.33 -1.77 -0.17 -1.37 - OH 49.06 8.02 -1.22 0.06 -1.22 -0.04 OMe - 1.97 -1.46 0.05 -1.16 -0.14 ONOj 62.96 3.40 -1.74 -0.46 -0.99 -0.26 OCOCFj - 4.72 -1.59 -0.42 -1.03 -0.24 CH, 8.62 7.74 0.24 -0.19 -0.54 0.13 CHjOH 14.93 1.98 0.17 -0.39 -0.37 0.07 CHO 23.07 0.30 0.03 -0.07 -0.02 0.18 COjMe 17.61 4.04 -0.01 -0.07 -0.07 0.06 CONH2 18.68 4.26 0.05 0.03 -0.18 - CN - 6.98 0.10 -0.26 -0.06 - Ph 17.49 8.22 0.26 0.13 -0.22 - NHCOCH, 28.84 7.28 -0.87 -0.25 -0.77 -0.13

(a) values relative to the chemical shift of dodeca hedrane (66.93 ppm).

The ester 175 (Figure 5) showed the opposite chemi cal shifts, i.e., the \-carbons appearing downfield and the 6-carbons upfield from 66.93 ppm. Thus, direct attachment of a heteroatom or a halogen to the dodecahedrane framework results in slight shielding of the Y~carbon with respect to all other carbons present in the molecule. The opposite 135 effect {downshift shift of ■*“) is observed with the derivatives in which a carbon functional group is attached to the framework. The e- and c-carbons were easily identified by peak heights, with the e-carbons being equal in height to the 3-carbons. The effects of substituents on the a- and 3-carbon chemical shifts depend on steric as well as on the elec tronic "through-bond" interactions and are not simply cor related to any single parameter characterizing their steric or inductive effects. Therefore, one is usually confined to comparative analysis and discussion of various analo- 108 109 gously substituted compounds. ' In substituted dodecahedranes, the a-effect corre lates well with the chemical shifts of the quaternary carbon atoms in tertiary butyl derivatives,110 as can be seen in Figure 8. Comparison of the 3-carbon shifts also leads to good correlation with the methyl groups of tert-butyl deriva tives (Figure 9). The paramagnetic shifts vary widely from +0.30 for aldehyde 199 to +12.99 for bromide 163. In studying the shifts of the ir-carbons, obvious differences in substitutent effects are seen in dodeca hedrane derivatives, unlike the correlations in 1-adamantyl or monocyclohexyl derivatives. All adamantyl derivatives show as a rule paramagnetic (downfield) shifting of the 140V 136

120

100

40 60 80 100 FIGURE 8 a-Effect. Chemical Shifts (ppm) of a-Carbons, DDH-X vs. t-Butyl-X 851

25 4535 FIGURE 9. p-Effect. Chemical Shifts (ppm) of p-Carbons, DDH-X vs. t-Butyl-X 137 109 ir-carbons whereas all the cyclohexane derivatives show a 108 diamagnetic (upfield) shift. in the dodecahedrane derivatives, a different pattern emerges. For derivatives where the substituent is a halogen or direct attachment of a heteroatom exists (Group 1), a large diamagnetic shift is observed. In carbon-functionalized derivatives (Group 2), a slight paramagnetic shift results. Correlation between the substituent effects and the inductive cj111 constant is shown in Figure 10. The correlation is modest if one takes all the substituents into account. However, if the constituents are divided into the groupings identified above, each set gives a good correlation. oj values were used in this work instead of o* values for two reasons. First, more oj values are available for our derivatives than o* values; second, better correlations were seen, possibly due to the first reason. In aliphatic compounds, no effect on 6-carbons is usually observed. The C-4 atoms of cyclohexyl derivatives are shifted to high field, by up to 2 ppm in the cases of 108 strongly elctronegative substituents. An analogous situation is observed in adamantyl derivatives, again with 109 relatively large shifts of up to 2.3 ppm. As seen with dodecahedryl derivatives, the shifts fluctuate from upfield to downfield depending on the substituent (Table 2 and Figure 11). Only a very small (maximum -0.46 ppm) shift is 138 0.6 Gr.2

- 0.2

- 1.0

- 1.8 0.2 0.4 0.6 FIGURE 10. Y-Effect. Shifts of Y-Carbons, Substituent Effects vs. Oj Values,

0.6

0.2

— —1 1 -0.2 • • • • ■ -0.6

a 0*2 0.4 0.6 FIGURE 11. 6-Effect. Shifts of 6-Carbons, Substituent Effects vs. oj Values, 139 observed in all the cases. Accordingly, the effect on the 6-carbon in monosubstituted dodecahedranes is very small relative to comparable systems. Shifts of the e-carbons again become very pronounced with all substituents exhibiting a diamagnetic shift with the values ranging from near zero to -1.37 for fluoride 176. Correlation again using oj values is shown in Figure 12. The c position wavers from diamagnetic to paramag netic shifts of relatively small values. Figure 13 shows the effect but as seen only a modest correlation results. In summary, interesting comparisons have resulted from the study of proton and carbon NMR spectra of the substituted dodecahedranes. The proton spectra are simple and very useful in studying the structure of the molecules, whereas the carbon data has been relatively complicated as compared with the adamantyl and cyclohexyl systems showing quite different and interesting trends. - a e

- 1 . 2

0 0.4 0.6 FIGURE 12. e-Effect. Shifts of e-Carbons, Substituent Effects vs. o r Values.

0.2

- 0.2

0 FIGURE 13. C-Effect. Shifts of c-Carbons, Substituent Effects vs. oj Values. 141 IV-8 PAVING THE WAY TO NEW MOLECULES.

The success achieved along the cationic pathway has allowed for the synthesis of many functionalized dodecahedrane derivatives. Of course, in the overall study of the dodecahedryl system we want to examine other reaction pathways available. The implementation of radical and anionic pathways has been briefly investigated and will continue to be in future work. This section will review these preliminary attempts as well as progress toward new "second generation" molecules. In attempts to prepare methyl ether 172, bromide 163 was irradiated in a methanol-petroleum ether-dichloromethane solvent system. The mixture of products which resulted was identified by GC/MS as 172, dodecahedrane (3),

+ 3 + 7H96I pentane

and an unknown compound. Literature precedence indicates that the reaction likely proceeds by way of a radical 112 process. The mass spectrum showed the unidentified 142 product to be a dodecahedryl derivative giving a base peak of 259 which correspsonds to the radical ion of a substi tuted dodecahedrane. This fragment can result in one of two ways: 1) a substituted dodecahedrane not showing an M+ peak, or 2) possibly the preparation of bis-dodecahedrane 196. Fragmentation of 196 would occur at the central bond giving rise to the mass peak observed. Further identification of this unknown has not been achieved in hopes of finding a more efficient manner of preparing 196. Treatment of bromide 163 with excess allyl tributyl- tin under irradiation surprisingly gave no allyldodecahe- 113 drane (197). Only reductive cleavage of the bromine

B r CH2^CHCH2SnBu3 hv 163 107 3

atom occurred giving dodecahedrane (3). Also, photolysis of 163 in the presence of hexabutylditin rapidly (as would be expected) gave the hydrocarbon as sole product. 143 Initial work using AIBN as initiator has provided similar results. With additional work, this pathway should give positive results under the proper conditions. Attempts at generating the anion of dodecahedrane have also been made. Metalation of 163 and trapping has to date given no evidence for anion formation. Often, dodeca hedrane and starting material result from these reactions. Two examples consist in the sonication of 163 with lithium dispersion and also with n-butyllithium. We turned to metal-halogen exchange with tert-butyllithium to avoid the complication of single electron transfer which may be occurring in the lithium and n-butyllithium experiments. Addition of tert-butyllithium to 163 at -78 °C resulted in the immediate formation of a yellow color; all attempts to trap the anion with ethyl chloroformate and methyl iodide failed. The hydrocarbon was isolated as the sole product. It is conceivable that the anion could be abstracting a proton from solvent or from another dodecahe drane moiety present. Alternatively, a single electron transfer process is being followed and reduction of the bromine atom is occurring. Attempts to prepare the Grignard reagent has only given dodecahedrane to date. Obviously, finding the appropriate conditions would open avenues for new and exciting chemistry. 144 The Grignard derivative has been of Interest In connection with the synthesis of 196, trivially named dumbellane. Coupling of the Grignard reagent with silver

1. Hg 2.~AgOTf*

163 196

triflate following a procedure developed by Whitesides 33 would give 196 if successful. Wurtz coupling attempts only gave 3. Another approach to 196 can be imagined, one originating from the photolysis of the trimethylstannyl derivative 198. This derivative should be able to be made

USn(CH3 ) ^

163 198 196

following a procedure proven successful in acquiring the 114 adamantyl, cubyl, and other bridgehead tin compounds. Coupling of two photolytically prepared radicals may give 196. The coupled molecule 200 is another interesting compound where synthesis is readily imaginable through a

CHO Me Murry Coupling

109 200

115 McMurry coupling of aldehyde 199. The aldehyde was pre pared uneventfully from alcohol 195 by PCC oxidation. One attempt to couple 199 resulted in recovery of the aldehyde. With the preparation of 200, study of the steric effects exerted by the spheres should prove interesting. Experi mental attempts continue to date. Also the synthesis of another interesting coupled molecule is conceivable with existing intermediates. Treatment of 163 with 164 in the presence of silver 146 triflate may give rise to bis-dodecahedryl ether 201, a procedure found successful In synthesizing the methoxy derivative 172. Attempts to isolate dodecahedrene (202) have also been initiated by this investigator and continued by others. The dehydration of 164 and dehydrobromination of 163 have failed as might be expected owing to the strain which would exist in 202. In studying the mass spectral data of many of the dodecahedrane derivatives, a base peak which corresponds to the olefin (M +258) often is seen. Compounds with carbonyl groups (such as ester 175 and amide 191) all give large m/z 258 peaks but the most obvious compound in which this occurs is the trifluoroacetate 165. No molecular ion peak is seen; only the 258 mass is observed. This is seen as an acetate pyrolysis, but all

165 202

attempts to duplicate this result experimentally have resulted in decomposition to date. 147 IV-9 SUMMARY

This chapter has shown the successful routes to functionalized dodecahedrane derivatives and has set the stage for development of new and exciting chemistry in the upcoming months and years. The advances in the synthetic route to the parent hydrocarbon have afforded relatively large amounts of dodecahedrane for this study to begin and continue. Undoubtedly, many interesting results will continue to be generated and this researcher hopes the work completed in this dissertation will act as the backbone for the future. EXPERIMENTAL SECTION

Wadsworth-Bmons Condensation of 13* A nitrogen-blanketed solution of diethyl cyanomethylphosphonate (5.95 g, 33.6 mmol) in anhydrous tetrahydrofuran (10 mL) was vigorously stirred at 0 °C while 15 n-butyllithium in hexane (20.9 mL of 1.55 M, 33.5 mmol) was introduced slowly via syringe. The resulting sludge was stirred at room temperature for 1 h, at which point a solution of 13 (1.36 g, 8.4 mmol) in dry tetrahydrofuran was added dropwise. The reaction mixture was heated at reflux for 84 h, cooled, and treated with water (30 mL). The solution was saturated with sodium chloride and extracted with ether (3 x 25 mL). The combined organic layers were washed with 6 M hydrochloric acid (25 mL) and saturated sodium bicarbonate solution (25 mL) prior to drying. Solvent evaporation gave an orange oil, which was taken up in ether and passed through a short Florisil column. The oil so obtained slowly crystallized

148 149 and was triturated with hexanes to give 15 (1.25 g, 80.6%) as a powdery white solid, mp 109-110 #C, after sublimation. The substance was approximately a 4:1 mixture of isomers. The following are the spectral properties of the major con stituent: IR (CHCl3, cm"1) 2980, 2150, 1350, 1080, 1070, 940; *H NMR (300 MHz, CDCla) 6 5.89 (d, J * 1.7 Hz, 2 H), 4.96 (s, 1 H), 4.90-4.88 (m, 1 H), 4.43 (t, J = 4.1 Hz, 1 H), 3.40-3.37 (m, 1 H), 3.01-2.97 (m, 1 H), 2.66 (br s, 1 H), 2.58-2.55 (m, 1 H), 1.94 (d, J = 12.2 Hz, 1 H), 1.62 (dt, J = 12.1, 4.0 Hz, 1 H); 1SC NMR (75 MHz, CDC1,) ppm 174.54, 134.04, 128.59, 116.55, 89.96, 87.90, 81.20, 53.86, 46,86, 46.52, 45.63, 33.32; m/z calcd (M+) 185.0840, obs

185.0843. Anal. Calcd for C 12H lxNO: C, 77.81; H, 5.99. Found: C, 77.80; H, 6.17.

Magnesium Methanol Reduction of 15. Magnesium turnings (6.9 g, 0.283 mol) were added to a stirred solution of 15 (1.3 g, 7.0 mmol) in dry methanol (138 mL). After the vigorous reaction had subsided, the 16 mixture was stirred at room temper ature for 24 h. The magnesium salts were dissolved by addition of 6 M hydrochloric acid (130 mL). When the solution had cooled, the product was extracted into ether (4 x 50 mL) and dichloromethane (1 x 50 mL, 3 x 25 mL). The combined organic layers were washed with brine (2 x 100 mL), dried and evaporated to leave a brown oil. This material was purified by passage through a short Florisil column (ether elution) and obtained as a colorless semi-solid (1.25 g, 95%), mp 60-69 °C, which was utilized directly in the next step; IR (CHC1,, cm"1) 2970, 2250, 1355, 1070, 990, 940, 880; *H NMR (300 MHz, CDCl,) 6 6.01-5.94 (m, 2 H), 4.89-4.86 (m, 1 H), 4.34-4.32 (m, 1 H), 2.99-2.82 (m, 2 H), 2.69-2.60 (m, 2 H), 2.41-2.23 (m, 2 H), 2.03-2.00 (m, 1 H),1.77 (d, J e 11.8 Hz, 1 H), 1.60 (dt, J = 11.8, 4.0 Hz, 1 H); 1SC NMR (75 MHz, CDC1S) ppm 134.24, 132.30, 119.50, 91.53, 82.99, 48.56, 45.34, 43.73, 43.01, 39.14, 38.39, 18.90; m/z calcd (M+ ) 187.0997, obs 187.0987.

Bpoxidation of 16. A cold (0 °C), magnetically stirred solution of 16 (1.25 g, 6.68 mmol) in dichloromethane (12 mL) was treated in one portion with

17 m-chloroperbenzoic acid (1.40 g 7.0 mmol) and the oxidation was 151 allowed to proceed for 1 h at 0 °C and for 6 h at room temperature. Water (50 mL) and dichloromethane (50 mL) were added and the layers were separated. The aqueous phase was extracted with dichloromethane (50 mL) and the combined organic solutions were washed with saturated sodium thiosulfate (2 x 30 mL) and sodium bicarbonate solutions (2 x 30 mL) prior to drying. Solvent evaporation afforded a pale yellow oil that crystallized upon cooling. This solid was dissolved in the minimum amount of dichloro methane and diluted with petroleum ether to yield 1.29 g (95.6%) of 17 as colorless crystals, mp 196.0-196.5 °C (from dichloromethane-petroleum ether); IR (CHC1S, cm***) 2980, 2960, 2250, 1390, 1360, 1250, 1105, 1075, 840; AH NMR (300 MHz, CDCl,) 6 4.82-4.79 (m, 1 H), 4.20 (br s, 1 H), 3.62 (d, J = 0.98 Hz, 1 H), 3.48 (d, J = 0.98 Hz, 1 H), 3.03 (d, J = 8.6 Hz, 2 H), 2.87-2.72 (m, 2 H), 2.44-2.43 (m, 2 H), 2.07-2.02 (m, 1 H), 1.96 (d, J * 12.5 Hz, 1 H), 1.72 (dt, J = 12.0, 3.6 HZ, 1 H ) ; l3C NMR (75 MHz, CDCla) ppm 118.69, 83.16, 79.24, 55.79, 54.38, 47.97, 45.84, 43.83, 41.65, 39.92, 38.61, 19.64; m/z calcd (M+ ) 203.0947, obs 203.0891. Anal. Calcd for C^H^NO*: C, 70.92; H, 6.45. Found: C, 70.92; H, 6.42. 152 Anionic Cyclization of 17.

Epoxide 17 (35.7 mg, 0.176 mmol) in dry tetrahydrofuran (1 mL) was added to a slurry of potassium hydride (0.1 mL, 6.15 mmol/mL) in the same solvent (4 mL) under nitrogen. The reaction mixture was heated at reflux for 30 min, cooled, and quenched with saturated ammonium chloride solution. Water (20 mL) was added and the product was extracted into dichloromethane (3 x 25 ml). The combined organic layers were dried, fil tered, and evaporated to leave a yellow oil which was passed through a silica gel column to give 27.7 mg (78%) of 18 as an oil that was tosylated directly; lH NMR (300 MHz, CDC1,) 54.94-4.91 (m, 1 H), 4.70-4.67 (m, 1 H), 3.71 (s, 1 H),3.33-3.32 (m, 1 H), 3.16 (s, 1 H), 3.02-2.99 (m, 1 H), 2.72-2.70 (m, 1 H), 2.39-2.34 (m,2 H), 2.17 (br s, 1 H), 2.06 (br s, 1H), 2.01 (d, J = 12.3 Hz, 1 H), 1.63 (dt, J =

12.3, 4.3 Hz, 1 H); 1SC NMR (75 MHz, CDCls) ppm 121.43, 86.14, 79.40, 54.05, 51.81, 51.02, 42.68, 41.01; 40.45, 40.24, 38.59, 36.61; m/z calcd (M+ ) 203.0945, obs 203.0937. 153 Tosylation of 18.

Cyano alcohol 18 (273 mg, 1.3 mmol) was dissolved In dry pyridine (2 mL) and treated with £-toluene- sulfonyl chloride (256 mg, 1.3 mmol). The reaction mixture was stirred under nitrogen at room temperature for 68 h, poured into water, and extracted with dichloromethane (3x). The combined organic layers were twice washed with 6 M hydrochloric acid and once with water, saturated sodium bicarbonate solution, and brine. Drying, solvent removal, and MPLC on silica gel (elution with 42% ethyl acetate in petroleum ether) gave the tosylate as a colorless solid, mp 118-120 °C (from dichloromethane-petroleum ether) (407 mg, 85%); XR (CHC1S, cm"1); 2950, 2860, 2240, 1600, 1490, 1445, 1360, 1165, 1080, 1070, 1030, 945, 820; lH NMR (300 MHz, CDC1S) 6 7.78 (d, J - 8.5

Hz, 2 H), 7.38 (d, J- 8.5 Hz, 2 H), 4.80 (t, J = 4.57 Hz, 1 H), 4.69 (t, J = 3.91 Hz, 1 H), 4.41 (s, 1 H), 3.37-3.32 (m, 1 H), 3.17 (s, 1 H), 3.02-2.96 (m, 1 H), 2.88 (d, J - 5.65 HZ, 1 H), 2.48 (s, 3 H), 2.40-2.39 (m, 1 H), 2.19-2.16 (m, 2 H), 1.98 (d, J = 12.49 Hz, 1 H), 1.64 (dt, J = 12.49, 3.83 Hz, 1 H); 1SC NMR (20 MHz, CDC1S) ppm 145.32, 134.00, 130.12, 127.61, 119.57, 87.15, 84.41, 79.55, 53.91, 51.01, 154 49.97, 46.47, 43.52, 40.78, 40.18, 39.91, 21.65; m/z calcd (M+ ) 357.1036, obs 357.1036.

Anionic Cyclization of 18 giving 21 and 28.

NC 21 28

A. Short Reaction Time. An excess of potassium hydride in mineral oil (27 mL of 24.6% by weight, 160 mmol) was washed under nitrogen three timeswith petroleum ether and dry tetrahydrofuran (5 mL). A solution of 20 (1.20 g, 3.39 mmol) in dry tetrahydrofuran (5 mL) was added and the mixture was heated at the reflux temperature for 30 min. Workup in the manner described below afforded 380 mg (60.5%) of 21; identical in all respects to the material whose spectral properties are given in Part B.

B. Longer Heating. An excess of potassium hydride in mineral oil (37.8 mL of 24.6% by weight, 225 mmol) was washed under nitrogen three timeswith petroleum ether and dry tetrahydrofuran (10 mL) was added. A solution of 20 (1.61 g, 4.51 mmol) in the same solvent (20 mL) was introduced and the mixture was heated at reflux for 1 h. 155 After cooling, saturated ammonium chloride solution was carefully added, to be followed by water (50 mL). The pro duct was taken up in dichloromethane (3 x 75 mL), washed with brine, dried, and evaporated. The residue was chroma tographed on silica gel (elution with 16% ethyl acetate in petroleum ether) to give 310 mg (37.2%) of 21 as a color less solid, rnp 116-118 °C; IR (CHClj, cm"1) 2970, 2840, 2220, 1445, 1340, 1100, 1065, 1055, 1000, 900; *H NMR (300 MHz, CDCla) 6 4.59-4.55 (m, 1 H), 4.45-4.42 (m, 1 H), 3.53- 3.47 (m, 1 H), 2.96-2.87 (m, 3 H), 2.63-2.53 (m, 2 H), 2.47-2.44 (m, 1 H), 2.05 (d, J = 12.5 Hz, 1 H), 1.55 (dt, J =12.5, 3.60 Hz, 1 H); lsC NMR (75 MHZ, CDCl,) ppm 119.85, 90.46, 80.18, 61.69, 56.82, 56.62, 53.39, 45.13, 40.72, 39.19, 36.44, 30.64; m/z (M+) 185.0840, obs 185.0806. Anal. Calcd for CiaH lxNO: C, 77.81; H, 5.99; N, 7.56. Found; C, 77.26; H, 5.89; N, 7.45. Also isolated was 80 mg (11.1%) of 28, a volatile colorless solid, mp 173.5-174.5 °C (following sublimation); IR (CHClj, cm"1) 3050, 2960, 2860, 1440, 1330, 1320, 1230, 1040; *H NMR (300 MHz, CDC1,) 6 4.46 (t, J = 5.3 Hz, 1 H), 4.34-4.31 (m, 1 H), 3.24-3.18 (m, 1 H), 2.73-2.65 (m, 1 H), 2.63-2.60 (m, 1 H), 2.56-2.50 (m, 1 H), 2.23-2.17 (m, 1 H), 2.09-1.95 (m, 2 H), 1.93-1.72 (m, 2 H), 1.45 (dt, J = 11.9, 3.7 Hz, 1 H); 13C NMR (20 MHz, CDC13) ppm 91.23, 80.43, 156 62.04, 57.05, 54.18, 53.86, 40.89, 37.25, 35.01, 34.50,

29.29; m/z calcd (M+ ) 160.0888, obs 160.0921. Anal. Calcd for C u H 120: C, 82.46, H, 7.55. Found: C, 82.47; H, 7.83.

Independent Decyanation of 21. A solution of 21 (30 mg, 0.162 mmol) in dry tetrahydrofuran (2 mL) was added to a slurry of petroleum ether-washed potassium hydride (approx.8 mmol) in the same solvent (5 mL). The reaction mixture was heated at reflux for 8 h, cooled, quenched with saturated ammonium chloride solution (20 mL), and extracted with dichloromethane (3 x 20 mL). The combined organic extracts were washed with brine (20 mL), dried and evapo rated. The residual yellow oil was purified by MPLC (silica gel, elution with 7% ethyl acetate in petroleum ether) to give 2.2 mg (8.5%) of 28 as a colorless solid, spectroscopicallyidentical to the substance described

above.

Cyclopropyl Ring Cleavage of 21 with TMSI. A mixture of hexamethyl- disilane (30 mg, 2.0 mmol) and iodine (500 mg, 2.0 mmol) was CN warmed at 65 °c until homogeneous, 25 then heated at reflux for 1.5 h. To the cooled black mixture was added 5 mL of carbon tetrachloride and 1.1 g of potassium carbonate. Cyano ether 21 (20 mg, 0.108 mmol) was added and stirring was maintained at room temperature for 36 h. Water was added and the product was extracted Into dichloromethane (3 x 30 mL). The combined organic layers were washed with sodium thlosulfate (30 mL) and sodium chloride solutions (30 mL), dried, and evaporated. The resulting yellow oil was purified by silica gel chromatography (dichloromethane elution) to give 29.4 mg (87%) of 25 as a pale yellow oil; IR (CHClj, cm"1) 2995, 2960, 2870, 2250, 1450, 1365, 1340, 1330, 1270, 1230, 1160, 1110, 1070, 1020, 820; JH NMR (300 MHz, CDCl,) 6 5.33 (t, J = 4.28 Hz, 1 H), 4.72 (s, 1 H), 4.65-4.62 (m, 1 H), 3.11-3.10 (m, 2 H), 2.99-2.91 (m, 2 H), 2.60 (br s, 1 H), 2.51-2.50 (m, 1 H), 2.31 (br s, 1 H), 1.95 (d, J = 12.5 Hz, 1 H), 1.46 (dt, J = 12.5, 3.9 Hz, 1 H); 15C NMR (75 MHz, CDCls) ppm 118.92, 86.96, 79.55, 53.59, 50.40, 50.08, 49.19, 46.96, 41.36, 40.88, 39.43, 28.66; m/z calcd (M+ ) 312.9963, obs 312.9964.

Recyclization of 25. Potassium hydride In mineral oil (5 mL of 24.6%, 0.8 mmol) was washed three times with petroleum ether and dry tetrahydrofuran (5 mL) was added. A solution of 25 (50 mg) in the same solvent (5 mL) was added and the reaction mixture was heated at reflux for 2 h 158 and stirred at room temperature for 36 h. Saturated ammonium chloride solution was carefully added and the product was taken up In dichloromethane (3 x 25 mL). The combined organic layers were washed with brine, dried, and evaporated to leave a colorless solid shown to be 21 by IR and XH NMR spectroscopy.

DIBAL Reduction of 21. A solution of 21 (40 mg, 0.22 mmol) in dry benzene (2 mL) was added to a solution of diisobutylaluminum hydride (0.26 mmol, hexane solution) in dry benzene (1 mL). The reaction mixture was stirred at room temperature for 48 h, during which time additional aliquots of the reducing agent were introduced. Saturated Rochelle's salt solution was added and after 1 h of stir ring, the product was extracted into dichloromethane (3 x 30 mL). The combined organic layers were washed with brine (30 mL), dried, and evaporated. The residue was purified by medium pressure liquid chromatography on‘ silica gel (elution with 20% ethyl acetate in petroleum ether) to give 29 as a clear, colorless solid (23 mg, 57.5%), mp > 225 °C (from dichloromethane-petroleum ether); IR (CHC1S, cm-1) 3100, 2970, 2880, 2740, 1700, 1485, 1345, 1235, 1125, 1100, 1070, 1060, 1010, 945, 835; *H NMR (300 MHz, CDC13) 6 8.98 (s, 1 H), 4.55 (t, J = 5.2 Hz, 1 H), 4.40-4.39 (m, 1 H), 3.39-3.35 (m, 1 H), 3.14-2.72 (m, 4 H), 2.23 (br s, 1 H), 1.99 (d, J = 12.6 Hz, 1 H), 1.96 (br s, 1 H), 1.60-1.47 (m, 1 H); 1SC NMR(75 MHz, CDC1,) ppm 196.95,90.52, 80.48, 60.68, 56.49, 53.01, 50.41, 46.85, 40.85, 40.40, 36.80 (quaternary carbon not observed); m/z. calcd (M+ ) 188.0837, obs 188.0830.

Air Oxidation of 29. When the benzene-d6 solution of 29 was allowed to stand open to the air, quantitative oxidation to the carboxylic acid occurred, mp HOOC 30 208-209 °C (from ethyl acetate- petroleum ether); IR (CHCl3, cm-1) 3300-2900, 2980, 1735, 1695, 1345, 1250, 1100, 1075, 1060, 1005, 995, 900; lH NMR (300 MHz, CDC1S 6 10.1-9.6 (br, 1 H), 4.63-4.57 (m, 1 H), 4.57-4.44 (m, 1 H), 3.47-3.37 (m, 1 H), 3.09-3.05 (m, 1 H), 2.88-2.61 (m, 4 H), 2.45-2.40 (m, 1 H), 2.06 (dd,J * 8.46, 12.5 Hz, 1 H), 1.56 (dt, J = 12.5, 3.9 Hz, 1 H); 13C NMR (75 MHz, CDC1S) ppm177.27, 90.71, 80.68, 60.75, 56.47, 53.72, 53.08, 48.17, 41.59, 41.33, 36.92, 29.77; m/z calcd (M+ ) 204.0786, obs 204.0792. 160 Anal. Calcd for C12H1203: C, 70.58; H, 5.92. Found: C, 70.69; H, 5.88.

Addition of NBS to 14. A magnetically stirred solution of 14 (50 mg, 0.31 mmol) in carbon tetrachloride (4 mL) under a nitrogen atmosphere was treated with N-bromosuccinimide (60.4 mg, 0.34 mmol) and stirred at room temperature for 15 h. The mixture was diluted with water and extracted with methylene chloride (3 x 20 mL). The organic extracts were washed once with brine and dried. Filtration and solvent removal afforded a pale yellow solid which was passed through a short silica gel column (elution with methylene chloride). Concentration yielded 74.9 mg (91.0%) of 32 as a colorless solid, mp 106-108 °C (from methylene chloride-petroleum ether); IR (CHCla, cnT*) 2970, 1770, 1345, 1080, 1065, 960, 885; *H NMR (300 MHz, CDCla) 6 6.07-6.01 (m, 2 H), 5.04-5.02 (m, 1 H), 4.64-4.61 (m, 1 H), 4.35 (s, 1 H), 3.27-3.22 (m, 1 H), 3.03-2.96 (m, 2 H), 2.95-2.93 (m, 1 H); m/z calcd (M+ ) 239.9786, obs 239.9774. Anal. Calcd for C10H9BrO: C, 49.82; H, 3.76. Found: C, 49.53; H, 3.84. 161 Addition of Lithio TMS AN to 4 giving 38 and 39.

CN TMS NC,

0 TMS 38 39

To a solution of diisopropylamine (1.87 mmol, 0.263 mmol) in dry tetrahydrofuran (2 mL) at 0 °C was added n- butyllithium (1.87 mmol, 1.21 mL, 1.55 M in hexane) and stirred for 30 min. The solution of lithium diisopropyl- amide was cooled to -78 °C after which time trimethylsilylacetonitrile (2.50 mmol, 0.340 mL) was added and stirred for 30 min. Dione 4 (50 mg, 0.313 mmol) in dry tetra hydrofuran (2 mL) was added at -78 °C and allowed to warm to room temperature for 18 h. The reaction mixture was treated with saturated ammonium chloride solution and extracted with methylene chloride (3 x 30 mL). The organic layers were washed with brine (50 mL), dried, and evapo rated to leave a brown oil. Purification and separation by medium pressure liquid chromatography (10% ethyl acetate- petroleum ether) gave 28.4 mg (30.5%) of 39 and 13.3 mg (15.6%) of 38 as colorless solids. Shorter reaction times give 38 in higher yields at the expense of 39.

38: m.p. 88-90 °C (sublimation); IR (CHC1,, cm"1) 3600-3400 (br), 2970, 2250, 1760, 1325, 1260, 1140, 960, 162

840; 1H NMR (300 MHz, CDCls) 6 6.70-6.68 (m, 2 H), 6.42- 6.41 (m, 2 H), 3.00-2.96 (m, 2 H), 2.79 (s, 1 H) , 2.78-2.75 (m, 2 H), 0.24 (s, 9 H); m/z calcd (M+ ) 273.1161, obs 273.1185.

Anal. Calcd for C 1BH 19N0 2Si: C, 65.90; H, 7.00. Found: C, 66.08; H, 7.15.

39: IR (CHClj, cm"1 ) 3600-3400, 2970, 2250, 2220,

1660, 1320, 1255, 1140, 840; XH NMR (300 MHz, CDCls) 6 6.40-6.38 (m, 2 H), 6.28-6.25 (m, 1 H), 6.18-6.15 (m, 1 H), 4.51 (s, 1 H), 3.30-3.28 (m, 1 H), 2.93-2.91 (m, 1 H), 2.86 (AB, J = 17.0 Hz, 2 H), 2.71-2.68 (m, 2 H), 0.24 (s, 9 H); m/z calcd (M+ ) 296.1345, obs 296.1341.

Treatment of 38 with Base. The mono-adduct product 38 (21.2 mg, 0.078 mmol) was dissolved in dry tetrahydrofuran (2 mL) and added to potassium hydride (0.388 mmol, 0.063 mL) which had been washed twice with petroleum ether. The suspension was stirred at room temperature for 1 h, refluxed for 6 h, and cooled. The excess potassium hydride was carefully quenched with saturated ammonium chloride solution and the product was extracted into methylene chloride (3 x 30 mL). The combined extracts were washed once with brine and dried. Solvent removal and analysis by 300 MHz *H NMR 163 showed a small amount unreacted starting material and diketone 4 to be present.

Lithium Acetonitrile Addition to 4. A nitrogen-blanketed solution of n-butyllithium (2.02 mL, 3.13 mmol, 1.55 M in hexane) in dry 0 tetrahydrofuran (4 mL) was stirred 41 magnetically at -78 °C while acetonitrile (0.163 mL, 3.13 mmol) was added and stirred for an additional hour. To the resulting suspension was added 4 (50 mg, 0,.313 mmol) in dry t'etrahydrofuran (4 mL). The reaction mixture was rapidly stirred for 3 h, quenched by the addition of saturated ammonium chloride solution (20 mL), and extracted with dichloromethane (3 x 20 mL). The combined organic layers were washed with brine (30 mL) and dried. Filtra tion and evaporation left a yellow solid which was passed through a short silica gel column (elution with dichloromethane) giving 54.4 mg (8 6 .6%) of 41 as a white solid; mp 147-148 ®C (from ethyl acetate-petroleum ether); IR (CHC1*, cm-1) 3600-3400 (br), 2960, 2940, 2880, 1770, 1460, 1340,

1110, 1065; lH NMR (300 MHZ, CDC1S) 6 7.16-7.15 (m, 2 H), 6.47 (s, 2 H), 4.46 (s, 1 H), 3.12-3.10 (m, 2 H), 2.88- 2.86 (m, 2 H),2.73 (s, 2 H); 1SC NMR (75 MHz, CDCl,) ppm 164 194.40r 140.25, 139.24, 117.60, 85.96, 51.99, 48.88, 31.49; m/z calcd (M+-H20) 183.0684, obs 183.0684.

Trapping the Alkoxide Anion with Acetic Anhydride. Acetate

42. Acetonitrile (0.163 mL, 3.13 O mmol) was added to a cold (-78 °C) OCCH solution of n-butyllithium (2.02 mL, 3.13 mmol, 1.55 M in hexane) in

42 dry tetrahydrofuran (2 mL), which was vigorously stirred for 1 h. To the suspension was added 4 (50 mg, 0.313 mmol) in the same solvent (2 mL) and stirring was maintained at -78 °C for 4 h. The alkoxy anion was trapped by addition at -78 °C of acetic anhydride (0.294 mL, 3.13 mmol), stirred for an additional hour, and allowed to warm to room temperature. The mixture was diluted with saturated ammonium chloride solution and extracted with dichloromethane (3 x 20 mL). The combined extracts were washed with brine (50 mL) and dried. Evaporation and purification by medium pressure chromatography (elution with 33% ethyl acetate-petroleum ether) gave 26.6 mg (35.0%) of 42 as a white solid; IR

(CHCl3, cm"1 ) 3040, 2940, 2860, 1770, 1740, 1375, 1325,

1240, 1070, 1060; NMR (300 MHz, CDCla) 6 6.70-6.69 (m, 2 H), 6.43-6.41 (m, 2 H), 3.31-3.29 (m 2 H), 3.04 (s, 2 H), 165 3.00-2.98 (m, 2 H), 2.05 (s, 3 H); m/z calcd (M+ ) 243.0895, obs 243.0907.

Treatment of 42 with Base. To clean sodium hydride (60 mg, 50% in oil) suspended in dimethoxyethane (2 mL) was added the acetate 42 (15 mg, 0.062 mmol) in the same solvent (2 mL). Stirring was continued for 15 min and the excess hydride was quenched with saturated ammonium chloride solution. The product was extracted into methylene chloride (3 x 20 mL) and dried. Solvent removal and analysis by NMR and TLC showed diketone 4 to be present as the sole product.

Trapping the Alkoxide Anion with Chlorodiethylphosphonate. Phosphonate 43. Acetonitrile (0.114 mL, 2.19 mmol) was added to a cold (-78 °C) OPlOEtl2 solution of n-butyllithium (1.41 mL, 2.19 mmol) in dry tetrahydro-

43 furan (2 mL) and stirring was continued for 1 h. To the anion was added 4 (35 mg, 0.219 mmol) in the same solvent (3 mL). After 5 h at -78 °C, the alkoxide was trapped by the addi tion of chlorodiethylphosphonate (0.36 mL, 2.50 mmol), diluted with saturated ammonium chloride solution, and extracted with ether (3 x 20 mL). The organic extracts were washed with brine (50 mL) and saturated sodium bicarbonate solution (50 mL), and dried. Evaporation and purification by preparative TLC on silica gel (elution with 20% ethyl acetate-petroleum ether) gave slightly impure 43 which remained at the baseline; *H NMR (300 MHz, CDCla) 6 6.79-6.77 (m, 2 H) , 6.46-6.45 (m, 2 H), 4.18 (q, J = 7.1 Hz, 2 H), 4.16 (q, J *= 7.1 Hz, 2 H), 3.18 (s, 2 H), 3.12- 3.10 (m, 2 H), 3.06-3.02 (m, 2 H), 1.39 (t, J = 7.1 Hz, 3 H), 1.38 (t, J = 7.1 Hz, 3 H); m/z calcd (M+ ) 337.1079, obs 337.1112.

Treatment of 41 with Sodium Hydroxide. Cyano alcohol 41 (62.8 mg, 0.313 mmol) was dissolved in a methanol-sodium hydroxide solution (1:1, 2 M aq NaOH, 10 mL) and 44 stirred at 60 °C for 2 h. The cooled mixture was diluted with water and extracted with methylene chloride (3 x 30 mL). The extracts were washed once with brine (30 mL) and dried. Concentration gave 58.4 mg (93.0%) of 44 as a white solid, mp 140.5-141.0 °C (from ethyl acetate-petroleum ether); IR (CHC1S, cm"1 ) 3020, 2980, 2870, 2200, 1770, 1680, 1640, 1615, 1450, 1415, 1345, 1230, 1120, 1100, 1070, 1015, 985, 167

975, 910, 870; lH NMR (300 MHz, CDClj) 6 6.25-6.21 (m, 1 H), 6.05-6.01 (m, 1 H), 5.03-5.01 (m, 1 H), 3.10-3.06 (m, 1

H), 2.93-2.87 (m, 1 H), 2.72 (AB, 2 H), 2.64-2.57*(m, 2 H), 2.03 (dd, J = 13.2, 3.7 Hz, 1 H), 1.73 (dt, J = 13.2, 3.7 Hz, 1 H); m/z calcd (M+ ) 201.0789, obs 201.0770.

Wadsworth-Emmons Condensation of 44. A solution of diethyl cyanomethylphosphonate (66.5 mg, 0.38 mmol) in dry tetrahydrofuran was stirred at 0 °C under a nitrogen atmosphere while n-butyllithium (0.365 mL, 0.38 mmol, 1.03 M in hexane) was introduced via syringe. The resulting solution was stirred for 30 min before 44 (37.8 mg, 0.188 mmol) dissolved in the same solvent (5 mL) was added. The reaction mixture was stirred at room temperature for 3 d, quenched with water, and extracted with dichloromethane (3 x 30 mL). The organic extracts were washed once with brine (30 mL) and dried. Solvent removal gave an oily solid which was passed through a plug of silica gel (methylene chloride elution) yielding 38.2 mg (90.7%) of 40 as a crystalline solid, mp 155-156 °C (from ethyl acetate- petroleum ether); IR(CHCl3, cm”1 ) 3000, 2907, 2940, 2220, 2000, 1670, 1640, 1610, 1450, 1415, 1340, 1230, 1130, 1110, 168 1070, 1015, 975, 800; *H NMR (300 MHz, CDCla) 5 6.10-5.85 (m, 2 H), 5.02-4.91 (m, 2 H), 3.55-3.54 (m, 1 H), 3.19-3.15 (m, 1 H), 2.92-2.89 (m, 1 H), 2.76-2.49 (m, 3 H),’ 2.00-1.94 (in, 1 H), 1.69-1.61 (m, 1 H); m/z calcd (M+ ) 224.0949, obs 224.0948.

Treatment of 39 with acid to give 40. The (3-hydroxy silane 39 (5 mg, 0.017 mmol) was dis solved in dry tetrahydrofuran (5 mL), treated with concen trated sulfuric acid (5 drops), and stirred for 50 h at room temperature. The solution was diluted with dichloromethane (50 mL), washed with saturated sodium bicarbonate solution, and dried. Filtration and solvent removal yielded 40 spectroscopically identical to the substance described above.

PCC Oxidation of 87. A slurry of pyridinium chloro- HC CH chromate (1.21 g, 5.65 mmol), sod ium acetate (0.719 g, 8.46 mmol), and Celite (1.5 g) in dry methylene

80 chloride (50 mL) was cooled to 0 °C under a nitrogen atmosphere. With vigorous stirring, a solution of diol 87 (350 mg, 1.41 mmol) in the same solvent (10 mL) was added, allowed to warm to room temperature, and stirred for 3 h. Ether (20

mL) was added dropwise and the mixture was placed on top of a silica gel column (15 g). Elution with 20% ethyl acetate-petroleum ether yielded 200 mg (58%) of 80 as a colorless oil; IR (CHClj, cm"1 ) 2980, 2740, 1730, 1470,

1400, 1250, 1035, 910; XH NMR (300 MHz, benzene-ds) 6 9.72 (t, J = 2.2 Hz, 2 H), 3.00-2.96 (m, 4 H), 2.62 (dd, J = 7.6 and 2.2 Hz, 4 H), 2.52-2.47 (m, 4 H), 2.40-2.27 (m, 2 H), 1.72-1.51 (m, 4 H); 1SC NMR (20 MHz, CDCl,) ppm 202.05, 48.80, 47.40 (2 C), 44.65, 28.04; m/z calcd (M+ ) 244.1463, obs 244.1471.

Photolysis of 80.

Dialdehyde 80 (68 mg, 0.279 mmol) was dissolved in dichloromethane (20 mL) and toluene (10 mL) and irradiated through a quartz immersion well at -50 °C using a 88 Philips H.P.K. 125 W lamp for 9.5 h. The solution was passed through a short silica gel column giving 28 mg (46.7%) of 88 as a colorless oil; IR (CHClj, cm"1 ) 3050, 2980, 2880, 1735, 1480, 1465, 1385,

1270, 1225; *H NMR (300 MHZ, CDC1S) 6 9.78 (t, J_= 2.3 Hz, 1 H), 2.99-2.98 (m, 4 H), 2.66 (dd, J = 7.6 and 2.3 Hz, 2 H), 2.61-2.55 (m, 2 H), 2.43-2.33 (series of m, 3 H), 170 2.06-1.99 (m, 1 H), 1.82 (dt, J ■ 15.3 and 2.1 Hz, 2 H),

1.56 (t, J = 11.5 HZ, 1 H), 1.51 (t, J = 11.5 Hz, 1 H), 1.03 (d, J = 7.1 Hz, 3 H); 1JC NMR (20 MHz, CDC1S) ppm 202.53, 50.33. 48.84, 47.71, 47.55, 44.91, 44.86, 44.73, 27.36, 14.21; m/z calcd (M+ ) 216.1514, obs 216.1494.

Saponification of 65. Dlester 65 (100 mg, 0.33 mmol) ,COOH HOOC was dissolved in 2 N aqueous sodium

hydroxide-methanol solution (1:1 by volume, 10 mL) and stirred at room

89 temperature for 18 h. The solution was acidified with 6 N hydrochloric acid, and extracted with ethyl ether (20 mL) and dichloromethane (3 x 20 mL). The organic extracts were washed once with brine (50 mL) and dried. Filtration and solvent removal yielded a white solid which was triturated with ether to give 83.8 mg (92.3%) of 89 as a white solid, mp 225 °C dec (from ethyl acetate); IR (KBr, cm"1) 3500-3000

(v br), 2980, 1720, 1420; AH NMR (300 MHz, DMSO-d6) 6 11.97 (br, 2 H), 2.92 (br s, 4 H), 2.45 (d, J = 16.2 Hz, 4 H), 2.22-2.02 (series of m, 4 H), 1.88-1.69 (series of m, 4 H), 1.54-1.42 (m, 2 H); 1SC NMR (20 MHz, DMSO-d6) ppm 173.37, 53.72, 48.04, 46.70, 34.43, 27.53; m/z calcd (M+ ) 276.1362, obs 276.1331. Addition of Phenylllthluin to 80. Dialdehyde 80 (200 mg, 0.82 .CHOHPh PhOHCH mmol) in dry tetrahydrofuran (10 mL) was cooled to -78 °C under a nitrogen atmosphere and phenyllithium (3.28 mmol, 0.45 M in 92 ether) was added and allowed to warm to room temperature. After 1.5 h, the reaction mixture was quenched with saturated ammonium chloride solution and extracted with ether (3 x 40 mL). The ether extracts were washed with water (50 mL) and brine (50 mL) prior to drying. The resulting oil was chromatographed on preparative TLC plates (elution with 20% ethyl acetate- petroleum ether) to remove side products. The baseline

material was rechromatographed through silica gel (20 g, elution with the same solvent system) affording 92 as an oil (240 mg, 73%); IR (CHC1S, cm-1) 3630, 2970, 1620, 1500, 1465; *H NMR (300 MHz, CDCl,) 5 7.36-7.28 (m, 10 H), 4.69- 4.67 (m, 2 H), 2.91 (br s, 4 H), 2.49 (very br s, 4 H),

2.08-1.87 (m, 8 H), 1.78 (br s, 2 H), 1.73-1.52 (m, 2 H); m/z calcd (M+ ) 400.2402, obs 400.2400. Manganese Dioxide Oxidation of 92

Diol 92 (170 mg, 0.425 mmol) .coc6h 5 was dissolved in dichloromethane (15 mL), treated with manganese dioxide (740 mg, 8.5 mmol) under a nitrogen atmosphere, and stirred for 4 d. The mixture was filtered through a Celite padand concentrated to furnish 100 mg (59%) of 81 which was recrystallized to give x-ray quality crystals, mp 131-132 °C (from ethyl acetate-petroleum ether); IR (CHC1S, cm-1) 2970, 1700, 1610, 1595, 1460,

1380, 1325, 1140, 1010; *H NMR (300 MHz, CDClj) 6 8.02-7.98

(m, 4 H), 7.59-7.44 (m, 6 H), 3.24 (d, J = 7.2 Hz, 4 H), 3.01-2.90 (m, 4 H), 2.67-2.63 (m, 4 H), 2.53-2.43 (m, 2 H), 1.95-1.59 (series of m, 4 H); 13C NMR (20 MHz, CDC1S) ppm 200.20, 137.47, 132.87, 128.65, 128.21, 48.80, 47.40, 46.70, 39.22, 28.43; m/z calcd (M+ ) 396.2089, obs 396.2102. 173 Dehydration of 87.

Dlol 87 (100 mg, 0.40 mmol) was dissolved In dry tetrahydrofuran-pyrldine (1:1, 10 mL) and o- nitrophenylselenocyanate (277 mg,

94 1.22 mmol) was introduced followed by dropwlse addition of tri-n-butyl phosphine (246 mg, 1.22 mmol) at room temperature. The solution was stirred for 3 h, cooled to 0 °C, and treated with hydrogen peroxide (30%, 137 mg, 1.22 mmol) in one portion. After an additional 3 h of stirring at room temperature, the solution was diluted with water (30 mL) and extracted with dichloromethane (3 x 100 mL). The combined extracts were washed with 1 N hydrochloric acid (2 x 10 mL), saturated sodium bicarbonate solution (100 mL), then dried. Filtration and concentration gave a yellow-red oil which was passed through a short silica gel column (elution with petroleum ether) to give 45.5 mg (53.2%) of

94 as a colorless oil; XH NMR (300 MHz, CDCl,) 6 6.14-6.03 (m, 2 H), 5.13-5.01 (m, 4 H), 3.03-3.02 (m, 4 H), 2.59 (bs,

6 H), 2.10 (d, J = 15.4 Hz, 2 H), 1.76-1.67 (m, 2 H); XSC NMR (20 MHz, CDC1S) ppm 140.03, 115.59, 55.23, 50.09, 47.74, 29.43; m/z calcd (M+ ) 212.1565, obs 212.1525. Addition of Phenyl Benzeneselenosulfonate to 94 and Elimination giving 96.

Phenyl benzeneselenosulfonate (336 mg, 1.13 mmol) and 94 (80 mg, 0.37 mmol) were dissolved in carbon tetrachloride (10 mL) and dichloromethane (10 mL) in a large pyrex 96 tube and irradiated for 24 h in a Rayonet photochemical reactor (253 nm lamps). The solution was concentrated and taken up in dichloromethane (10 mL) and cooled to 0 °C. Hydrogen peroxide (30%, 1.2 mmol, 0.12 mmol) was added rapidly, allowed to warm to room tempera ture, and stirred overnight. The reaction mixture was diluted with water (20 mL) and extracted with dichloro methane (3 x 30 mL). The combined organic extracts were washed with saturated sodium bicarbonate solution (30 mL), brine (30 mL), and dried. Evaporation yielded a yellow oil which was chromatographed on preparative TLC plates (elution with 50% ethyl acetate-petroleum ether). Recrys tallization (from chloroform-petroleum ether) gave 125 mg (67.2%) of 96 as a colorless solid; lH NMR (300 MHz, CDC1S)

6 7.92-7.86 (m, 4 H), 7.63-7.58 (m, 6 H), 7.14 (dd, J = 15.3 and 4.7 Hz, 2 H), 6.26 (dd, J = 15.6 and 1.4 Hz, 2 H),

3.07-3.05 (m, 4 H), 2.73-2.60 (br s, 6 H), 1.86-1.74 (m, 4 H); 1SC NMR (20 MHz, CDCl,) ppm 147.07, 136.35, 133.40, 175 131.54, 129.74, 127.55, 53.20, 49.37, 47.69, 30.62; m/z calcd (M+ ) 492.1429, obs 492.1477.

Methyl Wittig Addition to 5. Under an argon atmosphere methyl triphenylphosphonium iodide (2.2 g, 5.4 mmol) was suspended in dry tetrahydrofuran (20 mL) and dry 100 hexamethylphosphoramide (2.5 mL) with vigorous magnetic stirring. Potassium hexamethyldisilazide (5.6 mmol, 0.9 M in THF) was added at room temperature via syringe and stirred 15 min. The resulting yellow solution was cooled to -78 °C and 5 (250 mg, 1.33 mmol) in dry tetrahydrofuran (5 mL) was introduced and stirred cold 30 min, warmed to room tempera ture and further stirred 1.5 h. The reaction was quenched with addition of excess sodium sulfate hexahydrate, diluted with water, and extracted with ether (3 x 50 mL). The ether layers were washed with water and dried. Concentra tion and chromatography (SiOa, eluting with petroleum ether) yielded 220 mg (90.2%) of 100 as a colorless solid, mp sublimes >60 °C; IR (CHClj, cm"1 ) 3060, 2910, 2850, 1640, 1440, 1290, 1000, 955, 880; *H NMR (300 MHz, CDC1S) 5

4.87 (s, 4 H), 3.06 (br s, 8 H), 2.23-2.17 (m, 2 H), 2.04 (d, J = 12.2 Hz, 2 H); 13C NMR (75 MHz, CDCls ppm 168.04, 176 106.96, 51.56, 51.49, 47.11; m/z calcd (M+ ) 184.1252, obs 184.1259.

Anal, calcd for C 11(H16: C, 91.25; H, 8.75. Found: C, 91.11, H, 8.94.

Borane Addition to 100 and Oxidative Workup Affording 101. Olefin 100 (245 mg, 1.33 mmol) was taken up in dry tetrahydrofuran (10 mL) and cooled to 0 °C under argon. Borane-tetrahydrofuran (4

101 mL, 1 M in THF) was added dropwise and stirred at room temperature for 4 h. Water (1 mL) was very carefully added and a mixture of sodium hydroxide (1.6 g in 5 mL of H 20) and hydrogen peroxide (30%, 4.6 g) was introduced in one portion. The resulting solution was stirred for 1 h, diluted with water and extracted with ether (3 x 75 mL). The extracts were washed with brine and dried. Filtration and solvent removal gave 101 as a greasy oil (280 mg, 93.6%) which was normally used without further purification. Purification for spectral data was achieved by silica gel chromatography (25% ethyl acetate-petroleum ether); IR (CHC1S, cm”1 ) 3610,

2950, 1350, 1120, 970; lH NMR (300 MHZ, CDC13) 6 3.86 (d, J = 7.6 Hz, 4 H), 3.02 (m, 4 H), 2.65 (m, 4 H), 2.64-2.12 (m, 2 H), 1.81-1.59 (series of m, 4 H); 13C NMR (75 MHz, CDC1S) 177 ppm 63.37, 53.28, 47.16, 46.92, 27.43? m/z calcd (M+ ) 220.1454, obs 220.1464.

Jones Oxidation o£ 101. Impure diol 101 (280 mg) was HO2C co2h dissolved in acetone (20 mL) and cooled to 0 °c under argon. The solution was treated with excess 102 Jones reagent (2 mL), warmed to

room temperature, and stirred for 1 h. The orange-red solution was quenched by addition of excess isopropyl alcohol at 0 °C, diluted with water, and extracted with ether (3 x 50 mL). The organic layers were washed with brine and dried. Trituration of the residue with ethyl ether gave 166 mg (51.4%, total for two steps) of 102 as a white solid (yields of 70% were obtained starting with purified diol), mp >250 °C? IR (KBr, cm"1) 3600-2700 (br), 2980, 1685, 1415, 1285, 1000? *H NMR (300 MHz, DMSO-dJ 5 11.93 (br s, 2 H), 3.33 (s, 4 H), 2.86-2.75

(m, 6 H), 2.18-1.85 (m, 4 H)? 13C NMR (75 MHz, DMSO-d6) ppm 174.28, 56.79, 46.47, 46.18, 34.43? m/z calcd (M+-HaO) 230.0949, obs 230.0946. 178 Esterification of Diacid 102.

A suspension of 102 (32.5 mg,

CH3 O2C C02CH3 0.13 mmol) in dry benzene (10 mL)

was treated with 1 ,8-diazabicyclo- [5.4.0]undec-7-ene (0.5 mL) fol lowed by dry methyl iodide (0.5 103 mL). The resulting thick white suspension was rapidly stirred at room temperature for 2.5 h, subsequently diluted with water (20 mL), and extracted with ether (3 x 40 mL). The organic extracts were washed with saturated sodium bicarbonate solution, brine, and dried. Filtration and concentration gave 34.0 mg (94.0%) of 103 as a white solid, recrystallization from ethyl acetate-acetone gave colorless needles, mp 128.5-129.0 °C (sealed tube); IR (CHCls, cm"1 ) 2960, 1725, 1440, 1285,

1180; XH NMR (300 MHz, CDCl,) 5 3.67 (s, 6H), 3.02-3.01 (m,

4 H), 2.87 (br s, 6 H), 2.17-2.11 (m, 4 H); 1SC NMR (75 MHz, CDCla) ppm 173.90, 56.94, 51.11, 47.15, 46.81, 34.42; m/z calcd (M+ ) 276.1361, obs 276.1349. 179 Acyloin Condensation of 103. Trimethylsilyl chloride (1.03 ,o g, 9.5 mmol) was added to a sodium dispersion (194 mg, 8.45 mmol) in dry toluene (20 mL) under an argon

99 atmosphere followed by addition of 103 in the same solvent (10 mL). The mixture was refluxed for 21 h, filtered through Celite, and concentrated in vacuo to a yellow oil. The residue was dissolved in dry toluene (10 mL) and added dropwise to a solution of anhydrous ferric chloride (137 mg, 0.84 mmol) in dry ether (20 mL) containing concentrated hydrochloric acid (7 drops). The solution was gently refluxed overnight (16 h), cooled, poured into saturated ammonium sulfate solution, and extracted with dichloromethane (3 x 40 mL). The combined extracts were washed with saturated sodium bicarbonate solution, brine, and dried. Purification by MPLC (20% ethyl acetate-petroleum ether) gave 30.3 mg

(66.9%) of 99 as a yellow solid; IR (CHCls, cm-1) 2970,

2880, 1725, 1705, 1470, 1310, 1100; UV (XITiav, CH3CN) 222 ulaX nm; XH NMR (300 MHz, CDC13) 6 3.51 (t, J = 11.3 Hz, 2 H), 3.20-3.19 (m, 4 H), 3.08-3.04 (m, 4 H), 1.92 (d, J - 14.5 Hz, 2 H), 1.64 (dt, J = 14.5 and 6.5 Hz, 2 H); X3C NMR (75 MHz, CDClj) ppm 197.44, 55.69, 48.58, 47.85, 36.55; m/z calcd (M+ ) 214.0994, obs 214.1007. 180 Keto Phenoxymethyl Disecododecahedrane 143.

A mixture of epimeric alcohols 136 (345.4 mg, 0.89 mmol) in dry dichloromethane (5 mL) was rapidly added dropwise to a suspension of pyridinium chlorochrornate (481 mg, 1 4 3 2.23 mmol), dry sodium acetate (367 mg, 4.47 mmol), and dry Celite (400 mg) in the same solvent (10 mL) and stirred for 5 h under an argon atmo sphere. Ether (50 mL) was added dropwise to precipitate the inorganic salts, and the mixture was passed through a silica gel plug (elution with dichloromethane). Concentra tion gave ketone 143 as a white foam which was normally used without further purification. Purification for spectral data was achieved by MPLC (85% petroleum ether: 7.5% ethyl ether: 7.5% dichloromethane elution). Recrys tallization from acetone yielded colorless crystals, mp 61- 63 °C; IR (CHClj, cm"1 ) 2970, 1720, 1600, 1580, 1490, 1450,

1370, 1210, 1040; *H NMR (300 MHz, CDCl,) 6 7.26-7.21 (m, 2 H), 6.93-6.79 (m, 3 H), 4.18 (iABq, J « 7.7 Hz, 1 H), 3.63 (jABq, J = 7.7 Hz, 1 H), 3.63-1.50 (series of m, 21 H); 1SC NMR (75 MHz, CDCla) ppm 226.03, 158.96, 129.15, 120.69, 114.80, 76.73, 69.38, 67.62, 67.50, 67.29, 67.07, 64.95, 62.22, 61.35, 59.97, 58.39, 58.11, 53.87, 51.98, 51.81, 181 50.73, 49.56, 36.66, 30.79, 27.31; m/z calcd (M+ ) 384.2090, obs 384.2132.

Hydroxy Phenoxymethyl Secododecahedrane 144. Crude ketone 143 was taken up in a benzene:tert-butyl alcohol solution (20 mL, 4:1 volume) and

triethylamine (20 drops) was added. The solution was degassed with 144 argon for 30 min and irradiated using a 450 W Hanovia lamp for 13 h. Evaporation and purification by MPLC (elution with 85% petroleum ether: 7.5% dichloromethane: 7.5% ether) gave 246.7 mg (71.8% from epimeric alcohols) of alcohol 144, mp 111-113 °C (colorless crystals from acetone); IR (CHC1S, cm”1) 3540, 3000, 2930, 2860, 1595, 1580, 1485, 1230, 1025; *H NMR (300 MHz, CDC1S)

6 7.31-7.28 (m, 2 H), 6.98-6.92 (m, 3 H), 4.00 (s, 2 H), 3.56-2.89 (series of m, 19 H), 1.67-1.60 (m, 2 H); 1SC NMR (75 MHZ, CDCla) ppm 158.70, 129.35, 121.06, 114.87, 97.20, 82.24, 73.97, 69.08, 67.55, 65.57, 64.90, 64.19, 60.91, 60.88, 52.98, 48.94, 32.08; m/z calcd (M+ ) 384.2089, obs

384.2122. Phenoxymethyl Secododecahedrene 145.

Alcohol 144 (238.4 mg, 0.620

C H 2OPti mmol) was dissolved In dry benzene (35 mL) and dry pyridine (15 mL). This solution was treated with distilled phosphorous oxychloride

(4 mL) and heated at 55 °C for 10 h under an argon atmosphere. The solution was carefully poured onto cold dilute hydrochloric acid (20 mL) and extracted with methylene chloride (3 x 100 mL). The organic extracts were washed with brine and dried. Filtration and solvent removal gave an oil which was used without further purification. Recrystallization from acetone for spectral data yielded colorless crystals, mp 106-109 °C; IR (CHClj, cm"1 ) 3000, 2940, 2860, 1700, 1595,

1580, 1440, 1290, 1240, 1030; *H NMR (300 MHz, CDC1,) 6 7.28-7.22 (m, 2 H), 6.93-6.88 (m, 3 H), 3.92 (s, 2 H), 3.50-2.86 (m, 16 H), 1.96-1.92 (m, 1 H), 1.63-1.55 (m, 2 H); l3C NMR (75 MHz, CDC1S) ppm 159.68, 143.40, 141.85, 129.16, 120.19, 114.65, 82.34, 73.60, 66.98, 66.95, 66.45, 64.06, 63.86, 62.12, 61.21, 60.35, 60.30, 59.24, 59.09, 52.63, 51.83, 50.79, 48.08, 33.12, 30.45; m/z calcd (M+ ) 366.1983, obs 366.1947. 183 Phenoxymethyl Secododecahedrane 147

CH2OPh Olefin 145 was dissolved in dry tetrahydrofuran (90 mL) and glacial acetic acid (4 drops), saturated copper sulfate (4 drops),

147 and hydrazine hydrate (1.41 g, 28.2 mmol, 1.37 mL) were added. The mixture was vigorously stirred under argon and a solution of sodium periodate (753 mg, 3.52 mmol) in water (4 mL) was added over a 1 h period and stirred an addition 1 h. The reaction mixture was diluted with water (100 mL) and extracted with dichloromethane (3 x 150 mL). The organic layers were washed once with brine and dried. Evaporation and recrystallization from acetone gave 206.1 mg (90.2% overall from seco-alcohol 144) of 147 as colorless crys tals, mp 94-95 °C; IR (CHC1S, cm"1 ) 3000, 2830, 1595, 1590, 1485, 1410, 1360, 1220, 1085, 1030; XH NMR (300 MHz, CDC1S)

6 7.28-7.22 (m, 2 H), 6.93-6.88 (m, 3 H), 3.73 (s 2 H), 3.57-2.86 (series of m, 19 H), 1.61-1.51 (m, 2 H); 1SC NMR (75 MHz, CDC1,) ppm 159.77, 129.22, 120.20, 114.62, 82.65, 76.64, 69.26, 67.73, 65.86, 65.68, 64.72, 61.97, 53.34, 52.83, 52.60, 49.87, 32.17; m/z calcd (M+ ) 368.2140, obs

368.2164. Anal, calcd for C27H3B0: C, 88.00; H, 7.66. Found: C, 87.57; H, 7.70. 184 Hydroxyroethyl Secododecahedrane 153

The phenoxymethyl secododeca c h 2o h hedrane 147 (206.1 mg, 0.559 mmol) was dissolved in dry tetrahydro furan (80 mL) and added to dry (Na) ammonia (80 mL) at refluxing tem 153 peratures. Absolute ethanol (8 mL) was added followed by portionwise addition of sodium (643 mg, 28.0 mmol) to maintain the blue color.The ammonia was evaporated and the mixture was diluted with water (100 mL) and extracted with dichloromethane (3 x 100 mL). The combined extracts were washed with water and dried. Evaporation gave the dienol ether which was taken up in tetrahydrofuran (10 mL) and treated with 3 M hydrochloric acid (4 mL). The solution was stirred overnight (16 h), diluted with dichloromethane (20 mL), and washed with water and saturated sodium bicarbonate solution. Drying and evaporation yielded a white solid which was chromatographed over silica gel (10% ethyl acetate-petroleum ether elution) to give pure 153 (152.5 mg, 93.3%) as a white solid, mp >240 °C (from acetone),* IR (CHC13, cm--1-) 3350, 3005, 2910,

2840, *H NMR (300 MHz, CDCla) 6 3.54-2.69 (series of m, 22 H), 1.57-1.45 (m, 2 H) ,* l3C NMR (75 MHz, CDCla) ppm 84.70, 71.81, 69.27, 67.88, 65.79, 65.67, 64.15, 62.05, 52.95, 185

52.85, 52.71, 49.94, 32.25; m/z calcd (M+-H20) 274.1722, obs 274.1756.

Secododecahedrylcarboxaldehyde 154. Alcohol 153 (30.7 mg, 0.105 O ii CH mmol) In dry dichloromethane (3 mL) was added dropwise to a suspension of pyrldlnlum chlorochromate (40.7 mg, 0.189 mmol), dry sodium acetate 154 (25 mg, 0.305 mmol), and dry Cellte (60 mg) In the same solvent (3 mL) and stirred at room temperature for 3 h. Ether (10 mL) was added and the mixture was passed through a silica gel plug (10% ethyl acetate-petroleum ether elution) to give 27.9 mg (91.5%) of 154 as a white solid, mp > 240 °C (from acetone); IR (CHClj, cm-1) 3030, 2940, 2880, 2700, 1700; »H NMR (300

MHz, CDClj) 6 9.47 (s, 1 H), 3.60-3.03 (series of m, 19 H), 1.63-1.53 (m, 2 H); 1SC NMR (75 MHz, CDCls) ppm 201.32, 92.56, 69.43, 68.28, 65.86, 65.68, 62.62, 61.86, 53.16, 52.88, 50.17, 49.86, 32.10; m/z calce (M+ ) 290.1670, obs 290.1654. Carboxysecododecahedrane 155. Aldehyde 154 (19.0 mg, 0.065 mmol) was suspended in acetone (3 mL) at room temperature and dry tetrahydrofuran (approx. 3 mL) was added until dissolution of the 155 aldehyde was complete. The solu tion was cooled to -45 °C and Jones reagent (1 mL) was added. The reaction mixture was stirred for 1 h, quenched with excess isopropyl alcohol at -45 °C, allowed to warm to room temperature, and diluted with water (10 mL). The solution was extracted with ether (3 x 30 mL) and the combined extracts were washed with water and dried. Evaporation yielded 16.9 mg (84.5%) of 155 as a white solid, mp >240 °C (acetone); IR (CHC1S, cm"1 ) 3400-2700

(broad), 2930, 2860, 1680; lH NMR (300 MHz, CDC13) 6 3.54- 2.99 (series of m, 19 H), 1.59-1.54 (m, 2 H); 1SC NMR (75 MHz, benzene-d8) ppm 87.50, 69.52, 68.32, 66.53, 66.16, 65.92, 62.11, 55.54, 53.40, 53.14, 49.91,32.10 (acid carbon not observed); m/z calcd (M+ ) 306.1620, obs

306.1651. 187 Vinylsecododecahedrane 156 Methyl triphenylphosphonium pH=CH2 Iodide (197 mg, 0.487 nunol) was

suspended In dry tetrahydrofuran (6 mL) and hexamethylphosphoramide (0.6 mL) under an argon atmosphere. 156 To this mixture was added potassium

hexamethyldisilazide (0.468 mmol, 0.52 mL, 0.9 M In THF) at room temperature. After 15 min of stirring, the bright yellow solution was cooled to -78 °C and 154 (28.3 mg, 0.0975 mmol) In dry tetrahydrofuran (5 mL) was added. The reaction mixture was allowed to warm to ambient temperature and stirred for an additional hour. The reaction mixture was quenched by the addition of sodium sulfate hexahydrate and passed through a silica gel plug. Recrystalllzatlon from acetone yielded 24.8 mg (88.2%) of 156 as a colorless crystalline solid, mp > 230 °C; IR (CHC13, cm"1 ) 2950, 1630, 905; *H NMR (300 MHz, CDC1,) 5 6.12 (dd, J, - 10.4 and 17.2 Hz, 1 H), 4.87 (d, J = 17.2 Hz, 1 H), 4.68 (d, J = 11.1 Hz, 1 H), 3.56-2.94 (series of m, 19 H), 1.57-1.52 (m, 2 H); 13C NMR (75 MHz, CDC1S) ppm 151.30, 104.33, 69.29, 67.87, 67.63, 65.97, 65.74, 62.01, 56.42, 53.41, 52.83, 52.40, 49.93, 32.22? m/z calcd (M+ ) 288.1878, obs 288.1908. 188 Allyl Secododecahedranecarboxylate 159.

Acid 155 (10.6 mg, 0.0346 C0 2 mmol) was dissolved in dry benzene

(1 mL) and l,8-diazabicyclo[5.4.0]- undec-7-ene (2 drops) was added followed by allyl bromide (3 159 drops). The mixture was stirred vigorously at room temperature under argon for 1 h, at which time the mixture was passed through a silica gel column (10% ethyl acetate-petroleum ether elution) giving 11.7 mg of a white solid. Recrystallization from acetone yielded 10.6 mg (88.3%) of 159 as colorless crystals, mp 124.0-124.5 °C; IR (CHC1S, cm"1 ) 2930, 2860, 1740, 1100; XH

NMR (300 MHZ, CDCla) 6 5.98-5.86 (m, 1 H), 5.35-5.18 (m, 2 H), 4.56 (d, J = 5.4 Hz, 2 H), 3.57-2.98 (series of m, 19 H), 1.61-1.51 (m, 2 H); 1SC NMR (75 MHz, CDC1S) ppm 178.95, 132.74, 117.10, 87.27, 69.36, 67.99, 66.19, 65.74, 65.69, 64.86, 61.88, 55.21, 53.08, 52.93, 49.75, 31.95; m/z calcd (M+ ) 346.1933, obs 346.1922. 189 Dodecahedrane (3)

Alcohol 153 (5-6 mg, 23>2 mg total, 0.079 mmol) was placed In a 7 mm I.D. pyrex tube (4 tubes) along with 5% platinum on alumina 3 (420 mg). The glass tube was sealed under high vacuum to a length of approximately 5 cm. The four reaction vessels were vigorously shaken, wrapped in foil, and immersed in a 210 °C silicone oil bath behind a protective shield.

Heating continued for 8 h with interruption every 2 h for physical mixing of the contents. The cooled tubes were opened and the solid was placed directly on the top of a silica gel plug. Elution with benzene (50 mL) gave a white solid which was washed twice with ether to give pure 3 (11.5 mg, 55.6%).

Bromododecahedrane (163). Dodecahedrane (3, 12.7 mg, 0.488 mmol) was dissolved in neat bromine (3 mL) and stirred at room temperature for 24 h. After which 163 time the excess bromine was evapo rated under a stream of argon and the resultant solid was taken up in methylene chloride. The solution was swirled with powdered sodium thiosulfate

until colorless, dried by addition of anhydrous magnesium sulfate, and passed through a short silica gel column (elution with methylene chloride). Solvent removal and recrystallization from ethanol yielded bromododecahedrane (163, 15.7 mg, 95.2%) as colorless needles, mp > 240 °C

(sealed tube); *H NMR (300 MHz, CDCl,) 6 3.99-3.91 (m, 3

H), 3.60 (br s, 6 H), 3.38 (s, 10 H); 13C NMR (75 MHz,

CDCl3) ppm 96.98, 79.98, 66.87, 66.53, 66.09, 65.67; m/z calcd (M+2) 340.0650, obs 340.0628.

Hydroxydodecahedrane (164). To dodecahedrane (3, 3.9 mg, 0.015 mmol) dissolved in dry OH dichloromethane (1 mL) was added an excess amount of lead tetraacetate

164 (approx. 10 mg) followed by trifluoroacetic acid (1 mL) and a small amount of dry lithium chloride. The resulting mixture was stirred in the dark for 24 h at room tempera ture. The reaction mixture was diluted with ether (20 mL) and water (5 mL), the layers were separated, and the aqueous phase was further extracted with ether (3 x 20 mL). The combined organic extracts were washed with water (2 x) 191 and saturated sodium bicarbonate solution (1 x) prior to drying. Filtration and solvent removal gave the trifluoroacetate 165 which was taken up in benzene ( 2 mL) and heated to 100 °C for 2 h in the presence of 10% sodium hydroxide solution (2 mL). The cooled mixture was diluted with water (10 mL) and extracted with dichloromethane (3 x 30 mL). The combined extracts were washed with water and dried. Solvent removal gave a white solid which was triturated with ether (2 x) leaving pure hydroxydodecahe drane (164, 3.1 mg, 75.6%), mp > 250 °C (sealed tube); *H

NMR (300 MHZ, CDCl,) 6 3.55 (br s, 6 H), 3.35 (br S, 10 H) , 3.26-3.20 (m, 3 H), 1.76 (s, 1 H); 1SC NMR (75 MHz, CDCl,) ppm 115.99, 74.95, 66.97, 66.89, 65.71 (2C); m/z calcd (M+ ) 276.1515, obs 276.1482.

Independent Synthesis of Trifluoroacetoxydodecahedrane (165).116 Bromododecahedrane (163, 9.0 mg, 0.027 mmol) was dissolved in trifluoroacetic acid (5 mL), treated with silver trifluoroace- 165 tate (20 mg, 0.09 mmol), and stir

red at room temperature for 20 h. The trifluoroacetic acid was evaporated in vacuo and dichloromethane (15 mL) was added to the residue. The organic phase was washed with water (2 x), brine (1 x) and

dried. Solvent removal yielded a solid which was passed through a short silica gel column. Elution with benzene furnished 7.9 mg (84.0%) of 165 as a white solid, mp 183- 185 °C, dec 220 °C (from hexane); IR (CHC1S,- cm-1) 2940, 1730, 1355, 1340, 1305, 1300, 1230, 1160; *H NMR (300 MHz,

CDCl,) 6 3.58 (br s, 9 H), 3.40 (s, 10 H); ” C NMR (75 MHz, CDCl,) ppm 126.47, 71.65, 66.69, 66.51, 65.90, 65.34 (2

carbons not observed); m/z calcd (M-CF,C02H) 258.1408, obs 258.1407.

Chlorododecahedrane (169).116

A solution of bromide 163 Cl (11.2 mg, 0.033 mmol) in dichloro methane (10 mL) was stirred at room temperature for 4 h in the presence of anhydrous ferric chloride (40 mg). The solution was transferred to a separatory funnel and washed with water prior to drying. Filtration and solvent removal gave a white solid which was passed through a plug of silica gel. Elution with benzene furnished 8.8 mg (90.7%) of 169, mp > 280 °C (from benzene-hexane, 1:3); IR (CHC1,, cm"1 ) 2940, 1300,

1295, 870; lH NMR (300 MHz, CDCl,) 6 3.79-3.71 (m, 3 H),

3.62 (br s, 6 H), 3.38 (s, 10 H); 1SC NMR (75 MHz, CDCl,) 193 ppm 104.39, 78.40, 66.89, 6 6 .6 6 , 66.01, 65.81; m/z calcd (M+ ) 294.1175, obs 294.1140.

Phenyldodecahedrane (170). Bromododecahedrane (163, 4.0 mg, 0.0118 mmol) was dissolved In warm benzene (approx. 32 °C) under an argon atmosphere and treated

170 with anhydrous ferric chloride (-2 mg). The solution, which turned orange immediately, was stirred for 3 h. The mixture was diluted with ether (40 mL), washed with water followed by saturated sodium bicarbonate solution, and dried. Filtra tion and solvent removal yielded 3.75 mg (93.8%) of 170 as a colorless solid, mp 200-201 °C (from ethyl acetate - ace tone); IR (CHClj, cm"1 ) 2950, 2860, 1600, 1495, 1300; *H

NMR (300 MHz, CDCl,) 6 7.74-7.28 (m, 4 H), 7.15-7.09 (m, 1 H), 3.81-3.57 (m, 2 H ) , 3.44 (br s, 10 H); l5C NMR (75 MHz, CDCl,) ppm 154.28, 128.20, 124.80, 124.68, 84.42, 75.15, 67.19, 67.06, 66.71 (1 carbon not observed); m/z, calcd (M+ ) 336.1870, obs 336.1874. Methoxydodecahedrane (172). A. Treatment of 163 with Magic Acid and Methanol Quench.

Bromide 163 (4.2 mg, 0.0124 mmol) was dissolved In deuterated OCH3 chloroform (0.5 mL) and cooled to -78 °C under an argon atmosphere. 172 Sulfuryl chloride fluoride (1 mL) was added followed by fluorosulfonic acid (4 drops) and a spatula tip of antimony pentafluoride. The yellow solution was stirred at -78 °C for 15 min, then poured into a solution of sodium methoxide in methanol at -78 °C. The solution was allowed to warm to room temperature, diluted with water, and extracted with methylene chloride (3 x 30 mL). The dried extracts were concentrated and purified using preparative thin layer chromatography (elution with 5% ethyl acetate: petroleum ether). Rf>0.9: Dodecahedrane (3, 0.4 mg, 12.4%) Rf=0.3: Methoxydodecahedrane (172, 2.2 mg, 61.3%);

IR(CHClj, crn"1) 2930, 1270, 910, 900; lH NMR (300 MHz,

CDCl,) 6 3.50-3.37 (m with s at 5 3.37, 19 H), 3.24 (s, 3 H); 13C NMR (125 MHz, CDCl,) ppm 68.90, 66.90, 66.79, 65.77, 65.48, 51.19 (1 carbon not observed); m/z. calcd (M+ ) 290.1670, obs 290.1678. B. Treatment of 163 with Silver Trlflate in the Presence of Methanol.116 Bromide 163 (1.0 mg, 0.0030 mmol) was dissolved in methylene chloride (2 mL) and methanol (2 mL) and treated with silver trifluoromethanesulfonate (5.0 mg, 0.023 mmol). The mixture was stirred at ambient temperature in the dark for 18 h. Methylene chloride (15 mL) was added and the organic phase was washed with water (2 x) and dried. The solid obtained after concentration was purified using preparative TLC. Elution with 5% ethyl acetate-petroleum ether afforded 172 (0.6 mg, 69.8%); identical in all respects to the material obtained in Part A.

Carbomethoxydodecahedrane (175). Bromide 163 (4.4 mg, 0.013 mmol) was dissolved in methylene C02CH3 chloride (2 mL) and cooled to -78 "C. Fluorosulfonic acid (10 drops) was added followed by a spatula tip of antimony pentafluoride. The resulting yellow mixture was stirred for 15 min and then carbon monoxide gas was bubbled through the solution for 30 min at -78 °C. The mixture was stirred for an additional 15 min, poured onto ice water, and extracted with methylene chloride (3 x 30 mL). The extracts were washed with brine 196 and dried. The filtrate was concentrated to a volume of approximately 10 mL, diluted with ethyl ether (50 mL), and treated with excess diazomethane (generated from Diazald). After 1 h, the excess diazomethane was carefully destroyed with acetic acid and the solution was diluted with ether (100 mL). The ether layer was washed with water, saturated sodium bicarbonate solution and brine, prior to drying. Filtration and solvent removal gave the crude ester as a white solid. A total of 11.4 mg of 163 was treated in this manner (two reactions) and the combined products were submitted to MPLC purification. Elution with 2% ethyl acetate:petroleum ether yielded 6.3 mg (59.8%) of pure 175; mp 192-193 °C (from ethyl acetate); IR (CHCl,, cm-1) 2940, 2860, 1715, 1430, 1285; XH NMR (300 MHz, CDC1S) 5 3.77-3.66

(m, 3 H), 3.67 (s, 3 H), 3.50 (br s, 6 H), 3.41 (bs, 10 H); 1SC NMR (75 MHz, CDCl,) ppm 179.29, 84.54, 70.97, 66.99,

66.92, 66.86 (2 c), 51.99; m/z Cl (M+l) 319 (100%).

Acetylarainododecahedrane (166). Bromododecahedrane (163, 10.3

mg, 0.0304 mmol) was refluxed in NHCCH acetonitrile (10 mL) in the pre sence of silver trifluoromethane- 166 sulfonate (approx 150 mg) in the dark for 20 h. After cooling, 197 water (10 mL) was added and the solution was stirred 15 min followed by extraction into dichloromethane (3 x 30 mL). The combined extracts were washed with brine and dried. Filtration and solvent removal gave 9.6 mg (99.6%) of amide 166 as a white solid, mp > 250 °C (from ethyl acetate); IR

(CHCI3, cm-1) 2950, 1660, 1490, lH NMR (300 MHz, CDC13) 6

5.81 (bs, 1 H), 3.55 (bs, 6 H), 3.37-3.28 (m, with s at 3.37, 13 H), 1.92 (s, 3 H); 1SC NMR (125 MHz, CDCl,) ppm

168.95, 95.77, 74.21, 66.80, 6 6 .6 8 , 66.16, 66.07, 23.80; m/z calcd (M+ ) 317.1780, obs 317.1772.

Fluorododecahedrane (176).116 To bromide 163 (7.0 mg, 0.021 mmol) in methylene chloride (2 mL) and ether (2 mL) was added silver tetrafluoroborate (20 mg). The 176 mixture was stirred for 20 h in the dark after which time ether (5 mL) was added. The organic layer was washed twice with water and dried. Filtration and concentration yielded a solid which was passed through a plug of silica gel. Elution with benzene gave 5.1 mg (89.5%) of 176 as a white solid, mp > 260 °C (from chloroform); IR (CHCl,, cm"1 ) 2950, 1340,

1300; JH NMR (300 MHz, CDCl,) 6 3.55 (br s, 9 H), 3.37 (br s, 10 H); l,C NMR (125 MHz, CDCl,) ppm 137.64 (d, J_= 198 192.93 Hz), 71.26 (d, J = 23.61 Hz), 66.76, 65.68, 65.56, 65.17 (d, J = 3.01 H z ); l9F NMR (75 MHz, CDC1S) ppm 121.92; m/z 278.1487, obs 278.1476.

Methyldodecahedrane (152). Bromide 163 (7.0 mg, 0.021 mmol) was dissolved In hexane (5 mL) and treated withneat tri-

methylalumlnum (10 drops) at room 152 temperature for 24 h. The excess reagent was quenched by careful addition of methanol with cooling. The mixture was diluted with benzene and 3 N hydrochloric acid. The layers were separated and the organic phase was washed with water and dried. The residue obtained was passedthrough a silica gel plug (elution withbenzene). Concentration gave 5.1 mg

(90.1%) of 152 as a white solid; *H NMR (300 MHz, CDCl,) 6 3.38 (br s, 16 H), 2.92 (m, 3 H), 1.14 (s, 3 H); 1SC NMR (75 MHz, CDCl,) ppm 77.55, 74.67, 67.17, 67.06, 66.74,

66.39, 32.60. 199 Hotnoazadodecahedryl Azide 178 Bromododecahedrane (163, 4.0

£ a mg, 0.0118 mmol) was dissolved in deuterated chloroform (0.5 mL) and trimethylsilylazide (0.5 mL) was added under an argon atmosphere. 178 The solution was cooled to 0 °C and treated with tin tetrachloride (3 drops). After 15 min another aliquot of tin tetrachloride (3 drops) was added and stirred 45 min. The mixture was quenched by addition of ice water and extracted with dichloromethane (3 x 20). The combined extracts were washed with saturated sodium bicarbonate solution and dried. A total of 8.0 mg of 51 was treated in this manner (two reactions) giving 7.1 mg of a yellow solid. Trituration twice with ether gave 5.8 mg (77.8%) of pure 178 as a colorless solid, mp > 250 °C (from acetone); IR (CHCl,, cm-1) 2940, 2100, 1425, 1350, 900; XH

NMR (300 MHz, CDCls) 6 4.20 (t, J = 10.2 Hz, 1 H), 3.48-

3.33 (series of m, 13 H), 3.02-2.60 (series of m, 6 H); 1SC

NMR (75 MHz, CDCl,) ppm 87.76, 67.92, 6 6 .8 8 , 66.80, 64.37, 63.92, 63.28, 63.03, 59.43, 54.39, 49.90, 47.02; m/z calcd (M+-N2) 288.1626, obs 288.1674; FAB(M+1) 317.27. 200 (Dichloromethyl)dodecahedrane (182). 116 To a mixture of dodecahedrane (3, 4.5 mg, 0.017 mmol), 50% CHCI2 aqueous sodium hydroxide (5 mL), and triethylbenzylammonium chloride 182 (0.5 mg) in benzene heated to 50 °C

was added chloroform (2 mL) over 6 h. The mixture was further heated overnight. After cooling, the mixture was diluted with water and extracted with benzene (3 x 10 mL). The combined organic extracts were washed with brine and dried. The residue obtained was purified by preparative TLC. Elution with hexane afforded 2.5 mg (42.2%) of 182 as a colorless solid, mp 209-210 °c (hexane); IR (CHCl,, cm"1 ) 2950, 1300; XH NMR (300 MHz,

CDCl,) 6 5.71 (s, 1 H), 3.80-3.30 (m, 19 H); 1SC NMR (75 MHz, CDCl,) 89.53, 82.32, 66.98, 66.61, 66.30 (2 carbons not observed); m/js calcd (M+2) 342.0942, obs 342.0914. 201 Dodecahedryl Nitrate (184).

Dodecahedrane (3, 3.4 mg, 0.013 mmol) was taken up In deu- teriochloroform (3 mL) under an argon atmosphere and treated with an excess amount of nltronium tetrafluoroborate (heaping spa tula). The heterogeneous mixture was stirred vigorously for 68 h, then diluted with dichloromethane and water. The layers were separated and the aqueous layer was further extracted with dichloromethane (3 x 20 mL). The combined organic extracts were washed with brine and dried. Filtration and removal of solvent gave a yellow solid which was passed through a plug of silica gel (elution with dichloromethane). Recrystallization from ethyl acetate yielded 3.5 mg (83.5%) of 184 as colorless crystals, mp 220

°C dec; IR (CHCl,, cm”1 ) 3060, 2950, 1615, 1300, 1270; lH

NMR (300 MHz, CDCl,) 6 3.56 (br S, 9 H), 3.40 (br s, 10 H) ; 15C NMR (75 MHZ, CDCl,) ppm 129.89, 70.33, 66.67, 66.47, 65.94, 65.19; m/z CI(CHJ 320 (M-l, 1.57%), 275 (M-NO,, 10.45%), 259 (M-ONO,, 100%). 202 Triroethylsilyloxyazido Secododecahedrane 189 .

Aldehyde 154 (18.2 mg, 0.0627 mmol) was dissolved in dimethylformamide (1 mL) under an argon atmosphere and trimethylsilylazide (0.5 mL) was added along with zinc 189 chloride (2 crystals). The mixture was stirred at room temperature for 4 h and diluted with ether (50 mL). The ether layer was washed with water and saturated sodium bicarbonate solution followed by drying. Filtration and solvent removal gave siloxy azide 189 as a colorless oil which was used without further purification; IR (CHCl,, cm-1) 2940, 2860, 2100, 1255, 1135, 1110, 1100,

895, 880, 850; JH NMR (300 MHz, CDCl,) 6 4.68 (s, 1 H), 3.55-2.91 (series of m, 19 H), 1.56-1.50 (m, 2 H), 0.21 (s, 9 H); m/z calcd (M+-Na) 377.2175, obs 377.2197. 203 Aminomethyl Secododecahedrane Hydrochloride 185. To lithium aluminum hydride .c h 2n h 2 (90 mg) suspended in dry ethyl ether (3 mL) under an argon atmosphere was added dropwise a ... solution of crude siloxyazide 189 1 B5 in ether (5 mL). The mixture was stirred at room temperature for 1 h and refluxed for 4 h. After cooling, the excess hydride was quenched by careful addition ofsaturated ammonium chloride solution, and aqueous sodium hydroxide solution (10 mL, IN) was added until the mixture was strongly basic. The solution was extracted with ether (3 x 50 mL), and the organic extracts were washed with brine (30 mL) and dried. Filtration and solvent removal gave the free amine as a white solid which was taken up in dry ether (20 mL) and ethanol saturated with gaseous hydrochloric acid (0.5 mL) was added. Amine hydrochloride 185 precipitated out of solution immediately

and was collected after stirring for an additional 1 hour. The white solid was taken up in methanol and reprecipitated by the addition of ether. This procedure was repeated three times yielding 13.3 mg (64.5% from the aldehyde) of

185, mp > 240 °C; XH NMR (300 MHz, CDsOD) 6 3.62-3.73 (series of m, 24 H), 1.65-1.55 (m, 2 H); 1SC NMR (75 MHz, CDjOD) ppm 82.02, 70.42, 69.11, 67.29, 66.77, 66.48, 63.19, 204 55.33, 54.07, 53.72, 51.24, 51.08, 32.93; m/z calcd (M+ , FAB) 292.2066, obs 292.2081.

Amino Secododecahedrane Hydrochloride 186 .

Acid 155 (13.7 mg, 0.0447 .n h 2*hci mmol) was suspended in acetone (5 mL) and triethylamine was added until the acid completely dis solved. The solution was cooled to 186 0 °C and ethyl chloroformate (10 drops) was added along with water (2 mL). The solution was stirred for 1 h followed by the addition of an excess amount of sodium azide in water. After 1 h, a white solid appeared which dissolved upon the addition of ether. The layers were separated and the aqueous layer was extracted with ether (2 x 30 mL). The combined extracts were washed with water and dried. Filtration and addition of toluene (15 mL) followed by concentration in vacuo to approximately 10 mL gave the acyl azide which was slowly heated to 110 °C under argon and refluxed overnight. Concentration gave the isocyanate as a yellow oil; IR (CHC1S, cm-1) 2940, 2260, 1700, 1380, 1330. The oil was taken up in tetrahydrofuran

(3 mL) and 6 M hydrochloric acid (2 mL) was added, and the mixture was refluxed for 2 h. The tetrahydrofuran was distilled out of the reaction vessel and the mixture was heated to 90 °C for 6 h. Evaporation in vacuo gave a brown semi-solid which was triturated with hot acetone and methanol to give the amine hydrochloride as an off-white solid. Purification by reprecipitation from methanol and ether (3 x) yielded 10.3 mg (69.6%) of 186 , mp >240 °C;

NMR (300 MHz, CD,OD) 6 3.65-2.99 (series of m, 22 H), 1.69- 1.55 (m, 2 H); 1SC NMR (75 MHz, CD,OD) ppm 98.15, 69.98,

68.35, 68.06, 6 6 .8 6 , 66.34, 62.76, 57.40, 53.77, 53.16, 50.68, 32.51; m/z calcd (M+ , FAB) 278.1909, obs 278.1880.

Carboxydodecahedrane (168). Ester 175 (6.3 mg, 0.020 mmol) was dissolved in methanol (5 mL) c o 2h with the aid of gentle heating. A solution of sodium hydroxide (0.4 g) in water ( 2 mL) was added and stirred at room temperature for 18 h. The solution was acidified with 6 M hydrochloric acid, diluted with water (20 mL), and extracted with methylene chloride (4 x 30 mL). The combined extracts were washed with brine and dried. Filtration and solvent removal yielded a white solid which was triturated twice with ether to give pure acid 168 (5.7 mg, 95.0%), mp > 250 °C; *H NMR

(300 MHz, CDCl, and DMSO-d6) 6 3.22-3.20 (m, 3 H), 2.97-

2.94 (m, 6 H), 2.88 (br s, 10 H ) ; l3C NMR (75 MHz, CDC1S 206 and DMSO-dJ ppm 179.23, 83.14, 69.52, 65.52, 65.47; m/z calcd (M+ ) 304.1463, obs 304.1449.

Araidododecahedrane (191)

Carboinethoxydodecahedrane (175, 11.0 mg, 0.034 mmol) was taken up In dry dichloromethane (3 mL) and treated with a stock 191 solution of dimethylaluminum amide (3 mL, approx. 1 M solution in dichloromethane) at room temperature for 45 h. The mixture was quenched with methanol, diluted with saturated ammonium chloride solution and stirred to dissolve the aluminum salts. The layers were separated and the aqueous phase was further extracted with methylene chloride (4 x 20 mL). The combined organic extracts were washed with brine and dried. Filtration and solvent removal yielded a white solid which was triturated with hexane giving 9.1 mg (86.7%) of pure 191, mp > 250 °C; IR (CHC1S, cm"1) 3530, 3410, 2950, 1660,

1580, 1365; lH NMR (300 MHz, CDCl,) 6 5.22 (bs, 2 H ) , 3.77-

3.67 (m, 3 H), 3.53 (m, 6 H), 3.42 (s, 10 H); 1SC NMR (75 MHz, CDCl,) ppm 181.58, 85.61, 71.19, 66.98, 66.96, 66.75 (1 carbon not observed); m/z calcd (M+ ) 303.1623, obs

303.1640. 207 (tert-Butoxycarbonyl)aminododecahedrane (192)

Amide 191 (10.4 mg, 0.0343 mmol) was dissolved in tert-butyl

NHC02*"Bu alcohol (10 mL) and heated to 50 °C

in the presence of bis(trifluoroacetoxy)iodobenzene (50 mg) under 192 Ar for 22 h. The mixture was diluted with water and extracted with ether (3 x 20 mL). The combined extracts were washed with water and brine prior to drying. Filtration and solvent removal yielded a white solid. The product was isolated from this solid by trituration with hexane (3 x 15 mL). Concentration of the triturates yielded 9.3 mg of 192 which was used without further purification; IR (CHClj, cm-1) 3440, 2950, 1700,

1480, 1370, 1170; NMR (300 MHZ, CDCla) 6 5.30 (br s, 1

H), 3.55 (br s, 6 H), 3.36 (s, 10 H), 3.25-3.21 (m, 3 H), 1.45 (s, 9 H); m/z calcd (M+-t-Bu) 318.1515, obs 318.1494.

Aminododecahedrane Hydrochloride (187). Crude carbamate 192 (9.3 mg, 0.0338 mmol) was taken up in ether (2 mL) and treated with ethanol

saturated with hydrogen chloride (2 187 mL). The solution was stirred 2 h and concentrated. The residue was 208 triturated with ether (2 x) and reprecipitated with methanol and ether to give 4.5 mg (42.1%) of 187 as a white solid, mp > 250 °C? AH NMR (300 MHz, CDCl,) 6 3.64 (br s, 6 H), 3.46 (s, 10 H), 3.35-3.31 (m, 3 H); 1SC NMR (75 MHz, CD,OD) ppm 94.99, 73.14, 68.13, 67.99, 67.35, 67.26; m/z calcd (M+ , FAB) 276.29 (100%).

Cyanododecahedrane (171).

To amidododecahedrane (191, 2.0 mg, 0.0067 mmol) in pyridine (3 mL) was added thionyl chloride (10 drops) and stirred at ambient tem 171 perature for 16 h. The solution was diluted with ethyl ether, water, and 1 N hydrochloric acid. The layers were separa ted and the aqueous layer was further extracted with ether (3 x 20 mL). The combined extracts were washed with 1 N hydrochloric acid, water, brine, and dried. The resulting solid obtained was passed through a silica gel plug (elu tion with methylene chloride) to give the nitrile. Tri turation with hexane gave 1.7 mg (90.4%) of pure 171, mp > 250 °C; IR (CHCl,, cm-1) 2940, 2220; lH NMR (300 MHz,

CDCl,) 6 3.38-3.76 (m, 3 H), 3.59 (br s, 6 H), 3.42 (s, 10 H); 1SC NMR (125 MHz, CDCl,) ppm 128.00, 73.91, 67.03, 209 66.87, 66.67 (two carbons not observed); m/z, calcd (M+ ) 285.1517, obs 285.1520.

Hydroxyinethyldodecahedrane (195).

Carbomethoxydodecahedrane (175, 5.3 mg, 0.017 mmol) was c h 2o h dissolved in dry benzene (2 mL) and treated with diisobutylaluminum 195 hydride (15 drops). The solution was stirred for 23 h under an argon atmosphere after which time excess hydride was destroyed by addition of methanol. The mixture was diluted with 1 N hydrochloric acid (5 mL) and extracted with methylene chloride (3 x 20 mL). The combined extracts were washed with brine and dried. Filtration and solvent removal afforded a white solid which was triturated with hexane (2x) leaving 3.5 mg (72.9%) of pure 195, mp> 250 °C; IR (CHCl,, cm-1) 3500 (br), 2960, 1300; lH NMR (300 MHZ, CDCl3) 5 3.40 and 3.39 (two s, total 18 H), 3.09 (br s, 3 H); 15C NMR (125 MHz, CDC1S) ppm 81.86, 71.06, 68.91, 67.10, 67.02, 66.56, 65.54; m/z calcd (M+ ) 290.1681, obs

290.1659. Dodecahedrylcaboxaldehyde (199).

To a suspension of pyridinium

chlorochromate (10 mg), sodium acetate (10 mg), and Celite (10 mg) in dry methylene chloride was added igg alcohol 195 (7.2 mg, 0.0248 mmol) in the same solvent (2 mL) and stirred at room temperature for 1 h. The mixture was diluted with ether (5 mL) and passed through a plug of silica gel (elution with dichloromethane). Concentration yielded a white solid which was purified by trituration with hexane to give 4.7 mg (66.2%) of pure 199, mp > 250 °C; IR (CHCl,, cm-1) 2940, 2800, 2700, 1695, 1300, 905,

720; *H NMR (300 MHz, CDCl,) 5 9.48 (s, 1 H), 3.58 (br S, 3 H), 3.44 (s, 16 H); 13C NMR (125 MHz, CDCl,) ppm 201.49,

90.00, 67.23, 67.11, 66.96, 66.91, 6 6 .8 6 ; m/z calcd (M+ ) 288.1514, obs 288.1519. REFERENCES AND NOTES 1. a) Garratt, P.J.j White, J.F. J. Orq. Chem. 1977, 42. 1733. b) Nickon, A.; Pandit, G.D. Tetrahedron Lett. 1968, 3663.

2. Hirao, K.; Ohuchi, Y.; Yoneraitsu, O. J. Chem. Soc.. Chem. Comm. 1982, 99.

3. Doering, W. von E.; DePuy, C.H. J. Am. Chem. Soc. 1953, 25, 5955.

4. DePuy, C.H.; Ponder, B.W. J. Am. Chem. Soc. 1959, 81. 4629.

5. Klinsmann, U.j Gauthier, J.; Schaffner, K.; Pasternak, M.; Fuchs, B. Helv. Chim. Acta 1972, 55. 2643.

6. Pfund, R.A.j Ganter, C. Helv. Chim. Acta 1979, 62. 228.

7. a) Tombo, G.M.R.; Chakrabarti, S.; Ganter, C. Helv. Chim. Acta. 1983, 92, 914. b) Doecke, C.W.; Garratt, p.j. Tetrahedron Lett. 1981, 1051.

8. Garratt, P.J.; Doecke, C.W.; Weber, J.C.; Paquette, L.A. J. Orq. Chem. 1986, 51, 449.

9. Ammann, von W.; Ganter, C. Helv. Chim. Acta 1977, 60, 1924.

10. Profitt, J.A.; Watt, D.S.; Corey, E.J. J. Org. Chem. 1975, 40, 127. 11. Similiar cyclization pathway,seen-in.the synthesis of 5,12-Dioxapentacyclo[6 .4.0.0*' .0 ' .0*'3 ]dodecane (a), Ammann, von W.j Jfiggi, F.J.; Ganter, C. Helv. Chim. Acta 1980, 63., 2019.

a 211 212

12. Houk, K.N.; Brown, F., private communication. 13. a) Bhatt, M.V.; Kulkarni, S.U. Synthesis 1983, 24. b) Jung, M.E.; Lyster, M.A. J. Orq. Chem. 1977, 42., 3761. c) Funk, R.L.; Mossman, C.J.; Zeller, W.E. Tetrahedron Lett. 1984, 1655.

14. a) Arapakos, P.O. J. Am. Chem. Soc. 1967, 89, 6794. b) Marshall, J.A.; Bierenbaum, R. J. Ora. Chem. 1977, 42, 3309.

15. Miller, A.E.G.; Biss, J.W.; Schwartzman, L.H. J. Orq. Chem. 1959, 24, 627. 16. a) Baird, M.C.; Nyman, C.J.; Wilkinson, G. J. chem. Soc. A 1968, 348. b) Ohno, K.; Tsuji, J. J. Am. Chem. Soc. 1968, 90, 99. 17. a) Paquette, L.A.; Uchida, T.; Gallucci, J.C. J. Am. Chem. Soc. 1984, 106, 335. b) Lampman, G.M.; Aumiller, J.C. Orq. Svn. 1971, 51, 106; Bunce, N.J. J. Orq. Chem. 1972, 37.# 664. 18. Matsuda, I.; Murata, S.; Ishii, Y. J. Chem. Soc.. Perkin I 1979, 26.

19. Kaiser, E.W.; Hauser, C.R. J. Orq. Chem. 1968, 33., 3402. 20. Taguchi, H.; Shimoji, K.; Yamamoto, H.; Nozaki, H. Bull. Chem. Soc. Jpn. 1974, 47, 2529 and references cited therein. 21. Hudrlik, P.F.; Peterson, D. J. Am. Chem. Soc. 1975, 97. 1464 and references cited therein. 22. Magnus, P.D.; Roy, G. Organometallics 1982, jL, 553. 23. Paquette, L.A.; Browne, A.R.; Doecke, C.W.; Williams, R.V. J. Am. Chem. Soc. 1983, 105. 4113.

24. Paquette, L.A.; Fischer, J.W.; Browne, A.R.; Doecke, C.W. J. Am. Chem. Soc. 1985, 107. 6 8 6 .

25. a) Paquette, L.A.; Carr, R.V. J. Am. Chem. Soc. 108, 102. 7553. b) Paquette, L.A.; Carr, R.V.C.; charumilind, P.; Blount, J.F. J. Org. Chem. 1980, 45. 4922. 213

26. a) Prinzbach, H.; Sedelmeier, 6 .; Kruger, G.; Goddard, R.; Martion, H.-D.; Gleiter, R. Anqew. Chem.. Int. Ed. Engl.. 1978, 12, 171. b) Jones, G.,111; Becker, W.G.; Chiang, S.-H. J. Photochem. 1982, 19, 245. 27. a) Nagarkatti, J.P.; Ashley, K.R. Tetrahedron Lett. 1973, 4599. b) Mori, K. Tetrahedron 1978, 34* c) Kim, M.Y.; Starrett, J.E., Jr.; Weinreb, S.M. J. Org. Chem. 1981, 46, 5383. 28. Grieco, P.A.; Masaki, Y.J. J. Org. Chem. 1974. 39. 2135. 29. Paquette, L.A.; Ktinzer, H.; Green, K.E. J. Am. Chem. Soc. 1985, 107. 4788. 30. Brow, A.C.; Carpino, L.A. J. Org. Chem. 1985, 50. 1749. 31. Schaefer, J.P.; Bloomfield, J.J. Org. React. 1967, 15, 1. 32. Bloomfield, J.J.; Owsley, D.C.; Nelke, J.M. Org. React. 1976, 23, 259. 33. Whitesides, G.M.; Gutowski, F.D. J. Org. Chem. 1976, 41, 2882. 34. Paquette, L.A. Org. React. 1977, 25, 1. 35. See ref. 34, footnotes 107-116 for examples.

36. Borch, R.P. J. Org. Chem. 1969, 34.* 627. 37. a) Paquette, L.A.; Ternansky, R.J.; Balogh, D.W.; Kentgen, G. J. Am. Chem. Soc. 1983, 105, 5446. b) Iriarte, J.? Schaffner, K.; Jeger, O. Helv. Chlm. Acta 1963, 175. 1599. c) Mori, T.; Matsui, K . ; Nozak, H. Tetrahedron Lett. 1970, 1175. 38. DeBoer, C.D.; Herkstroeter, W.G.; Marchetti, A.P.; Schultz, A.G.; Schlessinger, R.H. J. Am. Chem. Soc. 1973, 95, 3983. 214 39. a) Bernassan, J.-M.; F&tizon, M. Synthesis 1975. 795. b) Wagner, P.J.; Hammond, G.S. J. Am. Chem. Soc. 1965, 87, 4009. c) Orban, X.; Schaffner, K.; Jeger, 0. J. Am. Chem. Soc. 1963, 85_, 3033. d) Yang, N.C.; Morduchowitz, A.; Yang, D.-D. H. Ibid. 1963, 85. 1017. 40. Paquette, L.A.; Balogh, D.W. J. Am. Chem. Soc. 1982, 105. 774 and references therein.

41. a) Padwa, A.; Eisenberg, W. J. Am. Chem. Soc. 1970, 92. 2590. b) Meador, M.A.; Wagner, P.J. J. Am. Chem. Soc. 1983, 105. 4484.

42. Jorgenson, M.J. Org. React. 1970, 18., 1. 43. Borowitz, I.J.; Anschel, M.; Firstenberg, S. J. Org. Chem. 1967, 32, 1723.

44. Engel, P.; Weber, J.C.; Paquette, L.A. 2. Kristalloqr. 1987, xx, xx.

45. Grieco, P.A.; Gilman, S.; Nighlzania, M. J. Org. Chem. 1976, 41, 1485.

46. Kricka, L.J.; Ledwith, A. Synthesis 1974, 539. 47. Paquette, L.A.; Coghlan, M.J.; Lin, H.-S. Org. Syn.. submitted for publication.

48. Sreekumar, C.; Darst, K.P.; Still, W.C. J. Org. Chem. 1980, 45, 4260.

49. Ono, N.j Yamada, T.; Saito, T.; Tanaka, K.j Kaji, A. Bull. Chem. Soc. Jpn. 1978, 51, 2401.

50. a) Verheijdt, P.L.; Cerfontain, H. Reel. Trav. Chim. Pays-Bas 1983, 102. 173. b) Sarphatie, L.A.; Verheijdt, P.L.; Cerfontain, H. Ibid. 1983, 102, 9. c) Urry, W.H.; Trecker, D.J.; Winey, D.A. Tetrahedron Lett. 1962, 609.

51. a) Ternansky, R.J.; Balogh, D.W.? Paquette, L.A. J. Am. Chem. Soc. 1982, 104. 4503.

52. a) Wyvratt, M.J. Ph.D. Dissertation, The Ohio State University, 1976. b) Balogh, D.W. Ibid.. 1980. c) Ternansky, R.J. Ibid.. 1982. 215

53. Fessner, von W.-D.j Murty, D.A.R.C.; Wdrth, J.; Hunkier, D.; Fritz, H.; Prlnzbach, H.; Roth, W.D.; Schleyer, P. von R.; McEwen, A.B.; Maier, W.F. Anqew. Chem. 1987, 99, 484.

54. Paquette, L.A.; Wyvratt, M.J.; Berk, H.C.; Moerck, R.E. J. Am. Chem. Soc. 1978, 100. 5845.

55. Paquette, L.A.; Wyvratt, M.J.; Schallner, O.; Muthard, J.L.; Begley, W.J.; Blankenship, R.M.; Balogh, D. J. Org. Chem. 1979, 44, 3616.

56. Paquette, L.A.; Miyahara, Y. J. Org. Chem. 1987, 52. 1265.

57. Paquette, L.A.; Balogh, D.W.; Ternansky, R.J.; Begley, W.J.; Banwell, M.G. J. Org. Chem. 1983, 48. 3282.

58. Paquette, L.A.; Miyahara, Y.; Doecke, C.W. J. Am. Chem. Soc. 1986, 108. 1716.

59. Paquette, L.A.; Balogh, D.W. J. Am. Chem. Soc. 1982, 104. 774.

60. Paquette, L.A.; Ternansky, R.J.; Balogh, D.W.; Taylor, W.J. J. Tun. Chem. Soc. 1983, 105. 5441.

61. a) Clark, R.J.H. in "Comprehensive Inorganic Chemistry," Vol. 3, Ch. 32, 1973. b) Fanelli, A.J.; Maeland, A.J.; Rosan, A.M.; Crissey, R.K. J. Chem. Soc.. Chem. Comm. 1985, 8 .

62. Corey, E.J.; Schmidt, G. Tetrahedron Lett. 1979, 399. 63. a) Lindgren, B.O.; Nilsson, T. Acta. Chemica. Scand. 1973, 27, 8 8 8 . b) Corey, E.J.; Myers, A.G. J. Am. Chem. Soc. 1985, 107. 5574.

64. Nwaukwa, S.O.; Keehn, P.M. Tetrahedron Lett. 1982, 3131.

65. For reviews on adamantane chemistry see; a) Fort, R.C., Jr.; Schleyer, P. von R. Chem. Rev. 1964, 64, 277. b) Sevost'yanova, V.V.; Krayushkin, M.M.; Yurchenko, A.G. Russ. Chem. Rev. 1970, 3 9 817. 66. Stetter, H.; Schwartz, M.; Hirschhorn, A. Chem. Ber. 1959, 92, 1629. 216 67. Partch, R.E. J. Am. Chem. Soc.. 1967, 89, 3662.

6 8 . Jones, S.R.; Mellor, J.M. J. Chem. Soc. Perkin I 1976, 2576. 69. a) Takaishi, N.; Fujlkura, Y.; Inamoto, Y. Synthesis 1983, 293. b) Tori, M.; Matsuda, R.; Asakawa, Y. Tetrahedron Lett. 1985, 227. 70. Stetter, H.; Mayer, J.; Schwarz, M. ; Wulff, K. Chem. Ber. 1960, 93, 226.

71. Koch, H.; Haaf, W. Ora. Svn. 1964, 44/ 1* 72. a) Olah, G.A.; Faroog, 0.; Prakash, G.K.S. Synthesis 1985, 1140. b) Reetz, M.T.; Chatziiosifidis, I.; Kfinzer, H.; Starke, H.M. Tetrahedron 1983, 961. 73. Jones, N.J.; Deadman, W.D.; LeGoff, E. Tetrahedron Lett. 1973, 2087. 74. a) Olah, G.A.; Lukas, J. J. Am. Chem. Soc. 1968, 90, 933. b) Olah, G.A. et al. Ibid 1985, 107.~2764. 75. a) Souma, Y. ; Sano, H. Bull. Chem. Soc. Jpn. 1976, 49. 3335. b) Yoneda, N.; Fukuhara, T.; Takahashi, Y.; Suzuki, A. Chem. Lett. 1983, 17. 76. Gallucci, J.C.; Doecke, C.W.; Paquette, L.A. J. Am. Chem. Soc. 1986, 108, 1343. 77. Takeuchi, K.j Moriyama, T.; Kinoshita, T.; Tachino, H.; Okamoto, K. Chem. Lett. 1980, 1395. 78. Fry, A.J.; Migrom, Y. Tetrahedron Lett. 1979, 3357. 79. Della, E.W.; Bradshaw, T.K. J. Ora. Chem. 1975, 40. 1638. 80. Prakash, G.K.S.; Stephenson, M.A.; Shih, J.G.; Olah, G.A. J. Org. Chem. 1986, 51, 3217. 81. a) Brown, H.C.; Midland, M.M.; Levy, A.B. J. Am. Chem. Soc. 1972, 94, 2114. b) Suzuki, A.; Sono, S.; Itoh, M.; Brown, H.C.; Midland, M.M. Ibid. 1971, 93, 4329. 217 82. a) Becker, K.B.; Gabutti, C.A. Tetrahedron Lett. 1982, 1883. b) Sasaki, T . ; Eguchi, S.; Okano, T. Ibid. 1982. 4969. c) Radziszewski, J.G.; Downing, J.W.; Jawdosink, M.; Kovacic, P.; Michl, J. J. Am. Chem. Soc. 1985, 107. 594.

83. a) Toda, M.; Hirata, Y. J. Chem. Soc.. Chem. Comm. 1970, 1597. b) Quast, H.; Eckert, P. Liebigs Ann. Chem. 1974, 1727.

84. Tabushi, I.; Yoshida, Z .; Takahaski, N . J. Am. Chem. Soc. 1970, 92, 6670.

85. Marino, J.P.; Laborde, E. J. Org. Chem. 1987, 52, 1. 86. Olah, G.A.; Lin, H.C. J. Am. Chem. Soc. 1971, 93. 1259.

87. a) Hoffman, C.E.; Haff, R.P.; Neumayer, E.M. Fed. Prod. 1964, 23, Abstract No. 1716. b) Jackson, G.G.; Muldoon, R.L.; Akers, L.W. Antimlcrob. Agents Chemother. 1963, 703. c) Lundahl, K.; Schut, J.; Schlatmann, J.L.M.A.; Paerels, G.B.; Peters, A. J. Med. Chem. 1972, 15, 129. d) Aldrich, P.E.; Hermann, E.C.; Meier, W.E.; Paulshock, M.; Prichard, W.W.; Snyder, J.A.; Watts, J.C. Ibid. 1971, 14,, 535. 88. Schwab, R.S.; England, A.C., Jr.; Poskanzer, D.C.; Young, R.R. J. Am. Med. Ass. 1969, 208, 1168.

89. Rail, D.P.; Zubrod, C.G. Ann. Rev. Pharmacol. 1962, 2, 115.

90. See reference 86 and references cited therein. 91. a) Birkofer, L.; Mliller, F.; Kaiser, W. Tetrahedron Lett. 1967, 2781. b) Evans, D.A.j Truesdale, L.K. Ibid. 1973, 4929.

92. Kyba, E.P.; John, A.M. Tetrahedron Lett. 1 9 7 7 . 2737. 93. Private communication, Eli Lilly and Company. 94. a) Kovacic, P.; Roskos, P.D. J. Am. Chem. Soc. 1969, 91, 6457. b) Kovacic, P.; Goralski, C.T.; Hiller, J.J., Jr.; Levisky, J.A.; Lange, R.M. Ibid. 1965, 87, 1262. 95. Sasaki, T.; Eguchi, S.; Kateda, T. J. Org. Chem. 1974, 39, 1239. 218

96. Aigami, K.; Inamoto, Y.; Takaishi, N. t Fujikura, Y. J. Med. Chem. 1976, 19, 536. 97. Khuong-Huu, Q.; Monneret, C.; Kabor&, I.; Goutarel, R. Tetrahedron Lett. 1971, 1935. 98. Fujinaga, M.; Matsushima, Y. Bull. Chem. Soc. Jpn. 1966, 39, 185. 99. Gutzwiller, J.; Uskokovic, M. J. Am. Chem. Soc. 1970, 92, 204. 100. a) Kishi, Y.; Fukuyama, T.; Aratani, M.j Makatsubo, F.; Goto, T.j Inoue, S.; Tanino, H.; Sugiura, S.; Kakoi, H. J. Am. Chem. Soc. 1972, 94* 9221. b) Paquette, L.A.; Kakihana, T.; Hansen, J.F.; Philips, J.C. Ibid. 1971, 93, 152. 101. Meerwein, H. Org. Syn. 1966, 46., 113. 102. a) Shioiri, T. ; Ninomiya, K.; Yamada, S. J. Am. Chem. Soc. 1972, 94, 6203. b) Eaton, P.E.j Shankar, B.K.R.; Price, G.D.; Pluth, J.J.; Gilbert, E.E.; Alster, J.? Sandus, O. J. Org. Chem. 1984, 49, 185. 103. Hogberg, T.; Strom, P.; Ebner, M.; Ramsby, S. J. Org. Chem. 1987, 52, 2033. 104. Basha, A.; Lipton, M.; Weinreb, S.M. Tetrahedron Lett. 1977, 4171. 105. a) Della, E.W.; Kasum, B.; Kirkbride, K.P. J. Am. Chem. Soc. 1987, 109. 2746. b) Radhakrishna, A.S.; Parham, M.E.; Riggs, R.M.; Loudon, G.M. J. Org. Chem. 1979, 44/ 1746. 106. a) Brown, H.C.; Heim, P. J. Org. Chem. 1973, 38, 912. b) Brown, H.C.; Narasimhan, S.; Choi, Y.M. Synthesis 1981, 996. c) Brown, H.C.; Choi, Y.M.; Narasimhan, S. J. Org. Chem. 1982, 47, 3153. 107. Reviews: Sternhell, S. Rev. Pure Appl. Chem. 1964, 14, 15 and Garbisch, E. J. Am. Chem. Soc. 1964, 84. 5561.

108. Pehk, T.; Lippmaa, E. Org. Magn. Resonance 1971, 3., 679. 219 109. Pehk, T.; Lippmaa, E.; Sevostjanova, V.V.; Krayuschkin, M.M.; Tarasova, A.I. Ora. Maon. Resonance 1 9 7 1 , 3., 783. 110. Kalinowski, H.-O.; Berger, S.; Braun, S. " 1,C-NMR- Spektoskopie", Geog Thieme Verlag, Stuttgart, 1984. 111. Hine, J. "Structural Effects on Equilibria in Organic Chemistry", Robert E. Krieger Publishing Company, Huntington, N.Y., 1981. 112. Perkins, R.R.; Pincock, R.E. Tetrahedron Lett. 197S, 943. 113. Keck, G.E.; Yates, J.B. J. Am. Chem. Soc. 1982, 104. 5829. 114. Della, E.W.; Patney, H.K. Aust. J. Chem. 1979, 32, 2243. 115. McMurry, J.E. Acc. Chem. Res. 1983, 16., 405. 116. Experiment performed by Dr. Tomoshige Kobayashi.