Citation
Heteroatom mediated modification of reactivity in carbon skeletons

Material Information

Title:
Heteroatom mediated modification of reactivity in carbon skeletons
Creator:
Sengupta, Saumitra, 1957-
Publisher:
[s.n.]
Publication Date:
Language:
English
Physical Description:
ix, 149 leaves : ill. ; 28 cm.

Subjects

Subjects / Keywords:
Alcohols ( jstor )
Amines ( jstor )
Carbamates ( jstor )
Carbon ( jstor )
Carbonates ( jstor )
Ethers ( jstor )
Lithium ( jstor )
Protons ( jstor )
Pyridines ( jstor )
Tetrahedrons ( jstor )
Alcohols ( lcsh )
Amines ( lcsh )
Carbanions -- Reactivity ( lcsh )
Chemistry thesis Ph. D
Dissertations, Academic -- Chemistry -- UF
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1987.
Bibliography:
Bibliography: leaves 136-148.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Saumitra Sengupta.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
022774810 ( ALEPH )
18763059 ( OCLC )

Downloads

This item has the following downloads:


Full Text









HETEROATOM MEDIATED MODIFICATION OF REACTIVITY
IN CARBON SKELETONS



BY

SAUMITRA SENGUPTA


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY




UNIVERSITY OF FLORIDA


1987





























To my Father

so close, yet so far

















ACKNOWLEDGEMENTS


It was a memorable experience to work under the

supervision of Prof. Alan R. Katritzky. I am most grateful for his thoughtful suggestions and untiring persuasion which ultimately led to fruitful results. I thank him, especially, for believing in me and for the independence I received during the later phases of my research.

I was rather fortunate to carry the blessings of Prof. Merle A. Battiste, a person whom I could approach both as a friend and as a teacher and from whom I learned most of my chemistry. Prof. W. M. Jones, Prof. W. R. Dolbier and Prof. J. F. Helling also extended their helping hands whenever needed and I thank them for their support. The generosity of Prof. K. S. Schanze and Prof. J. A. Deyrup is deeply appreciated as well.

I am very much indebted to Dr. Steve Cato who took great pains in reading through the entire manuscript and correcting my "Indian English." Dr. Ramiah Murugan and Sutha deserve special recognition for providing confidence, easing many of my anxieties and for making me feel at home. Jamshed Lam gave me much needed company both in and out of the lab











and I thank him for all his help. I also thank all my fellow group members and our secretary Ms. Dawn Sullivan for being on my side all these years and especially Dr. Rick Offerman and Mike Walker for creating such a wonderful working atmosphere in CRB 320. As a friend in need, Dr. Radi Awartani went out of his way to help me on numerous occasions and I am very much thankful to him.

Without the wholehearted support of my mother and my sister I wouldn't have been here in the first place. They gave me all that one could ask for. Words cannot describe the sacrifice they undertook and that, too, during a very difficult phase of their lives. I am grateful, yet proud of them and only wish that may I be born in such a loving family again.














TABLE OF CONTENTS


page

ACKNOWLEDGEMENTS ...................................... iii

ABSTRACT .............................................. v

CHAPTERS

I. GENERAL INTRODUCTION ............................. 1

1.1. Modification of Reactivity by Heteroatoms ... 1 1.2. Amines . ...................................... 2
1.3. Alcohols .................................... 4
1.4. Vinylic Amines .............................. 5
1.5. Preview of the Work ......................... 8

II. a-AMINOCARBANIONS ................................ 9

2.1. Introduction .................................... 9
2.1.1. Indirect Routes to cc-Aminocarbanions 10
2.1.2. Protection-Activation Route to
a-Aminocarbanions .................... 11
2.1.3. Carbon Dioxide as a
Protecting-Activating Group ............ 16
2.1.4. Aims of the Work ..................... 23
2.2. Results and Discussion .......... ........ 24
2.2.1. o-Metallation Studies on Lithium
1,2,3,4-Tetrahydroquinolinyl Carbamte. 24
2.2.2. a-Metallation Studies on Lithium
Indolinyl Carbamate ..................... 32
2.2.3. Role of Carbon Dioxide in the
c-Metallation of Lithium Carbamates 35
2.3. Conclusions ................................. 37
2.4. Experimental ................................ 38

III. a-HYDROXYCARBANIONS .............................. 45

3.1. Introduction ................................ 45
3.1.1. Indirect Routes to
a-Hydroxycarbanions ..................... 46
3.1.2. Protection-Activation Route to
a-Hydroxycarbanions ..................... 47
3.1.3. Carbon Dioxide as a
Protecting-Activating Group ............ 53
3.1.4. Aims of the Work ..................... 57











3.2. Results and Discussion ...................... 59
3.2.1. =-Lithiation Studies on Lithium
Methyl Carbonate ..................... 59
3.2.2. a-Lithiation Studies on Lithium
Trimethylsilylmethyl Carbonate ....... 60
3.2.3. x-Lithiation Studies on Lithium
Cinnamyl Carbonate ...................... 68
3.3. Conclusions ................................. 69
3.4. Experimental ................................ 70

IV. THERMAL BEHAVIOR OF N-VINYL-I,2-DIHYDROPYRIDINES . 76

4.1. Introduction ................................ 76
4.1.1. 1,2-Dihydropyridines .................... 76
4.1.2. Synthetic Routes to
1,2-Dihydropyridines .................... 77
4.1.2.1. Reaction of Nucleophiles
with Pyridines ................. 77
4.1.2.2. Reaction of Nucleophiles
with Pyridinium Salts ......... 79 4.1.3. Reactivity of 1,2-Dihydropyridines ... 80
4.1.3.1. Cycloaddition Reactions ..... 81 4.1.3.2. Electrocyclic Reactions ..... 82
4.1.4. Effects of N-Substituents on the
Thermal Behavior of
1,2-Dihydropyridnes ..................... 84
4.1.4.1. N-Alkyl and N-Aryl
Substituents ................... 84
4.1.4.2. N-Vinyl Substituents .......... 85
4.1.5. Aims of the work ..................... 87
4.2. Results and Discussion ...................... 89
4.2.1. Synthetic Strategy ...................... 89
4.2.2. Preparation of
N-(Cycloalken-3-on-l-yl)pyridinium
Salts . ............................... 90
4.2.3. Preparation of
N-(Cycloalken-3-on-l-yl)-1,2dihydropyridines ..................... 92
4.2.4. Pyrolysis of
N-(Cycloalken-3-on-l-yl)-4-phenyl1,2-dihydropyridine ..................... 95
4.2.5. Preparation of N-(Cycloalken-l-yl)pyridinium Salts ..................... 97
4.2.6. Preparation of N-(Cycloalken-l-yl)1,2-dihydropyridines ................... 101
4.2.7. Pyrolysis of N-(Cycloalken-l-yl)4-phenyl-l,2-dihydropyridine .......... 109











4.3. Conclusions ................................. 109
4.4. Experimental ................................ Ill

V. THERMAL BEHAVIOR OF 1,2-DIHYDROPYRIDINES WITH AN
UNSATURATED 4-SUBSTITUENT ........................ 120

5.1. Introduction . ................................ 120
5.1.1. Aim of the Work ...................... 121
5.2. Results and Discussion ...................... 123
5.2.1. Synthetic Strategy ..................... 123
5.2.2. Preparation of 4-Ethoxycarbonylaminopyridinium Salt ...................... 125
5.2.3. Attempted Sodium Borohydride
Reduction of 4-Ethoxycarbonylaminopyridinium Salt ...................... 125
5.2.4. Preparation of
4-Methoxycarbonylmethyl-l-methylpyridinium Salt ...................... 126
5.2.5. Sodium Borohydride Reduction of
4-Methoxycarbonylmethyl-l-methylpyridinium Salt ...................... 126
5.2.6. Preparation of l-Methoxycarbonyl-4methoxycarbonylmethyl1,2-dihydropyridine .................... 128
5.2.7. Thermolysis of 1-Methoxycarbonyl-4methoxycarbonylmethyl1,2-dihydropyridine .................... 129
5.3. Conclusions ................................. 130
5.4. Experimental ................................ 130

VI. SUMMARY .......................................... 133

BIBLIOGRAPHY .......................................... 136

BIOGRAPHICAL SKETCH ................................... 149


vii












Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

HETEROATOM MEDIATED MODIFICATION OF REACTIVITY IN CARBON SKELETONS


BY


SAUMITRA SENGUPTA

DECEMBER, 1987


Chairman: Prof. Alan R. Katritzky Major Department: Chemistry


The generation of c-aminocarbanions from secondary

amines is a matter of long-standing interest. The usual way to generate such species involves a three-step sequence: protection of the amino proton, a-lithiation and electrophilic substitution followed by deprotection. As a superior alternative, carbon dioxide was investigated as a transient protecting group, which enabled electrophilic substitution at the a-position of tetrahydroquinoline and indoline in a one-pot sequence. However, the results show that the a-activating effect of lithium carbamates is small and requires strongly basic reagents for a-deprotonation.

Carbon dioxide has also been investigated as a

protecting group for generating a-hydroxy carbanions. The a-activating power of lithium carbonates was found to be


viii











indeed very small, even in cinnamyl system. In view of this, a trimethylsilyl group has been used as an easily removable a-activating auxilliary. Thus, 1-trimethylsilylmethanol via its lithium carbonate was successfully exploited to give a novel methanol dianion synthon in a one-pot sequence.

There are a few examples of thermal electrocyclic

ring-opening of 1,2-dihydropyridines to the 1-azatrienes, a process considered uphill by thermodynamic considerations. It was of interest whether such ring-opening would be favored in N-vinyl-1,2-dihydropyridines. In the latter case, it was argued that the incipient 3-azatetraene may be induced into further non-degenarate reactions, thus driving the ring-opening equilibrium toward the right. Some novel N-cycloalkenyl-l,2-dihydropyridines were synthesized for this purpose and their thermal behavior studied. However, they preferably underwent oxidative fragmentation to the corresponding pyridine bases and no products arising out of their ring-opening was observed. Similar failures were also recorded for a 4-vinyl-l,2-dihydropyridine precursor.

















CHAPTER I

GENERAL INTRODUCTION



1.1. Modification of Reactivity by Heteroatoms



In general, saturated hydrocarbons are very inert

molecules. They show little tendency to take part in most types of chemical reaction. Thus they do not behave as an electrophile nor can they be used as nucleophiles in a reaction. However, introduction of a heteroatom in an all-carbon framework dramatically alters the situation. By including a nitrogen atom (as in amines) or an oxygen atom (as in alcohols) many transformations can now be achieved at the carbon atoms as well as on the heteroatom. This is mainly due to the large electronegativity difference between carbon and the heteroatoms which induces considerable polarization in such carbon-heteroatom bonds.

Unsaturated carbon skeletons like alkenes and alkynes are more reactive than their saturated counterparts. By











virtue of their loosely bound n-electrons they undergo reactions with electrophiles and again, introduction of a heteroatom like nitrogen next to a a-bond greatly enhances their electrophilicity. The effect of the lone-pair on nitrogen is conjugatively transmitted to the n-electrons thus increasing their nucleophilic character which is well documented in enamine chemistry [82T1975, 82T3363].



1.2. Amines


The reactivity of the -carbons in amines is dominated by nucleophilic attack at that center [68MIl] which shows that amines normally prefer a electropositive center next to the nitrogen. However, being more electronegative than carbon, the nitrogen atom in amines polarizes the C-N bonds to induce a partial positive character on the adjacent carbon, thus rendering the a-hydrogens somewhat acidic. This, in turn, opens up the possibility of generating a carbanionic center next to the nitrogen and hence to the prospect of electrophilic attack at that center. This type of reactivity umpolung at the a-carbon of amines was forseen by Seebach [79AG(E)239,77S357] and is most desirable since a-functionalization by nucleophilic attack has many limitations. For example, it would take a multistep sequence to construct a 0-hydroxy amine skeleton which, instead, can be easily prepared in one step via the reaction of an a-amino carbanion with a suitable electrophile (Scheme 1.1).













\/ R\
N-CH2-CH N-CH + RCHO
OH


Scheme 1.1


In practice, generation of a-amino carbanions proved to be a rather difficult task, especially for primary and secondary amines. In these cases the amino groups contain more acidic hydrogens and require protection during the a-deprotonation step. Here the principal problem lies in the choice of a suitable protecting group which should be easily introduced and subsequently removed as well as be inert to the deprotonating agent. However, with time, it became well recognized that the a-protons in amines are not acidic enough to be abstracted, even by strong bases and therefore needed additional activation, namely a-activation. This has been achieved in various ways, most commonly through the choice of an a-activating protecting group [84CRV471). Yet another problem associated with a-amino carbanions is that, once generated, they face strong repulsion from the adjacent lone-pair on the nitrogen and their stability is greatly reduced. As a result, many a-amino carbanions show tendencies to undergo rearrangements whereby the negative charge is tranferrerd to the nitrogen atom, leading to a thermodynamically more stable situation.











Since a-functionalization in amines has been a matter of long-standing interest, research in the generation and utilization of a-amino carbanions continues. Now that generation of a-amino carbanions has been successfully demonstrated, greater stress has been applied to find novel yet simpler routes to these species.



1.3. Alcohols


Since oxygen is more electronegative than carbon, the oxygen atom enjoys the greater share of the bonding electrons in a C-0 bond. Hence the a-carbon atom in an alcohol attains some electropositive character. Being more electronegative than nitrogen, the degree of polarization in hydroxy derivatives is much more than in amines and nucleophilic attack on the a-carbon atom is much more facile in oxygenated compounds [71MI1]. On the other hand, inductive electron-withdrawal by oxygen makes the a-hydrogens acidic so that generation of an a-oxy carbanion is feasible. However, the generation of a-oxy carbanions faces similar problems as encountered with a-amino carbanions. Thus, the more acidic hydroxy proton needs prior protection and although the kinetic acidities of a-hydroxy protons are slightly higher than in amines, the process of a-activation is again necessary. The stability of a-oxy carbanions is even lower and they are very prone to 1,2-anionic migrations, most notably the Wittig rearrangement [70AG(E)763J.











As a demonstration of reactivity umpolung at the

a-carbon of alcohols, a-oxy carbanions has been successfully generated and utilized to construct key segments which otherwise require lengthy procedures. However, in contrast to the amines, reactivity umpolung at the a-carbon of hydroxy compounds has found less prominence since several other synthetic sequences are available which formally implement the concept. Nevertheless, a-oxy carbanions are synthetically important intermediates and are attractive targets for synthetic chemists.



1.4. Vinylic Amines



Pericyclic reactions are one of the most interesting

classes of reactions associated with alkenes and dienes; of these, the thermal electrocyclic reactions of polyenes have created much interest [80MI11 both from the mechanistic and the synthetic viewpoints. Cyclic dienes have attracted considerable attention as substrates for thermal electrocyclic reactions. Suitable 1,3-cyclohexadiene moieties undergo 6n cycloreversion to the corresponding 1,3,5-hexatriene system (Scheme 1.2). Although this transformation is considered an equilibrium, the forward reaction can rarely be used preparatively due to unfavorable energetics [80M121.














R R



Scheme 1.2


The formation of 1,3-cyclohexadiene from

cis-l,3,5-hexatriene has an activation enthalpy of 29.0 Kcal/mole [64JCS3080]. The heats of formation of 1,3-cyclohexadiene and cis-l,3,5-hexatriene are reported as 40.6 and 25.4 Kcal/mole, respectively [69CC833]. These figures establish the AHm for the ring opening of cyclohexadiene at 44.2 Kcal/mole which means that very high temperatures would be necessary to achieve ring-opening of 1,3-cyclohexadiene type systems. Thus, in contrast to numerous instances of electrocyclizations of 1,3,5-hexatrienes, one finds only a few examples of cycloreversion in 1,3-cyclohexadiene systems.

In 1,3-cyclohexadiene, substitution of a methylene group by a heteroatom such as nitrogen transforms it to a 1,2-dihydropyridine which can be looked upon as a dienamine. Both enamines and dienamines are more polarized than their parent alkenes. Although they participate in pericyclic reactions, the reactions are mostly polar in nature and quite often occur in stepwise fashion. Thus it can be











expected that the diene moiety in 1,2-dihydropyridinies would be perturbed by the nitrogen so as to aid in its cycloreversion. Experimental evidence shows this not to be of considerable extent and 1,2-dihydropyridines are much more stable than the 1-azatrienes [80M13]. However, in contrast to 1,3-cyclohexadienes, examples of ring-opening of 1,2-dihydropyridines are more numerous and the process is indeed facile in those cases where the resulting azatrienes enjoy extensive charge delocalization. This has been elegantly shown by Maas [85AG(E)511] (Scheme 1.3).



N NA

HC (CN)2 H NN CH(CN)2
251C, 2 hr
Me 2N + NMe Me2 NNMe2
2 OTf OTf





CN
/ CN
Me N

MeN
Me N
Me2N


Scheme 1.3











The above reaction, together with several examples of the ANRORC (Addition of Nucleophile--Ring Opening--Ring Closure) type reactions of pyridinium salts [80S589) show that, given the proper scenerio, 1,2-dihydropyridines are much easier to ring-open than 1,3-cyclohexadienes.



1.5. Preview of the Work



In Chapter II of this dissertation, a novel methodology for the generation of a-carbanions from secondary amines is described. In that section, stress has been given to the role of the protecting group; also, the concept of transient a-activation via anionic protection has been introduced. The same concept was next applied to the generation of c-hydroxy carbanions (Chapter III) which in the process also saw the utilization of the trimethylsilyl group as an easily removable a-activating auxilliary.

In Chapter IV, thermal electrocyclic reactions of a

cyclic dienamine, a 1,2-dihydropyridine were evaluated. The synthesis and thermal behavior of some novel N-cycloalkenyl-l,2-dihydropyridines have been investigated and their results are presented.

















CHAPTER II

a-AMINOCARBANIONS



2.1. Introduction



In amines, the inductive electron withdrawal by nitrogen generates a small positive charge on the a-carbon, rendering the a-hydrogen(s) acidic. As a result, the generation of a-aminocarbanions becomes feasible and leads to the prospect of electrophilic substitution at the a-carbon of amines. However, the generation of a-aminocarbanions has several associated problems. Primary and secondary amines possess more acidic hydrogen(s) on the heteroatom and, in these cases, generation of a-aminocarbanions would require prior protection of the amino group. Furthermore, the a-aminocarbanions, once generated, experience strong repulsion from the adjacent lone-pairs on nitrogen, thus destabilizing them.











Chemical transformations at the cx-carbon of an amine is a matter of long-standing interest and in this regard,

-aminocarbanions were expected to play a powerful role. Thus in the last decade much research has been devoted to the generation of a-aminocarbanions, as a result of which several x-aminocarbanion synthons are now available. These synthons have been critically assesed and several reviews have appeared on these and related topics [75AG(E)15,77S357, 78CRV275,79AG(E)239,8OT2531,84CRV471,85MIlI.



2.1.1. Indirect Routes to a-Aminocarbanions

Some c-aminocarbanions can be generated as their lithio salts without using an a-deprotonation step in the sequence. These methods are termed here as "indirect methods," which utilize either trans-metallation (usually tin-lithium exchange) or lithium-metalloid exchanges [80T2531]. In the latter class, the name "metalloid" implies elements such as S, Se, and the heavier members of the halogen family.

Tin-lithium exchange has been successfully utilized to generate a-lithio-tertiary amines. Thus the readily-available dialkylaminomethylstannanes (2.1), on treatment with n-buatyllithium, gave rise to the dialkylaminomethyllithium reagents (2.2). The latter upon reaction with different electrophiles produced the expected











a-alkylated amines (2.3) (Scheme 2.1) [86TL2361, 84MI1, 71JA4027]. That this type of transmetallation produces a free organolithium (e.g. 2.2), and not any pentavalent "tin-ate" complex, can be nicely shown by 119Sn-NMR studies [85TL1141,86JA2102].


BuLi E+
R2N-CH2-SnBu3 > R2N-CH 2-Li > R2N-CH2-E


(2.1) (2.2) (2.3)


Scheme 2.1

Although quite rewarding in its own merit, this

transmetallation route suffers from a few drawbacks. The process fails completely when the a-carbon is either secondary or tertiary in nature [84MII]. Similar transmetallations for a-stannyl-primary and -secondary amines are yet to be reported [85TL1141].

Lithium-metalloid exchange may not be a wise choice for generating o-lithioamines. The starting materials (e.g. a-haloamines) needed for this purpose are rather unstable and are often difficult to prepare.



2.1.2. Protection-Activation Route to a-Aminocarbanions

This section deals with the more direct approach toward a-aminocarbanions, formed as their lithio salts. We would focus our attention on the c-lithio secondary amines where this methodology has found considerable success. In the











"protection-activation" route, the c-lithioamines are generated by direct abstraction of an c-hydrogen using a strong lithiating base. In order to do so, the more acidic amino hydrogen needs protection. The protecting groups are so chosen that they are capable of performing two functions simultaneously. Firstly, the more acidic amino proton is protected from reacting with the lithiating base. Secondly, they play a major role in activating the a-position toward hydrogen abstraction. This process of cc-activation is necessary since the c-hydrogens of amines are not acidic enough to be abstracted by strong bases. If "G" be such a hypothetical protecting-activating group, c-alkylation of a secondary amine can be visualized to occur as depicted in Scheme 2.2.

+ G-X R"Li
RN-CH -R' > RN-CE -R' > RN-CH-R'
1 2 + 2- 1
H -H G G'''Li (2.4) (2.5) (2.6)






-G
RN-CH-R' I I +I I H E + H G E (2.8) (2.7)


Scheme 2.2











The key step in the sequence in Scheme 2.2 is the

formation of the c-lithio species (2.6) from 2.5, since it involves the process of a-activation. There are several possibilities by which the protecting group (in 2.5) could activate the c-position. According to Beak and Meyers [86ACR356], prior complexation of lithium (from the lithiating base) with "G" would activate the lithiating agent as well as direct the latter on to the c-carbon via a rigid chelate. In doing so, a partial dipole might also be created resulting in increased kinetic acidities of the c-protons (Figure 2.1).




RN -CH-R' RN+-+ CH-R'
%I I
H~H


Li-R L R"



Figure 2.1. c-Activation of Secondary Amines


In all probability the processes of prior complexation and dipolar activation would work in unison and complement each other. The a-lithio species once formed owes its stability to a great extent on the co-ordinating power of the protecting group so as to (partly) satisfy the co-ordination sites for lithium. For an all round success,












the protecting group itself should not contain any acidic hydrogen, should be inert to the metallating agent and, of course, be easily introduced and removed.

Based on these concepts several protecting-activating groups have been successfully employed in generating a-lithio secondary amines. Some of the versatile a-lithio secondary amine precursors are shown in Figure 2.2. A comprehensive picture can be obtained from Beak's recent review on this subject [84CRV471].



R-N-CH R' R-N-CH R' R-N-CH R'
1 2 1 2 1 2
N=O C=O C=O
I I
C(Et)3 C(Ph)3 (2.9)
(2.10) (2.11)



R-N-CH R' R-N--CH R'
1 2 1 2
P(NMe2)2 HC
0 N-CMe3


(2.12) (2.13)


Figure 2.2. a-Lithio Secondary Amine Precursors


For the N-nitrosoamines (2.9) [75AG(E)15], resonance

stabilization of the a-carbanion is thought to be the chief source of a-activation [78CRV275]. Lithium diisopropylamide (LDA) has been used as the lithiating agent. The remaining examples in Figure 2.2 operate on the "prior complexation-











activation" concept [86ACR356] and stronger metallating agents like alkyllithiums were necessary for their success. However, alkyllithiums are strong nucleophiles as well and the use of such strongly nucleophilic agents warrants that the protected center be sufficiently hindered so as to prevent attack at that position. Thus the sterically crowded triethylacetamides (2.10) [84JA1010], and triphenylacetamides (2.11) [83T1963] have been chosen to protect their carbonyl centers from any nucleophilic attack. By comparison, phosphonamides (2.12) [83T1963] and formamidines (2.13) [85MII] are relatively inert to attacks by organolithiums.

At present, the N-nitrosoamines (2.9) and the

formamidines (2.13) stand out as the most useful precurssors to a-lithio secondary amines. Both such systems encourage a-lithiation on 1 , 2 and 3 carbons and alkylation has been achieved with a wide variety of electrophiles. However, the N-nitrosoamines are notoriously toxic and hazardous. The formamidines, on the other hand, are easy to prepare, nontoxic and can be deprotected (by hydrazinolysis or LiAlH4 reduction) to liberate the a-alkylated secondary amines. In addition, chiral formamidines derived from secondary amines served as substrates for asymmetric a-alkylation and have











been utilized for the synthesis of various alkaloids in high (90-98%) enantiomeric excesses [87JA1263, 87JA1265, 86JOC3108, 86JOC3076]. The c-aminocarbanion precursors 2.102.12 require very harsh deprotection steps and are thus not of much practical use.


2.1.3. Carbon Dioxide as a Protecting-Activating Group

As just described, a-alkylation of secondary amines can be achieved in a three-step sequence: protection, a-lithiation and electrophile addition followed by deprotection (Scheme 2.2); the principal focus being in the .c-lithiation step. In order to accomplish this, special protecting groups have been employed. In some cases (e.g.

2.10 and 2.11), preparation of the protecting group itself involved multistep synthesis while in others, deprotection needed extreme conditions (e.g. 2.11 requires dissolving metal reduction) thereby limiting their synthetic utility.

An additional problem encountered in the generation of a-aminocarbanions is the 1,2-migration of the N-protecting group. This type of anionic migration is extremely facile since it involves an entropically favored three-membered transition state. The equilibrium in Scheme 2.3 demonstrates such a rearrangement [76CC339].













Ph

IN
Ph '>'C=Ot
1~


Ph
C-N -R "/I Ph C=O
OEt


Scheme 2.3

Another example of such 1,2-anionic migration is found in the case of N-nitroso(2-methylamino)acetonitrile (2.14) (Scheme 2.4). In basic methanol, 2.14 gives rise to the a-carbanion (2.15) which triggers off a facile nitroso-transfer, ultimately leading to the amidoxime (2.17) [62HCA2426, 64HCA33, 78JOC3617]. Similar nitroso-transfer has also been observed during the o-lithiation of N-(t-butyl)-N-(nitroso)trimethylsilylmethylamine [77CB1852].


base
NC-CH2-N-Me > NC-CH-N-Me N=O MeOH -N=O (2.14) (2.15)



1


NC-C-NH-Me
II
N-OH
(2.17)


NC-CH-N-Me
I
N=O
(2.16)


Scheme 2.4











This type of migration of the N-protecting group is aided by the fact that the nitrogen, in most cases, is attached to an electrophilic center (either a carbonyl for the amides or a nitroso for the N-nitrosoamines). On the other hand, such electrophilic centers on nitrogen is essential both for a-activation and easy deprotection. This requirement withholding, a decrease in the electrophilicity of the N-bound center appeared to be a viable solution to the 1,2-migration problem. An anionically protected amine, e.g. (2.18) was thus considered as the remedy (Figure 2.3).

R N-C 0

I XLi+
CH2


(2.18)


Figure 2.3. Anionically Protected Secondary Amine It was envisaged that the negative charge at the

protected end of 2.18 would decrease the electrophilicity of the carbonyl group and hence, the N -> C migration could be minimized. Furthermore, the N-bound center in 2.18 remains electronically protected from nucleophilic attack by the lithiating agent. The process of a-activation could still be operative through prior complexation of the lithiating agent











either to "X" or to the carbonyl oxygen. The biggest advantage of using 2.18 lies in the prospect of its facile deprotection. This should occur readily since the protecting group would be lost as a neutral molecule (O=C=X).

The proposed synthon (2.18) could possibly be made via the reaction of an N-lithio amine with a suitable cumulene of general stucture O=C=X (Scheme 2.5). In choosing the appropriate cumulene, preference should be given to those in which X is a heteroatom so that the negative charge in 2.18 is better accomodated. Also, a heteroatom at that position may be neccessary for the process of a-activation through prior complexation with the lithiating agent.


R
N-Li 0 R 0
+ C >N-C
II I \ _CH2 XI X Li
Rf "ICH2 R'

(2.18)


Scheme 2.5


One such cumulene which meets all these requirements is carbon dioxide. It is readily available, relatively non-toxic and has the proven ability as an electrophile. A thorough search of the literature revealed that various N-lithio derivatives of secondary amines react with CO2 to











form the corresponding lithium carbamates [87H1333]. These salts, although stable in the absence of moisture, readily loose CO2 on mild hydrolysis or simple heating [87H1333]. This latter property, we thought, could be efficiently exploited to achieve a very mild deprotection step. Given the hope that a-lithiation of lithium carbamates could be achieved, a novel synthon for c-aminocarbanions was in the offering.

In a preliminary study carried out by previous workers in this group, CO2 has shown considerable promise as a protecting-activating group for a-lithiation of a primary amine. Benzylamine (2.19) was investigated as a model compound which formed its lithium carbamate (2.20) on being treated with n-butyllithium (1 equivalent) followed by CO2. Treatment of 2.20 with two equivalents of t-butyllithium gave rise to the a-lithio species (2.21) which was subsequently trapped with various electrophiles (D20, CO2, MeI, PhCO2Et, Bu tNCO, PhNCO) providing the c-substituted carbamates (2.22) (Scheme 2.6). The latter, without the need of isolation, was readily deprotected under acidic work-up to give the a-substituted benzylamines (2.23) in good yields [87S415]. As an added advantage, the whole set of operations can be executed in the same reaction vessel.













1) BunLi 2 ButLi
PhCH2NH2 > PhCH2NHCO2Li > PhCH-N=C OLi
2) CO2 THF/- 200C Li
(2.19) (2.20) (2.21)




H+/H20

PhCH(E)NH < PhCH-N=COLi 21
-CO2 E
2 OLi
(2.23) (2.22)


Scheme 2.6


The directing power of the lithium carbamate (2.20) can be well appreciated if one considers that mono- and disubstituted benzylamines undergo exclusive ortho-lithiation in the phenyl ring [71JOC1607,63JOC663, 63JOC3461,67JOC1479J (Scheme 2.7). For ortho-lithiation to occur in 2.20, a seven-membered chelate would be neccesary; instead a more rigid five-membered chelate in 2.21 directs the lithiation at the benzylic carbon.


2 BuLi
CH -NHR >CH NR
2 2

[a L Li


Scheme 2.7











Previous workers have also studied the lithium carbamate route for a-lithiation of a secondary amine, namely tetrahydroisoquinoline (2.24). Most gratifyingly, the derived lithium carbamate (2.25) underwent facile lithiation at the 1-position to produce the a-lithio species (2.26). The latter upon reaction with different electrophiles followed by acidic work-up directly afforded the 1-substituted tetrahydroisoquinolines (2.27) (Scheme 2.8) [86T2571]. Again, the whole sequence can easily be carried out in the same reaction vessel.



NH 1) i> N-CO 2Li 2) CO2
(2.24) (2.25)



ButLi THF/-200C





1) E'
NH < NN-CO 2Li
E 2) H /H20 Li

(2.27) (2.26)


Scheme 2.8











With these preliminary results in hand, a detailed study of this novel a-lithiation process was most desirable.



2.1.4. Aims of the Work

The aim of the present investigation was to explore various other lithium carbamates as c-aminocarbanion synthons, in particular, those derived from unactivated secondary amines. a-Lithiation of benzylamine and tetrahydroisoquinoline via their lithium carbamates can be considered as examples where the respective a-hydrogens are fairly activated, being benzylic in nature. Therefore, the role of CO2 in the c-activation step could not be critically assessed. Hence it was of interest to study the a-lithiation of those lithium carbamates where any kind of inductive activation of the a-hydrogens is absent. The secondary amines chosen for the study were 1,2,3,4-tetrahydroquinoline and indoline.

a-Lithiation of tetrahydroquinoline and indoline has previously been achieved only via their formamidines (81TL5119]. In that study, the respective a-lithio species were trapped with alkyl halides (MeI, Bul) and benzaldehyde to give the a-alkylated formamidines in good yields. However, deprotection of the latter required rather harsh











conditions such as refluxing with aqueous KOH or hydrazine, especially for the benzaldehyde adducts which could only be deprotected by LiAlH4 in refluxing THF. In view of the mild deprotection procedure for the lithium carbamates, succesful a-lithiation of the tetrahydroquinoline and indoline via their lithium carbamates appeared as a potentially superior method for the a-functionalization of these amines.



2.2. Results and Discussion


2.2.1. c-Metallation Studies on Lithium
1,2,3,4-Tetrahydroquinolinyl Carbamate

1,2,3,4-Tetrahydroquinoline (2.28) was easily converted to its lithium carbamate (2.29) by sequential addition of n-butyllithium followed by CO2 gas (Scheme 2.9). In order to achieve the desired a-lithiation, 2.29 was exposed to t-butyllithium in tetrahydrofuran. After 2 hrs at -200C, p-tolualdehyde was added as the electrophile. However, after acidic work-up with 2N HCl, formation of the desired a-adduct 2.31 was not observed; the starting materials were recovered unchanged. Thus it was evident that the a-lithio species 2.30 was not generated under these conditions.













1) BUnLi I
N
II
H 2) CO2 C 2Li
(2.28)
(2.29)


Bu tLi



Me


P / 1 -MeC 6H 4CHO
N- -- - N Li
I \I

H OH 2) H+/H20
/H20 C02Li (2.31) (2.30)


Scheme 2.9


In order to achieve the desired a-lithiation, 2.29 was

subjected to various other reaction conditions. It was found that solvent had no effect on the lithiation process. Thus the reaction failed when carried out in ether as well as in hexane (heterogeneous system). Complexing agents [86TL331] like hexamethylphosphoramide (HMPA) or tetramethylethylenediamine (TMEDA) also had no effect on the reaction and in each case, the starting materials were recovered unchanged. These failures, however, shed some light on the role of CO2 in the a-lithiation step: the lithium carbamate 2.29 had no perceptible activation toward











cc-lithiation. At this stage it was felt that the source of a-activation for lithium carbamates lies somewhere else, probably in the use of a stronger metallating agent. This received some support from the recent studies by Schlosser [84TL741] on the superior metallating power of alkyllithiums in an admix with potassium alkoxides. More pertinent to the present study was the report by Ahlbrecht [84TL13531 who described the successful a-metallation of N-methyl tertiary amines using a mixture of potassium t-butoxide and s-butyllithium, otherwise a poor yielding process with alkyllithiums alone. On the basis of these reports, we proceeded to reinvestigate the e-metallation of the lithium carbamate 2.29 using Schlosser's mixed base system.

The lithium carbamate 2.29 was generated as described

before in Scheme 2.9. When 2.29 was treated at -78*C with an equimolar mixture of t-butyllithium and potassium t-butoxide in tetrahydrofuran, a wine-red solution was immediately formed. Addition of benzaldehyde discharged this color and after stirring overnight at room temperature, the reaction was worked-up with 2N HCI. This time the crude product mixture afforded the desired c-adduct 2.33 in 55% yield as a mixture of diastereomers (Scheme 2.10).















ButLi/KOBut I CO2Li THF, - 500C CO2Li
(2.29) (2.32)



) PhCHO
2) 2N HCl






aN Ph
H OH


(2.33) (55%)
(erythro:threo = 40:60)


Scheme 2.10

Optimization of the reaction conditions gave the best results when t-butyllithium was used as the alkyllithium component and the metallation being carried out at -500C for 2 hrs. The reaction worked best in tetrahydrofuran since the lithium carbamate 2.29 was found to be insoluble in ether solvents such as hexane or ether.

The diastereomers of 2.33 were separated by careful

column chromatography (silica gel) to give a 40:60 ratio of











erythro:threo isomers and were characterized individually by their 1H and 13C-NMR spectra. Their stereochemical assignments were based, tentatively, on the nature of the a-hydroxy proton in the IH-NMR spectra (200 MHz) (Fig. 2.4) and were in agreement with the trend which is observed for similar c-adducts of tetrahydroquinoline [81TL5119] and piperidine [79TL771, 85JOC1019]. According to these reports the a-hydroxy proton of the erythro-adducts usually appears at a more downfield region and with a lower coupling constant than that of the threo-isomers. Thus, in our hands, the product-isomer whose a-hydroxy proton appeared at 6 4.97 (d, J 6 Hz) was assigned to erythro-2.33; the corresponding proton for the other isomer was found at 8 4.97 (d, J 10 Hz) 13 M
and was assigned to threo-2.33. In their - C-NMR spectra the a-hydroxy carbons were found at 8 75.1 and 76.0 wheras the a-amino carbons appeared at 8 41.6 and 43.0, respectively, for threo- and erythro-2.33. It should, however, be noted that both threo-2.33 (m.pt. 147�-149�C, from benzene pet.ether) and erythro-2.33 (m.pt. 105'-107'C, from ether pet.ether) were isolated as crystalline solids. This is in contrast to the only previous report [81TL5119] of their preparation where they were described as the hydrochloride salt (m.pt. 138*-141*C, from ether - ethanol) of the diastereomeric mixture.








29













A



I

























, I I I i 7 -7i Iw1
5 6 4 2




1
Figure 2.4. H-NMR Spectra of the threo- and erythro-2.33
A = threo; B = erythro











The c-metallo species 2.32, generated as in Scheme 2.10, reacted with other carbonyl electrophiles such as cyclohexanone and benzophenone to give the corresponding adducts 2.34 and 2.35 in 52 and 55% yield, respectively (Scheme 2.11). These novel adducts were each characterized by their IH and 13C-NMR spectra as well as elemental (C,H,N) analyses.





N M
I
CO2Li


(2.32)



1) Ph2CO 2) 2N HCl
S2 ) 2N HCI




OH Ph NPh H OH (2.34) (2.35)


Scheme 2.11











Our attempts at the a-functionalization of

tetrahydroquinoline via 2.32 had its share of failures also. Of particular significance are the failures when diphenyl diselenide or dimethylformamide were used as electrophiles. These electrophiles are known to be highly lithiophilic and the failure in their reaction with 2.32 speaks in favor of formation of an a-potassio species, rather than an x-lithio species. Reaction of 2.32 with benzoyl chloride, methyl chloroformate or ethyl p-toluate produced complex product mixtures from which only a small amount of the desired products could be detected in their 1H-NMR spectra. Interestingly, when diphenylcarbamoyl chloride was used as the electrophile, it produced 2.36 (35%), the product of N-carbamoylation. This latter result indicates partial cleavage of the N-protecting group in 2.29 during the a-metallation step which might account for the above failures as well as for the modest yields of the products

2.33-2.35 obtained so far.






NN


0 N(Ph)2
2.36











Similar results were obtained when methyl iodide was

used as the electrophile. Although in the 1H-NMR spectrum of the crude product the desired 2-methyltetrahydroquinoline could be observed, it was inseparable from the starting material due to very close Rf values. What could be separated by chromatography was shown to be N-methyltetrahydroquinoline (22%), the N-alkylation product.



2.2.2.ox-Metallation Studies on Lithium Indolinyl Carbamate

Similar to the case of tetrahydroquinoline, the lithium carbamate (2.38) of indoline was prepared by sequential addition of n-butyllithium to the amine followed by CO2 gas. As was expected, 2.38 did not undergo a-lithiation with t-butyllithium alone under a variety of reaction conditions. However, when 2.38 was treated at -781C with an equimolar mixture of t-butyllithium and potassium t-butoxide in tetrahydrofuran, formation of the characteristic wine-red solution of the =-metallo species 2.39 was observed. The color was discharged upon addition of p-tolualdehyde and acidic work-up (with 2N HCI) after 12 hrs gave the a-adduct

2.40 in 52% yield (Scheme 2.12). The incipient ot-metallo species 2.39 could also be trapped with cyclohexanone to give the adduct 2.41 in 57% yield (Scheme 2.12).














) BunLiN ButLi/KOBut


H 2) CO2 CO2Li THF, -500C
(2.37)
(2.38)



-IR 1) RCHOI
N OH < N M
I I
H 2) 2N HC1 CO2Li
(2.40) (52%) (2.39) (erythro:threo = 55:45)



2) 2N HCI
R = p-Me-C6 H4

00H
N
H
(2.41) (57%)


Scheme 2.12


The a-adduct 2.40 was obtained as a mixture of

diastereomers. They were separated by careful column chromatography over silica-gel to give a 55:45 ratio of erythro:threo isomers. The stereochemical assignments were again based on the chemical shift and coupling constant of











their a-hydroxy proton in the IH-NMR spectra. In accordance to the literature data [81TL5119], the isomer whose a-hydroxy proton appeared at 8 4.85 (d, J 6 Hz) was assigned to erythro-2.40. Accordingly, the c-hydroxy proton for threo-2.40 appeared at a relatively upfield position, 6 4.6 and with a slightly larger coupling constant (J 8 Hz). In their 13C-NMR spectra, both the isomers showed the requisite number of carbons, the a-hydroxy carbons being found at

6 75.2 and 75.6, whereas the a-amino carbons appeared at

8 48.6 and 49.3, respectively, for threo- and erythro-2.40. Again it should be mentioned that the benzaldehyde a-adduct of indoline was previously reported [81TL5119] only as the hydrochloride salt of the diastereomeric mixture.

The novel cyclohexanone c-adduct 2.41 (m.pt. 1081-1101C, from ether - hexane) was obtained in 57% yield after column chromatography of the crude product mixture. It was characterized by its 1H and 13C-NMR spectra as well as by elemental (C,H,N) analysis.

Just like its tetrahydroquinoline counterpart, the a-metallo species 2.39 was also found to be inert to diphenyl diselenide and dimethylformamide, whereas reactions with ethyl p-toluate and methyl chloroformate gave complex product mixtures.











2.2.3. Role of Carbon Dioxide in the a-Metallation of
Lithium Carbamates

In the previous two sections, a novel one-pot sequence for m-metallation of tetrahydroquinoline and indoline via their lithium carbamates was described. However, the principal question which remains unanswered is that concerning the role of lithium carbamates toward a-activation. In this section, we attempt to clarify this through some experimental observations.

We begin with Ahlbrecht's results [84TL1353] where the N-methyl groups of some tertiary amines were successfully metallated with Schlosser's mixed base system. In these cases, the substrate did not posses any a-activating group on the nitrogen. However, it should be noted that the proton in question was a primary hydrogen which would be expected to be more acidic than those encountered in our study, i.e. secondary hydrogens. Unfortunately, Ahlbrecht's report [84TL1353] provides no information on attempted a-metallation of secondary carbons in tertiary amines.

Unable to find a direct comparison, we set out to study the a-metallation of protected secondary amines where the protecting groups do not provide any kind of a-activation whatsoever. The substrate chosen for this purpose was the N-trimethylsilylindoline (2.42). The latter was prepared in situ from indoline (n-butyllithium - chlorotrimethylsilane)











and treated first with a mixture of t-butyllithium and potassium t-butoxide followed by benzaldehyde as the electrophile (Scheme 2.13). However, after acidic work-up (with 2N HCl), none of the desired a-adduct (2.44) could be detected. Indoline and benzaldehyde were recovered unchanged.




CnN 1~) Bu Li>N
N N
H 2) Me3SiCl SiMe3
(2.37) (2.42)




Bu tLi KOBut






Ph 1) PhCHO jM


H OH 2) 2N HCI SiMe
3

(2.44) (2.43)


Scheme 2.13


The above result, indirectly shows that lithium

carbamates indeed have an a-activating effect. We believe that dipole creation (as in 2.45) is the major factor behind











such a-activation resulting in increased acidity of the a-protons. However, the extent to which this happens must be small since it required a very strong base like Schlosser's mixed base system for proton abstraction.




/ (CH 2)n


N H

LIo to n - 1,2
LiO 0


2.45




2.3. Conclusions


Tetrahydroquinoline and indoline, via their lithium

carbamates underwent successful c-metallation when treated with a mixture of t-butyllithium and potassium t-butoxide. The resulting a-metallo species were trapped with aldehydes and ketones to give the respective a-alkylated products in good yields, although alkyl halides and acyl chlorides failed in this respect. The nature of "M" in the a-metallo species 2.32 and 2.39 could not be ascertained at this stage. However, on the basis of Schlosser's [84TL7411 and Phlbrecht's [84TL13531 accounts, it was tentatively assigned











as potassium, although the results of attempted transmetallation to the a-lithio species remain inconclusive; thus treatment of 2.32 with lithium, bromide, prior to the addition of benzaldehyde, led to much poorer yields of -alkylation.

The advantages of the present methodology include a very mild deprotection step and the convenience of a one-pot operation. These score over the only reported procedure [81TL5119] for c-alkylation of tetrahydroquinoline and indoline which involved a conventional three-step sequence as well as very harsh deprotection conditions.



2.4. Experimental



Melting points were determined on a Bristoline hot-stage microscope and are uncorrected. 1H and 13C-NMR spectra were recorded on a Varian XL200 spectrometer. Chemical shifts for the protons are reported in 6 (ppm) downfield from Me4Si whereas those for the carbons were determined with reference to the solvent peaks (CDCI3, 6 77.0 or DMSO-d6, 6 39.5). The nature of the carbon signals in the 13C-NMR spectra were determined by the Attached Proton Test [81CC150, 82JMR5351. Mass spectra were recorded on a AEI MS instrument attached to a DS 55 database. Elemental analyses (C,H,N) were performed by Dr. Roy W. King of this department. All











manipulations involving organolithium compounds were carried out in oven-dried (1200C, overnight) apparatus under a slight positive pressure of dry argon. Transferrring operations were done either by syringe techniques or via canula. Tetrahydrofuran (THF) and ether were freshly distilled from sodium-benzophenone ketyl. Chlorotrimethylsilane, hexamethylphosphoramide (HMPA) and tetramethylethylenediamine (TMEDA) were distilled from calcium hydride under reduced pressure and stored over molecular sieves. Petroleum ether (360-60*C), hexane and triethylamine were distilled from calcium hydride and stored over potassium hydroxide. Column chromatography was carried out on MCB silica gel (230-400 mesh) and the solvents used for chromatography were routinely distilled prior to use.


General Procedure for a-Alkylation of Indoline and Tetrahydroquinoline

n-Butyllithium (2.5 M in hexanes, 2 ml, 5 mmole) was

added dropwise at -780C, under argon, to a solution of the amine (5 mmole) in dry THF (35 ml) in a 250 ml Schlenk type reactor. It was then slowly warmed to room temperature and dry CO2 gas rapidly bubbled through it. The solution turned colorless (a colorless solution at this stage is vital for ultimate success; otherwise, poor yields were obtained). All volatile contents were then removed under 0.5 mm leaving a white powdery residue of the lithium carbamates. A preheated











(1400C, 0.5 mm, 2 hrs) sample of potassium t-butoxide (0.57 g, 5 mmole) was then quickly added, under argon, to the reactor. After redissolving in dry THF (40 ml), t-butyllithium (1.7 M in pentane, 3 ml, 5 mmole) was slowly added at -78�C. The resulting wine-red solution was stirred for 2 hrs at -500C and the appropriate electrophile was then added at -780C. Stirring was continued at room temperature for 12 hrs; the solvent was then removed at 0.5 mm and the residue treated with hydrochloric acid (2N, 10 ml). After gas evolution ceased, it was basified with solid sodium carbonate and was extracted with ethyl acetate or methylene chloride (3 x 20 ml). The combined organic layer was washed with brine and dried (Na2So4). Removal of solvent followed by column chromatography (silica-gel, 2-15% gradient of ethyl acetate in hexane containing 10-15% triethylamine) gave the products which were characterized according to the individual cases described below:


l-Phenyl-l-(l,2,3,4-tetrahydroquinolin-2-yl)carbinol (2.33).

Yield 55%; erythro:threo = 40:60.

threo-(2.33): Rf 0.5 (silica-gel, 10% ethyl acetate in hexane containing 10% triethylamine); m.p. 147-149*C (needles from benzene-pet.ether); Found: C, 80.52, H, 7.02, N, 5.80. C6H17 NO requires C, 80.30, H, 7.16, N, 5.85%. 8H











(CDC13/DMSO-d6/D20) 1.6(2H, m), 3.1 (3H, m), 4.65 (1H, d, J 10 Hz), 6.4-6.6 (2H, m), 7.0 (1H, m) and 7.2-7.5 (6H, m). 6C 22.0 (CH 2), 36.6 (CH2), 41.6 (CH-N), 75.1 (CH-OH), 112.4 (aryl CH), 113.7 (aryl CH), 119.6 (aryl C), 125.6 (aryl CH), 125.7 (aryl CH), 125.8 (aryl CH), 126.7 (aryl CH), 130.1 (aryl CH), 142.8 (aryl C) and 143.6 (aryl C).



erythro-(2.33): Rf 0.4 (silica-gel, 10% ethyl acetate in hexane containing 10% triethylamine); m.p. 105-107*C (needles from ether-pet.ether); Found: C, 80.30, H, 7.31, N,

5.55. C16H17 requires C, 80.30, H, 7.16, N, 5.85%. 6H (CDC13/D20) 1.65 (1H, m), 2.2 (1H, m), 3.0 (1H, m), 3.2 (1H, m), 3.4 (1H, m), 4.97 (1H, d, J 6 Hz), 6.4-6.7 (3H, m), 6.9 (1H, m) and 7.2-7.3 (5H, m). 8C 21.7 (CH2), 39.2 (CH2), 43.0 (CH-N), 76.0 (CH-OH), 114.3 (aryl CH), 116.7 (aryl CH), 120.4 (aryl C ), 126.3 (aryl CH), 127.1 (aryl CH), 127.4 (aryl CH), 128.1 (aryl CH), 129.2 (aryl CH), 142.5 (aryl C) and 145.6 (aryl C).



1-(1,2,3,4-Tetrahydroquinolin-2-yl)cyclohexanol (2.34).

Rf 0.5 (silica-gel, 12% ethyl acetate in hexane

containing 10% triethylamine); yield 52%; viscous oil; M+ (HRMS) found: 231.1632. C15H21 requires 231.1624. 8H (CDC13/D20) 1.4-1.7 (11H, m), 2.2-2.3 (1H, m), 2.7 (1H, m), 3.3 (1H, m), 3.55 (1H, m), 6.4-6.6 (2H, m) and 6.9-7.0 (2H, M). 6C 22.0 (CH2), 22.2 (CH 2 ), 25.8 (CH2), 34.1 (CH2), 37.6











(CH2), 39.2 (CH2), 44.5 (CH-N), 73.8 (C-OH), 113.9 (aryl CH), 115.6 (aryl CH), 119.8 (aryl C), 127.6 (aryl CH), 131.5 (aryl CH) and 145.3 (aryl C).



1,1-Diphenyl-(1,2,3,4-tetrahydroquinolin-2-yl)carbinol

(2.35).

Rf 0.5 (silica-gel, 10% ethyl acetate in hexane

containing 10% triethylamine); m.p. 128-1291C (plates from ether-hexane); Found: C, 83.81, H, 6.97, N, 4.09. C22H21NO requires C, 83.78, H, 6.71, N, 4.44%. 8H (CDC13/D20) 1.8 (1H, m), 2.1 (1H, m), 3.0 (1H, m), 3.4 (1H, m), 4.05 (1H, m), 6.2-6.55 (3H, m), 6.95 (1H, m) and 7.1-7.8 (10H, m). &C 24.1 (CH2), 39.6 (CH2), 43.9 (CH-N), 81.5 (C-OH), 114.4 (aryl CH), 116.4 (aryl CH), 118.9 (aryl C), 125.8 (aryl CH), 126.1 (aryl CH), 126.2 (aryl CH), 127.8 (aryl CH), 128.0 (aryl CH), 130.8 (aryl CH), 145.3 (aryl C), 146.3 (aryl C) and 147.8 (aryl C).



1-(Diphenylcarbamoyl)-1,2,3,4-tetrahydroquinoline (2.36)

Rf 0.6 (silica-gel, 10% ethyl acetate in hexane

containing 10% triethylamine); yield 35%; m.p. 120-122*C (needles from ether-hexane); Found: C, 80.50, H, 6.23, N, 8.19. C22H20N20 requires C, 80.46, H, 6.14, N, 8.53%. 8H (CDC13) 1.9 (2H, m), 2.5 (2H, m), 3.6 (2H, m), 6.8-7.3 (13H,











m) and 7.4 (1H, d, J 12 Hz). 6C (CDCl3) 23.4 (CH 2), 26.5 (CH2), 45.4 (CH 2), 122.5 (aryl CH), 123.3 (aryl CH), 124.8 (aryl CH), 125.1 (aryl CH), 126.0 (aryl CH), 128.1 (aryl CH), 128.7 (aryl CH), 129.8 (aryl C), 138.7 (aryl C), 143.9 (aryl C) and 159.3 (C=O).



1-[(4'-methyl)phenyl]-l-(indolin-2-yl)carbinol (2.40).

yield 52%; erythro:threo = 55:45.

threo-(2.40): Rf 0.4 (silica-gel, 15% ethyl acetate in hexane containing 15% triethylamine); m.p. 130-1330C (reprecipitated from chloroform-hexane); M+ (HRMS) found: 239.1331. C16H17NO requires 239.1310. 6H (CDCI3/D20) 2.3 (3H, s), 3.0-3.3 (2H, m), 3.55 (IH, m), 4.6 (1H, d, J 8 Hz),

6.5-6.7 (2H, m) and 7.0-7.4 (6H, m). 8C 20.4 (CH3), 48.6 (CH-N), 49.1 (CH2), 75.2 (CH-OH), 108.9 (aryl CH), 117.6 (aryl CH), 125.4 (aryl CH), 126.0 (aryl CH), 126.9 (aryl CH), 128.0 (aryl CH), 129.9 (aryl C), 136.0 (aryl C), 139.9 (aryl C) and 151.3 (aryl C).



erythro-(2.40): Rf 0.3 (silica-gel, 15% ethyl acetate in hexane containing 15% triethylamine); m.p.123-125*C (reprecipitated from benzene-pet.ether); M+ (HRMS) found: 239.1316. C16H17No requires 239.1310. 6H (CDC13/D20) 2.35 (3H, m), 3.5-3.7 (3H, m), 4.85 (1H, d, J 6 Hz), 6.5-6.65 (2H, m) and 7.0-7.3 (6H, m). 6C 21.1 (CH3), 49.3 (CH-N),











49.7 (CH2), 75.6 (CH-OH), 110.0 (aryl CH), 118.5 (aryl CH), 125.0 (aryl CH), 126.5 (aryl CH), 128.1 (aryl CH), 128.6 (aryl C), 129.0 (aryl CH), 137.3 (aryl C), 139.5 (aryl C) and 152.2 (aryl C).



1-(Indolin-2-yl)cyclohexanol (2.41).

Rf 0.5 (silica-gel, 15% ethyl acetate in hexane

containing 15% triethylamine); yield 57%; m.p. 108-110*C (needles from ether-hexane); Found: C, 77.36, H, 9.03, N,

6.14. C14H19NO requires C, 77.38, H, 8.81, N, 6.45%. SH (CDCI/D20) 1.3-1.8 (10H, m), 3.25 (1H, t, J 7 Hz), 3.6 (2H, d, J 7 Hz), 6.6-6.7 (2H, m) and 7.0-7.3 (2H, m). 8C 21.6 (CH2), 25.7 (CH2), 33.8 (CH2), 34.4 (CH2), 48.8 (CH2), 52.9 (CH-N), 73.6 (C-OH), 109.6 (aryl CH), 118.1 (aryl CH), 126.3 (aryl CH), 127.9 (aryl CH), 128.5 (aryl C) and 152.5 (aryl C).

















CHAPTER III

a-HYDROXYCARBANIONS



3.1. Introduction


In alcohols, the inductive electron withdrawal by oxygen induces a polarization of the carbon-oxygen bond. As a result, the a-carbon bears a small positive charge, rendering the attached hydrogens acidic. This could allow the generation of an c-hydroxycarbanion, leading to the possibility of electrophilic substitution at the c-carbon atom of the alcohol. In practice, however, a-hydroxy carbanions are very difficult to generate. First, the more acidic hydroxyl proton needs suitable protection. A more severe problem appears to be the instability of carbanions adjacent to oxygen. Such carbanions, once generated, experience strong repulsion from the oxygen lone pairs making them energetically unfavorable. Hence, a-hydroxycarbanions, or their synthons, remain as challenging targets to synthetic chemists.











a-Hydroxycarbanions (or their synthons) hold tremendous promise in their synthetic utility. Such synthons offer easy access to 1,3- and 1,4-dihydroxy carbon fragments which are key segments in numerous antibiotics and, until now, have only been made by lengthy synthetic operations.



3.1.1. Indirect Routes to a-Hydroxy Carbanions

The e-hydrogens of an alcohol are not acidic enough to be abstracted by a strong base. Therefore, indirect methods have been adopted to generate a-hydroxy carbanion synthons. The most notable approach is the transmetallation (tinlithium) between tributylstannylmethanol (3.1) and n-butyllithium which was reported by Seebach [80CB1290]. This actually gives rise to a pentavalent "tin-ate" complex (3.2) (and not LiCH2OLi) [80CB1290] which reacts with electrophiles to yield the a-substituted methanol derivatives (Scheme 3.1).

2 n-BuLi E- E+


Bu3SnCH2-OH > 1Bu4Sn\/O Li > ECH2OH


(3.1) (3.2)


Scheme 3.1

Lithium-metalloid exchange (e.g. reductive cleavage of monothioketals [84JAI1301 or lithiodehalogenation of a-chloro ethers [64TL1503]) have also been used in generating -lithio ethers but with limited success.











3.1.2. Protection-Activation Route to a-Hydroxycarbanions

The protection-activation route to a-hydroxycarbanions is a direct approach to a-hydroxycarbanion synthons. This involves three distinct steps (Scheme 3.2):

i) protection of the more acidic hydroxyl proton

ii) a-lithiation and subsequent alkylation with an

electrophile, and finally

iii) deprotection of the hydroxy group.


step (i) step (ii) Li
I
RCH2-OH > RCH 2-0-P > RCH-O-P
+ P



E step (iii) E RCH-OH < RCH-0-P
- P


P = Protecting Group


Scheme 3.2


The crucial step in this sequence is the generation of

the a-lithio species via lithium--a-hydrogen exchange. Since the a-hydrogen(s) of alcohols are not sufficiently acidic to be abstracted by strong base the process of c-lithiation requires additional activation. This can be achieved in various ways. The most common practice is to choose a











choose a suitable protecting group which simultaneously protects the hydroxy function and directs the lithiation on to the a-carbon atom. As would be expected, the choice of an appropriate protecting group is vital to the overall success of the process. This has been nicely demonstrated by Beak in his study on the a-lithiation of methanol and ethanol via their aryl esters (3.3) (77JA5213] (Scheme 3.3). In that report, the alcohol was first protected in the form of a highly hindered ester (3.3). a-Lithiation was achieved with s-butyllithium in presence of tetramethylethylenediamine to form the a-lithio species 3.4. Alkylation of the latter, followed by deprotection with lithium aluminum hydride, liberated the a-alkylated alcohols 3.6. Thus, the a-lithio species (3.4) behaved as an a-hydroxy carbanion equivalent.


0 s-BuLi Li.0 .-.. 0
R-CH 0--C /H(3.6) __ > R-CH-0-C / -(3.3) (3.4) 1E+



R-H()-H LiAlH4 E 0 R-CHE)-H (3.5)


R = H, CH3


Scheme 3.3











Several features of this work are worthy of discussion. It should be noted that a highly hindered aryl group was used. This was necessary to prevent any nucleophilic attack on the carbonyl group of the aryl ester (3.3). Ironically, this became a serious problem in the latter stage when deprotection of 3.5 by conventional hydrolysis could not be achieved. A very harsh deprotection step involving LiAlH4 had to be employed which is a serious drawback to this method. Nevertheless, the a-lithiation step (3.3 -> 3.4) deserves special attention since it involves the process of

-activation. According to Beak, -activation is derived through a "prior complexation" phenomenon where the carbonyl oxygen of (3.3) first coordinates with the lithiating base (s-butyllithium). This evidently directs the delivery of the s-butyl residue on to the a-hydrogens of 3.3. Similar arguments have been forwarded to explain the regio- and stereochemistry of various lithiation processes [86ACR356].

a-Lithiation of an alcohol is much easier when an

activating (mesomeric) group is present in the a-position. Thus upon treatment with a strong base (t-butyllithium) benzyltrimethylsilyl ether gives rise to the a-lithio species (3.7) [74JA3214]. Allyl silyl ethers can also be metallated to generate the delocalized lithio species (3.8) [76JOC3620].












Li Li
Ph-CH-OSiMe3 -Z OSiMe 3 3

(3.7) (3.8)




In recent years, great strides have been made in the

utilization of silicon in organic synthesis (79MI1, 82MI1]. Organosilicon chemistry has shown two general trends. A silicon atom usually favors a positive charge "beta" to itself whereas it is capable of stabilizing an a-carbanion. Both these effects are due to the availability of d-orbitals on silicon. Thus the negative charge on an cc-silyl carbanion can be effectively delocalized into the vacant d-orbitals of silicon. This has been amply demonstrated by Magnus who led the field in the application of a-silyl carbanions [80M14]. A plethora of a-silyl carbanions are currently known, many of which also possess an a-heteroatom. Some of these, which have been useful to synthetic chemists, are shown below [80M14].

Li Li Li

Me 3Si OMe Ph,-''SiMe 3 Cl SiMe3


(3.11)


(3.9)


(3.10)











To us, the most encouraging example was that of a-lithio trimethylsilylmethyl methyl ether (3.9) [79CC822, 820M5531, since it shows that an =-trimethylsilyl group can also provide enough activation to generate z-oxy carbanions. The o-lithio species (3.9) was shown to react with carbonyl electrophiles (aldehydes and ketones) and the derived products were generally obtained in fair to good yields [820M5531.

Returning to the problems associated with

a-oxycarbanions, one is faced with the facile 1,2-migration these species undergo via the Wittig rearrangement [70AG(E)763J. Thus, in sharp contrast to the c-lithio-methyl ether derivative (3.9), the corresponding a-lithio-aryl ether species (3.13) undergoes Wittig rearrangement when kept at room temperature [82JOC5051] (Scheme 3.4).


s-BuLi Li Ar
_ _ I
ArO-CH2 SiMe3 > Ar-O-CH-SiMe3 -> LiO-CH SiM e3
(3.12) (3.13)
(3.14)


i)H20
ii)KF

HO-CH2-Ar

(3.15)


Scheme 3.4











1,2-Anionic migrations in c-oxycarbanions are also

common. Thus, both 1,2-acyl migration (e.g. 3.16 - > 3.17) [81JOC2363] and 1,2-phosphoryl migration (e.g. 3.18 - > 3.19) [80S289, 76TL471 have been reported for the respective a-oxycarbanions (Scheme 3.5). These rearrangements involve the transposition of the oxygen protecting group to the a-carbon. Therefore, the choice of an appropriate protecting group is of utmost importance for generating an c-oxy carbanion, especially via the protection-activation methodology.


0 11
C-R

OLi (3.17)


Li
_ O-C-R
II 0
(3.16)



Li

2
0


0
II


OLi (3.19)


(3.18)


Scheme 3.5











The current status of the synthetic art concerning a-hydroxycarbanions can be summarized as follows:

a) In the absence of a suitable activating group,

a-hydroxycarbanions are very difficult to generate.

b) Once generated, c-oxycarbanions tend to undergo

1,2-migrations either via the Wittig rearrangement or

by a 1,2-anionic shift.

These limitations are borne out by the experimental facts; until now only two a-hydroxy carbanion synthons, (3.2) and (3.4), have been reported which are derived from unactivated alcohols.



3.1.3. Carbon Dioxide as a Protecting-Activating Group

The 1,2-migrations, which are so problematic in the generation of a-oxy carbanions, could be minimized, in principle, if a-lithiation of anionically protected alcohols could be achieved. A proposed sequence is shown in Scheme

3.6. First the alcohol is protected by using a suitable protecting group (preferably a cumelene, Y=X=Z) so as to generate an anionically protected species like 3.20. Successful a-deprotonation of 3.20 would give rise to the a-lithio species 3.21, in which any 1,2-anionic migration would be electronically disfavored. Alkylation of 3.21 followed by simple hydrolysis would afford the desired a-alkylated alcohols.












i) Base Y R"Li Li Y
RCH2 OH > R-CH20-X<-- > R-CH-O-X
Ii) E+

ii) H20/H+


E
I
R-CH-OH


Scheme 3.6


Earlier in section 2.2.3, carbon dioxide was introduced as a novel protecting group for amines. The resulting lithium carbamates proved to be useful substrates for oc-lithiation of amines. Carbon dioxide appeared equally promising as a protecting group for alcohols. The derived hemicarbonate salts are very unstable in presence of moisture and rapidly lose carbon dioxide to return to the alcohol. In spite of their facile deprotection, the inherent instability of the hemicarbonates has restricted their synthetic utility as protected alcohols. To us, however, the hemicarbonate salts appeared ideally suited for ot-lithiation studies: (a) carbon dioxide is a non-toxic, cheap commercial material, (b) the hemicarbonate salts aptly qualify as anionically protected alcohols and (c) their deprotection would be extremely rapid in the presence of aqueous acid.











The above hypothesis has already been put to test by previous workers in this group. Benzyl alcohol (3.22) was chosen by those workers as the model hydroxy compound: treatment of this alcohol with n-butyllithium followed by CO2 readily gave the corresponding lithium carbonate (3.23). When exposed to t-butyllithium, the latter underwent smooth metallation giving rise to the x-lithio species (3.24) which was trapped with different electrophiles (D20, CH3I, C02, PhCO2Et, Ph2CO). Deprotection occurred spontaneously during acidic work-up and the a-alkylated benzyl alcohols (3.25) were isolated in fair to good yields [87S4151 (Scheme 3.7). As an added advantage, the entire sequence was carried out in the same reaction vessel.




i) n-BuLi 0 t-BuLi
PhCH2OH > PhCH2O-C/ > Ph-CH-0
2 I \ ii) CO2 OLi Li C-OLi ~0
(3.22) (3.23)
(3.24)



IE+


H + /H 20
H/20
PhCH-OH < Ph-CH-OCO Li
1 1 2
E -CO2 E
(3.25)


Scheme 3.7











Successful -lithiation of lithium benzyl carbonate

(3.23) is in sharp contrast to Seebach's study [80CB1304], the latter showing that benzyl alcohol on treatment with two equivalents of n-butyllithium afforded ortho-lithiation on the phenyl ring (Scheme 3.8).


2 n-BuLi Li... OLi OH
TMEDA







E

CL-1 OH


Scheme 3.8

In this context, a-lithiation of benzyltrimethylsilyl ether (3.26) is also worthy of discussion. c-Lithiation of

3.26 can be achieved with t-butyllithium to produce the a-lithio species (3.7), which however, undergoes a rapid 1,2-silyl migration as shown in Scheme 3.9 [74JA3214]. Such 1,2-migration was completely suppressed during the a-lithiation of lithium benzyl carbonate (3.23).












t-BuLi / - SiMe3
PhCH2-O-SiMe3 > PhtH--B)
(3.26) (3.7)



SiMe
/ 3
PhCH (3.27)
0

Scheme 3.9


3.1.4. Aims of the Work
As discussed above, anionically protected alcohols have been introduced as superior substrates for a-lithiation and initial results with benzylalcohol via its lithium carbonate have shown considerable promise [87S415]. However, benzylalcohol is regarded as an activated system since the a-carbanion is stabilized by the adjacent phenyl group. Therefore, in the present work, it was our aim to probe into the limits of a-lithiation on lithium carbonates and for possible extension of this methodology to other alcohols.

The lithium carbonate of methanol was chosen as the

primary target, for two main reasons. Firstly, x-lithiation would generate a primary carbanion in this case. Any other alcohol would give rise to either a secondary or a tertiary carbanion which are more difficult propositions, due to electronic as well as steric reasons. Secondly, a-lithiation studies on lithium methyl carbonate would provide a direct comparison with Beak's aryl ester methodology [77JA52131 depicted in Scheme 3.3. The role of the lithium carbonates,











if any, in the process of a-activation could thus be properly evaluated.

a-Lithiation of 1-trimethylsilylmethanol (via its

lithium carbonate) appeared very promising for two main reasons. Firstly, the trimethylsilyl group has the proven ability to stabilize an a-carbanion. Secondly, a judicious choice of the electrophile would enable the trimethylsilyl auxilliary to be readily removed from the product. The latter prediction has strong literature precedents [74ACR771 in that a silicon atom is extremely labile when adjacent to carbonyl functions. Thus the successful a-lithiation of lithium 1-trimethylsilylmethyl carbonate would formally constitute the generation of a methanol dianion synthon.

a-Lithiation of allylic alcohols has been the subject of numerous synthetic efforts in the past and present [84AG(E)932]. The generated a-carbanions can be effectively delocalized to the y-carbon and, as such, varying degrees of selectivity (a vs. y) are observed in the alkylation of these anions. Several oxygen protecting groups with varying co-ordinating powers have been investigated to control this selectivity [84AG(E)932]. In this perspective it was of considerable interest to investigate the directing power of











an allyl lithium carbonate. A strong chelating effect (five-membered chelate) from the carbonate function would be expected to increase the proportion of cx-alkylation over the y-attack. The extent to which this selectivity could be achieved was the principal objective in this case.



3.2. Results and Discussion



3.2.1. a-Lithiation Studies on Lithium Methyl Carbonate

Lithium methyl carbonate (3.29) was prepared in a manner similar to that of benzyl alcohol. Thus methanol (3.28) on treatment with n-butyllithium followed by carbon dioxide produced the lithium methyl carbonate (3.29) (Scheme 3.10). a-Lithiation of 3.29 was attempted with t-butyllithium in tetrahydrofuran at -201C for 2 hrs. Subsequent addition of benzaldehyde followed by acidic work-up did not afford the desired 1,2-diol (3.31), benzaldehyde being recovered unchanged. A repeat reaction replacing benzaldehyde with ethyl benzoate failed, as well. Addition of HMPA or TMEDA to the reaction mixture also showed no promise; it was inferrred that the desired lithium lithiomethyl carbonate (3.30) was not formed under these conditions.













i) n-BuLi t-BuLi
CH3OH > CH3OCO2Li /7-> LiCH2OCO2Li (3.28) ii) CO2 (3.29) (3.30)


i) PhCHO
ii) H'


PhCH-CH2OH
OH
(3.31)


Scheme 3.10

A comparison of these results with Beak's aryl ester methodology (cf. Scheme 3.3) shows that the lithium carbonates have no perceptible a-activating effects. Thus it became evident that the hemicarbonate route to a-oxy carbanions can be applied only to activated systems. This led to the prospect of utilizing a trimethylsilyl group as an easily removable a-activating auxilliary.


3.2.2. a-Lithiation Studies on Lithium Trimethylsilylmethyl

Ca rbonate

Commercially available 1-trimethylsilylmethanol (3.32) was easily converted to its lithium carbonate (3.33) following the standard procedure (n-butyllithium - C02). The latter was then treated with s-butyllithium in











tetrahydrofuran at -781C. After stirring the virtually colorless solution for 2 h at -250C, ethyl p-toluate was added as the electrophile and the resulting mixture stirred at room temperature for 8 hrs. Subsequent work-up with 10% aqueous hydrochloric acid afforded the a-hydroxyacetophenone derivative (3.36) in 68% yield (Scheme 3.11). The use of s-butyllithium in the metallation step was found to be essential as inferior yields were obtained when n-butyllithium or t-butyllithium were used.


i) n-BuLi 0 s-BuLi Ii
Me3SiCH 2OH > Me3SiCH2OCOLi
(3.32) ii) CO2 (3.33) THF/-250C



0
Lf C -OLi
Me 3SiCH-O
(3.34)


ArCO2Et


H+/H20
ArCO-CH2OH < Me 3SiCH-OCO2Li
(3.36) COAr (3.35)
Ar = p-Me-C6 H4


Scheme 3.11











Formation of the a-hydroxyketone (3.36) can be

rationalized via the intermediacy of the a-lithio species (3.34). Reaction of 3.34 with the ester electrophile presumably led to the a-silyl ketone derivative (3.35), which on hydrolysis lost CO2 with concommitant protiodesilylation to give the hydroxymethylketone (3.36) (Scheme 3.11). The facile desilylation of 3.35 is to be expected since a-silyl carbonyl compounds are known to be readily desilylated upon hydrolysis [74ACR771.

Other carbonyl electrophiles reacted in the same way as ethyl p-toluate. The results are summarized in Table 3.1. Thus N,N-dimethylbenzamide and benzoyl chloride led to a-hydroxyacetophenone, as expected. The low yield obtained with benzoyl chloride was probably caused by undesired side reaction involving metal-halogen exchange. Benzonitrile was also successfully employed as the electrophile; the intermediate imine adduct hydrolyzed to the ketone on acidic work-up. Quite significant are the results when a,O-unsaturated esters were used as electrophiles (entries 5 and 6, Table 3.1). In both cases, carbonyl addition was the exclusive pathway giving rise to the hydroxymethyl vinyl ketones. The 1,4-addition mode was not observed.











Table 3.1. Reaction of the a-Lithio Lithium Carbonate (3.34)
with Electrophiles

Electrophile Product Yield(%)

1. P-Me-C6H4CO2Et p-Me-C6H4COCH2OH 68 2. PhCOCI PhCOCH2OH 23 3. PhCONMe2 PhCOCH2OH 63 4. PhCN PhCOCH 2OH 65

5. PhCH=CHCO 2Et PhCH=CHCOCH2OH 45 6. Me2C=CHCO2Et Me 2C=CHCOCH2OH 60




All the hydroxymethyl ketones prepared in this study were characterized by their NMR spectra ( H and 13C) and through matching melting point data (in case of known compounds). The 1H-NMR spectra of these compounds showed a characteristic two-proton singlet around 4.8 ppm which was assigned tc the methylene protons. Their 13C-NMR spectra displayed the keto-carbonyl, usually near 198 ppm, whereas the methylene carbon appeared around 67 ppm.

Saturated aliphatic esters such as ethyl propionate or ethyl butyrate, however, failed to provide homogenuous products when used as electrophiles. Similar failure was also observed with other potential acylating agents like propionic anhydride and N,N-dimethylpropionamide. Dimethyl carbonate and phenyl isocyanate also failed to give the desired products.











The results obtained thus far (Table 3.1) did not

necessarily prove the formation of the x-lithio species (3.34). There are reports that alkyllithiums attack the silicon center in certain Si-C bonds giving rise to the C-Li species [80MI4]. Such an attack on the lithium carbonate (3.33) could potentially give rise to the lithiomethyl lithium carbonate (3.30) (Scheme 3.12). The latter, upon reaction with esters, would afford the same results.





Li 2 C-OC 2-Si e 3 > LiO 2 C-OCH 2Li
(3.33) (3.30)


1 RCOX



RCOCH2OH


Scheme 3.12


The nature of the a-lithio species was decisively

settled by using benzaldehyde as the electrophile. It was envisaged that reaction of benzaldehyde with the silyl-a-lithio species (3.34) would initially produce a 0-hydroxysilyl adduct (3.37). The latter, after Peterson olefination [84S384] and subsequent hydrolysis of the incipient styryl carbonate (3.38), should lead to phenylacetaldehyde (3.39) (Scheme 3.13). On the other hand,











if a-lithiomethyl lithium carbonate (3.30) was indeed the reactive species, then its reaction with benzaldehyde would produce the diol (3.31).


Me3SiCH2 0-CO2Li (3.33)


s-BuLi


Li
Me3 CHO-CO2Li
(3.34)


PhCHO


Me 3SiCHO-CO2Li Lio Ph (3.37)


LiCH2O-CO2Li
(3.30)


PhCHO

V

PhCH-CH OH
1 2
OH
(3.31)


-OCO 2Li


Ph Il


H +/H2 0
> PhCH2CHO

(3.39)


(3.38)


Scheme 3.13











In practice, a-lithiation of the lithium carbonate (3.33) with s-butyllithium (-25*C, 2 hrs) followed by addition of benzaldehyde produced phenylacetaldehyde (3.39) in 70% yield, as the sole product. This proved that the silyl-a-lithio species (3.34) is indeed formed and no lithiodesilylation (to give 3.30) had taken place. This also reinforced the earlier view that with the other carbonyl electrophiles a-silyl-a-hydroxy ketones are the initial products which protiodesilylate upon hydrolysis.

The fact that lithium a-lithiotrimethylsilylmethyl carbonate (3.34) converted benzaldehyde to phenylacetaldehyde immediately opens up a novel homologation process for aldehydes and ketones (Scheme 3.14). These transformations, although very common in multistep synthesis, suffer from the lack any general preparative method [76T1943]. The synthon (3.34) promised a one-pot procedure to achieve such transformations.


Me 3SiCH- i + R 2 > Me 3SiCHOCO 2Li Li 0 LiO
M3i-OCO2Li+ R-c-R2 2i


(3.34) R1 R2


~1

R1Hf+/H 0 _-OCO Li 'CH-CHO 2 11 R2/ - CO2 Ri/N R2 (3.40)


Scheme 3.14











In practice, reaction of the a-lithio species (3.34) with cyclohexanone resulted in a complex product mixture. Addition of MgBr2 or BF3.Et20 to the reaction mixture did not lead to the desired cyclohexane carboxaldehyde. In view of the enolization problems with cyclohexanone, benzophenone was next tried as the electrophile; however, the desired product could not be obtained. Further study with 3.34 as a reductive formyl anion equivalent was therefore abandoned.

At this time our attention was drawn to the recent

fluoride induced desilylative alkylation of a-silylbenzyl alcohol [84CL1803]. However, no reaction was observed when benzaldehyde and the lithium carbonate (3.33) were together treated with (Bu)4N+F- (in tetrahydrofuran, 250C, 24 hrs) or with CsF (in hexamethylphosphoramide, 60�C, 6 hrs). In another study, Sakurai has shown that one of the Si-C bonds in bis(trimethylsilyl)methane could be cleaved by sodium methoxide in hexamethylphosphoramide to produce the trimethylsilylmethyl carbanion [73TL4193]. A similar treatment of the lithium carbonate (3.33) was not fruitful and returned the starting materials. Generation of the methanol dianion equivalent directly from the lithium carbonate (3.33) was therefore deemed impossible.











3.2.3. c-Lithiation Studies on Lithium Cinnamyl Carbonate

Previous work in this group showed that allyl lithium carbonate could not be a-lithiated with t-butyllithium [86MI11. Hence, a more activated system such as cinnamyl alcohol (3.41) was chosen for the present work. The required lithium cinnamyl carbonate (3.42) could be prepared very easily as shown in Scheme 3.15. The lithium carbonate (3.42) was next treated with t-butyllithium followed by addition of D20. Hydrolytic work-up followed by 1H-NMR analysis of the product showed no deuterium incorporation in the molecule. This evidently showed that the desired allyl carbanion (3.43) could not be generated.


i) n-BuLi t-BuLi
Ph>'. OH > Ph oCo2Li //
ii) CO2
(3.41) (3.42)

LiO2C-O Ph

Li
(3.43)


Scheme 3.15

Since other methods exist which allows for the generation of x-oxyallyl carbanions [84AG(E)932, 81AG(E)127], further investigation on the allyl hemicarbonates was discontinued.











3.3. Conclusions



The advantage of using carbon dioxide as a hydroxyl

protecting group has been efficiently demonstrated. It is easily introduced and subsequent deprotection of the resultant hemicarbonates can be achieved under extremely mild conditions. The lithium carbonates are virtually inert to most metallating agents and the process of a-lithiation is devoid of the notorious 1,2-anionic migration. However, the lithium carbonates play no role in the process of a-activation. Presumably, the delocalized negative charge on the carbonate moiety reduces its co-ordinating power and the carbonyl oxygen fails to activate the lithiating agent (via the prior complexation mechanism [86ACR356]).

Lithium carbonates with appropriate a-activating auxilliaries promise very interesting synthetic transformations. As a result of the present investigation, lithium c-lithiotrimethylsilylmethyl carbonate (3.34) has emerged as a novel c-hydroxycarbanion synthon. In spite of its limitations, a novel synthesis of azhydroxyketones has been developed by proper utilization of the synthon 3.34 and that too in a one-pot sequence starting from a commercially available material. a-Hydroxyketones are traditionally prepared via electrophilic hydroxylation of ketone enolates,











a procedure not without complications. Direct oxygenation often results in complex product mixtures due to overoxidation [62JI578,68JOC3294,68JOC3695]. The MoO5.Py.HMPA complex has also been used for this purpose but the products from methylketone enolates were found to undergo further condensation with the starting material, leading to poor yields [78JOC188]. However, Rubottom's m-CPBA oxidation of silyl enol ethers [78JOC1599] or Moriarty's method using hypervalent iodine oxidation [84TL691] are quite effective. By contrast, our route to hydroxymethyl ketones utilizes a novel sequence which is conceptually opposite to the existing ones and demonstrates the viability of an umpolung approach to these compounds.



3.4. Experimental


Melting points were determined using a Bristoline

hot-stage microscope and are uncorrected. NMR spectra were recorded on a Varian XL200 instrument using Me4Si as the internal standard for protons and CDCI3 (77.0 ppm) as reference for carbon resonances. The nature of the carbon signals in the 13C-NMR spectra were determined by the











Attached Proton Test [81CC150, 82JMR535]. Mass spectra were recorded at 70 ev on a AEI MS instrument attached to a DS 55 database. All reactions were carried out in oven-dried (120*C, overnight) apparatus under a slight positive pressure of dry argon and transferring operations done using syringe techniques or via cannula. Tetrahydrofuran (THF) was freshly distilled from sodium-benzophenone ketyl just prior to use. Hexamethylphosphoramide (HMPA) and tetramethylethylenediamine (TMEDA) were distilled over CaH, under reduced pressure and stored under nitrogen. Flash chromatography [78JOC2923] was carried out over silica gel (MCB, 230-400 mesh) using pre-distilled solvents.



General Procedure for the Preparation and a-Alkylation of Lithium Trimethylsilylmethyl Carbonate (3.33): A 250 ml Schlenk reactor (oven-dried at 120*C overnight) was evacuated under 0.5 mm for 2 hrs and purged with argon. 1Trimethylsilylmethanol (3.32) (Aldrich or Petrarch) (0.52 g,

5.0 mmole) was added followed by dry THF (35 ml). After cooling to -781C, n-BuLi (2.1 ml, 2.5 M in hexanes, 5.2 mmole) was added dropwise. The reactor was warmed to room temperature and held there for 10 min. It was then recooled











to -250C and dry CO2 gas bubbled through it. The cooling bath was removed and while CO2 was still bubbled in, the reactor was allowed to warm to room temperature. After a further 5 min all volatile contents were evacuated at 0.5 mm leaving a white powdery residue of the lithium trimethylsilylmethyl carbonate (3.33). The reactor was again purged with argon and dry THF (45 ml) was added to redissolve the white residue. After cooling to -78'C, s-BuLi (4.2 ml, 1.3 M in cyclohexane, 5.5 mmole) was added dropwise. Stirring was continued at -250C for 2 h. To this virtually colorless solution was then added the appropriate electrophile (5.2 mmole) in dry THF (15 ml). The reactor was warmed to room temperature and the contents stirred for 30 min. The THF was removed under reduced pressure (on a rotary evaporator) and the residue treated with 10% aqueous HCI. After stirring for 15 min the aqueous phase was extracted with ether or ethyl acetate (3 x 15 ml). The combined organic layers were washed sequentially with water (10 ml), saturated NaHCO3 solution (10 ml) and brine (10 ml) and dried over anhydrous Na2So4. Removal of solvent under reduced pressure gave the products which were purified as described for the individual cases below.











The following compounds were made according to the above procedure:

2-Hydroxy-(4'-methyl)phenylethanone

Ethyl p-toluate (0.82 g) was used as the electrophile. Silica gel chromatography using petroleum ether - 25% ethyl acetate afforded the product (0.50 g, 68%); m.p. 86-89*C (lit. (06CB3757] m.p. 89-900C); SH (CDCI3) 7.9 (2H, d, J 8 Hz), 7.4 (2H, d, 1 8 Hz), 5.0 (2H, s), 3.5-3.2 (1H, broad) and 2.6 (3H, s); 6C (CDCI3) 197.9 (C=O), 145.3 (aryl CH), 130.8 (aryl CH), 129.6 (aryl C), 127.8 (aryl C), 65.2 (CH2) and 21.8 (CH3).



2-Hydroxyphenylethanone

(a) Benzoyl chloride (0.74 g) was used as the

electrophile. Silica gel chromatography using petroleum ether - 20% ethyl acetate afforded the product (0.16 g, 23%); m.p.86-880C (lit. [12LA(394)42] m.p. 89-90*C); 6H 8.07.9 (2H, m), 7.7-7.5 (3H, m), 4.9 (2H, s) and 3.5-3.4 (1H, broad).











(b) N,N-Dimethylbenzamide (0.74 g) was used as the electrophile. Silica gel chromatography using petroleum ether - 20% ethyl acetate afforded the same product (0.42 g, 63%) identical in melting point and 1H-NMR data to above.

(c) Benzonitrile (0.54 g) was used as electrophile.

Silica gel chromatography using petroleum ether - 25% ethyl acetate afforded the same product (0.44 g, 65%) identical in melting point and 1H-NMR data to above.



trans-l-Hydroxy-4-phenylbut-3-en-2-one

trans-Ethyl cinnamate (0.88 g) was used as the

electrophile. Silica gel chromatography using petroleum ether - 20% ethyl acetate afforded the product (0.34 g, 42%); m.p. 65-68C (lit. [84TL691] m.p. 69-70.5C); 6H (CDCl3) 7.9-7.5 (6H, m), 6.8 (1H, d, J 15 Hz), 4.5 (2H, s) and 2.9 (broad, OH); 8C (CDCI3) 198.1 (C=O), 144.1, 133.4, 131.1, 129.0, 128.4, 121.0 and 67.0 (CH2).



l-Hydroxy-4-methylpent-3-en-2-one

Ethyl 3,3-dimethylacrylate (0.64 g) was used as the electrophile. Silica gel chromatography using petroleum ether - 10% ethyl acetate afforded the product (0.34 g, 60%)











as an oil; 8H (CDCI3) 5.9 (1H, s), 4.2 (2H, s), 2.3 (broad, OH), 2.2 (3H, s) and 1.9 (3H, s); 6 C (CDCl3) 198.0 (C=O), 158.9 (C-3), 119.0 (C-2), 68.6 (CH2), 27.8 (CH3) and 21.4 (CH3); HRMS: M+" found 114.0728, C6H1002 requires 114.0681.



Phenylacetaldehyde

Benzaldehyde (0.52 g) was used as the electrophile.

Filtration through a short column of neutral alumina using hexane-2% ethyl acetate afforded the product (0.52 g, 78%); 6H (CDCI3) 9.74 (1H, t, 1 2 Hz), 7.4-7.2 (5H, m) and 3.67 (2H, d, J 2 Hz).


















CHAPTER IV

THERMAL BEHAVIOR OF N-VINYL-1,2-DIHYDROPYRIDINES



4.1. Introduction



4.1.1. 1,2-Dihydropyridines

Of the five possible classes of dihydropyridine

(4.1-4.5), the 1,2- (4.1) and 1,4-dihydropyridines (4.2) have attracted most attention. The rest have been less studied because of their instability and the lack of good preparative methods.


I

R (4.1)


I

R (4.2)


N



(4.3)


(4.4)


No



(4.5)


The chemistry of dihydropyridines has been extensively reviewed [72CRV1,82CRV223,82MII]. As compared to











1,4-dihydropyridines, which have been widely exploited as NADH mimics, 1,2-dihydropyridines until recently, have found less prominence. This may be partly attributed to their lower stability [72JA5926]; 1,2-dihydropyridines are readily oxidized to the corresponding pyridines. In recent years, however, 1,2-dihydropyridines have found widespread use in natural product synthesis; these highly reactive compounds have proved to be important building blocks in the synthesis of various alkaloids, especially for the indoloquinolizidine [75ACS(B)655,76ACS(B)251,80TL23411, elaeocarpine [71TL4395] and cantharanthine (78TL5157,79TL2485,80JOC3382] families.



4.1.2. Synthetic Routes to 1,2-Dihydropyridines

Due to their inherent instability there are only few methods of preparing 1,2-dihydropyridines. These are described below.



4.1.2.1. Reaction of Nucleophiles with Pyridines

Strong nucleophiles such as phenyllithium add to

pyridine exclusively at the 2-position to form the highly reactive N-lithio-l,2-dihydropyridine (4.6) [71TL4961]. This is currently the only practical way of synthesizing 1,2-dihydropyridines directly from pyridines (Scheme 4.1).
















+ PhLi ether, 00C N Ph
NI
Li

(4.6)


Scheme 4.1


Normally pyridines (unless highly deactivated by

electron withdrawing groups) are not reduced by sodium borohydride. However, in presence of alkyl chloroformates, sodium borohydride reduces the incipient pyridinium salt to give high yields of N-carbamoyl-l,2-dihydropyridines [72JOC1321] (Scheme 4.2). In an analogous manner, other nucleophiles such as Grignard reagents [82JOC4315,83TL1801, 84TL4867] and cyanide ion (to form Reissert compounds) [85CI(L)125] have also been successfully employed.





I + NaBH4 >
N CH3OH, -700C N
CO R
CO2R


Scheme 4.2











4.1.2.2. Reaction of Nucleophiles with Pyridinium Salts

Pyridinium salts are more susceptible to nucleophilic

attack than the corresponding pyridines. However, depending on the nucleophile, the regioselectivity of the attack (1,2vs. 1,4-) varies and, quite often, mixtures of 1,2- and 1,4-dihydropyridines are obtained. This problem has been addressed with reference to the "hardness" or "softness" of the nucleophile [65TL4615] and there is evidence that in some cases a mixture of 1,2- and 1,4-isomers are formed under kinetic control which subsequently equilibrates to the thermodynamically more stable 1,4-isomer (79JOC1757]. Enhanced 1,2-regioselectivity is generally observed in the reaction of pyridinium salts with organometallic reagents. Thus Grignard reagents generally attack at the 2-position to form 1,2-dihydropyridines [74TL59,71JOC772] (Scheme 4.3). Good yields of 1,2-dihydropyridines are also obtained when phenyllithium or t-butyllithium is used [72CZ411,74JOC59] but simpler alkyllithiums tend to give very low yields. Organocuprates usually attack pyridinium salts at the 4-position [74CJC3563]. Nucleophilic attack on a 4-substituted pyridinium salt invariably occurs at the 2-position. One exception being benzylic organotin reagents, which, in the presence of chloroformates, always attack the 4-position of a pyridine, even when a 4-substituted pyridine is used [86TL211].















RM
N >N R
I +>
R R


R M = PhMgBr 70% EtMgBr 70%
PhLi 80%
n-PrLi 20%


Scheme 4.3


Sodium borohydride reduction of pyridinium salts is difficult to arrest at the dihydropyridine stage and over-reduction to tetrahydropyridines usually occur[53BSF53, 70JOC2809,86AHC(39)l].To summarize, the most synthetically useful methods for preparing 1,2-dihydropyridines are:

i) Reaction of pyridines with sodium borohydride in

presence of chloroformates, or

ii) Reaction of pyridinium salts with Grignard

reagents.



4.1.3. Reactivity of 1,2-Dihydropyridines

The reactivity of 1,2-dihydropyridines ranges from that of an enamine [75CJC2305,76JHC789,79JCS(PI)3082,75JOC569, 78CJC1026,79CJC2342,79JHC409] to that of a cyclic 1,3-diene.











4.1.3.1. Cycloaddition Reactions

Depending on the nature of the N-substituent, 1,2-dihydropyridines behave as either a 4n or a 2n component in cycloaddition reactions. When the N-substituent is an electron withdrawing group, 1,2-dihydropyridines behave as the 4n component in cycloaddition reactions. There are numerous reports of intermolecular Diels-Alder reactions where 1,2-dihydropyridines were effeciently used as the diene component [76JHC481,80JA6157,80CJC2447,81JOC4836, 82TL2527,83TL2927,85JOC3236,85TL2617]. The intramolecular version is also known [74HCA1204,83TL2711].

In absence of electron withdrawing N-substituents,

1,2-dihydropyridines behave more like enamines and hence undergo thermal [2+2] cycloadditions. An example of thermal [2+2] cycloaddition of 1,2-dihydropyridines is shown in Scheme 4.4. Thus 1-phenyl-l,2-dihydropyridine (4.7) on heating with dimethyl acetylenedicarboxylate gave rise to the 1,2-dihydroazocine derivative (4.9). The formation of the latter was rationalized through an initial [2+2] adduct (4.8) which underwent further cycloreversion to produce the observed product [74JCS(PI)2496].
















N
I
Ph
(4.7)


C2Me C2 Me


,CO2 Me CO2Me


(4.8)


CO2Me CO2Me


(4.9)


Scheme 4.4


4.1.3.2. Electrocyclic Reactions

The ANRORC reaction (Addition of Nucleophile - Ring Opening - Ring Closure) is probably the most studied electrocyclic reaction involving 1,2-dihydropyridines. A crucial step in this reaction sequence is the electrocyclic ring opening of the intermediate 1,2-dihydropyridine (4.11) to the triene (4.12) (Scheme 4.5). A detailed mechanistic discussion can be found in the recent reviews on ANRORC-type reactions [80M13,80S589,81T34231. Similar ring openings of pyridine N-oxides are also known [67T2775,69T4291,71JOC1705, 74T4055,76TL4717].


















RNH2


NHR


(4.11)


N NHR
I
Ar


(4.12)


+ C
N
+~


Ar = 2,4-dinitrophenyl


Scheme 4.5

Thermally-induced electrocyclic reactions are very rare for 1,2-dihydropyridines. One such example is the thermal isomerization of l-ethyl-l,2-dihydropyridine (4.13) to the 2,3-dihydropyridine derivative (4.14) (Scheme 4.6) (78JA6696]. The lack of many such examples probably reflects the greater thermodynamic stability of 1,2-dihydropyridines as compared to 1-azatrienes [80MI31.


N
1+
Ar


(4.10)


ArNH2


















N
CH2CH3


[1,7-H]


N CH3


(4.13)


CH3 N CH3


(4.14)


Scheme 4.6


4.1.4. Effects of N-Substituents on the Thermal Behavior of
1,2-Dihydropyridnes

4.1.4.1. N-Alkyl and N-Aryl Substituents

N-Aryl-1,2-dihydropyridines are more stable under pyrolytic conditions than the corresponding N-alkyl derivatives [79JCS(PI)442]. The latter, on pyrolysis, yield the parent pyridine together with the hydrocarbon derived from the N-substituent. This has been effectively exploited












by Katritzky and co-workers for the three-step reductive deamination of primary amines [84JCS(PI)1671] (Scheme 4.7). Although usually limited to allyl, benzyl or heterobenzyl amines, in some cases, it can be extended to alkyl amines.


NaBH4 Ph >


I
R (4.16)


(4.15)


(4.17)


R-H


R-NH2


R - allyl, benzyl or heterobenzyl


Scheme 4.7


4.1.4.2. N-Vinyl Substituents

In a manner similar to the N-alkyl-l,2-dihydropyridines, pyrolysis of a N-vinyl-l,2-dihydropyridine might be expected to yield an olefin corresponding to the N-vinyl substituent; however, this was not the case. When the dihydropyridine (4.18) was pyrolyzed for 2 hrs (1801C, 0.1 mm),











m-chlorostyrene and 2,4,6-triphenylpyridine (4.17) were formed [82JOC492] (Scheme 4.8).



Ph Ph



N Ph Ph N Ph + Ar
/ CH=CH2

Ph "Ar (4.17)


(4.18)


Ar = 3-chlorophenyl


Scheme 4.8

Apparently the result of a rearrangement, the above transformation was rationalized as shown in Scheme 4.9. After initial electrocyclic ring opening of the 1,2-dihydropyridine (4.18), the resultant 3-azatetraene (4.19) could lead to the bicyclic intermediate (4.21) via two consecutive electrocyclizations. The intermediate (4.21) underwent fragmentation and consequent aromatization to afford the observed products. Pyrolysis of some other N-styryl-l,2-dihydropyridines followed the same fragmentation pattern [82JOC492].













Ph


Ph


Ph Ar Ph Ph


(4.18) (4.19) (4.20)






1

Ph


Ar
CH + Ph N Ph <
IIPh N Ph Ar CH2

(4.17) (4.21)


Scheme 4.9


4.1.5. Aims of the work

As discussed above, pyrolysis of

N-vinyl-l,2-dihydropyridines led to an interesting fragmentative rearrangement. The aim of the present project was to extend this observation to 1,2-dihydropyridines having an N-cycloalkenyl substituent. Thus the latter class











of dihydropyridines (e.g. 4.22), on pyrolysis, would be expected to give pyridines 4.23, having long side chains (Scheme 4.10).


K
No
C
* CH


(CH2) (4.22)


N (CH2)nCH (.CH 2
(4.23)


Scheme 4.10

As a further extension, by capping the dihydropyridine

nucleus with cycloalkyl groups as in 4.24, medium-ring fused pyridines such as 4.25 might be obtained (Scheme 4.11).


(4.24)


(4.25)


Scheme 4.11











4.2. Results and Discussion



4.2.1. Synthetic Strategy

We intended to prepare the required

N-cycloalkenyl-l,2-dihydropyridines by sodium borohydride reduction of the corresponding pyridinium salts. Thus our primary synthetic targets were the N-cycloalkenylpyridinium salts. At the time of this work, no such N-vinylpyridinium salts were known where the vinyl grouping was part of a carbocycle. Our strategy for the preparation of such salts comprised of two basic pathways as illustrated in Scheme 4.12. In Path A, it was envisioned that a suitable pyridine (4.26) would displace the halogen from 0-chlorocycloalkenones (4.27), thus leading to N-cycloalkenonylpyridinium salts (4.28). Path B, on the other hand, requires the preformation of a suitable pyridinium salt (4.29) with a O-substituted cycloalkyl group as the N-substituent. The latter, on elimination of HX, might afford the desired N-cycloalkenylpyridinium salts (4.30). The latter pathway is quite similar to the one generally followed for the preparation of acyclic N-vinylpyridinium salts [83JOC40171.















R




N.




(4.28)


Path B:


- HX


X = Cl, Br, OH


Scheme 4.12


4.2.2. Preparation of N-(Cycloalken-3-on-1-yl)pyridinium
Salts

To test the viability of Path A in Scheme 4.12, the

reaction of pyridine and 0-chlorocyclohexenone [76JOC636] was attempted. The reaction was monitored by 1H-NMR (CDC13, 551C) and was found to be quite slow. It was only after


Path A:


R



N
(4.26)


Cl





(4.27)


(4.29)


(4.30)











48 hrs that the a-protons of the pyridinium salt (4.31a) appeared in the spectrum (at - 10.1 ppm ). However, with more nucleophilic pyridines such as p-dimethylaminopyridine (DMAP) the reaction was much faster and the derived pyridinium salt (4.31b) could be isolated in very good yield (94%) (Scheme 4.13). 4-Phenylpyridine could also be used which gave the corresponding pyridinium salt (4.31c) in 57% yield; 4-picoline, however, gave an intractable product mixture. In these reactions, the best yields were obtained when benzene was used as the solvent; in chloroform or acetonitrile the yields were much lower.



R
R Cl
-~ benzene/ I + > N + Cl
N 00



(4.31)
a, R = H; b, R = NMe2; c, R = Ph


Scheme 4.13

When the tricyclic pyridine (4.32) [71JPS(A-I)(9)1807] was reacted with 0-chlorocyclohexenone, the reaction resulted in a complete failure, even when the components were heated in a sealed tube at 200*C for 24 hrs; starting




Full Text

PAGE 1

HETEROATOM MEDIATED MODIFICATION OF REACTIVITY IN CARBON SKELETONS BY SAUMITRA SENGUPTA A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA

PAGE 2

To my Father so close, yet so far

PAGE 3

ACKNOWLEDGEMENTS It was a memorable experience to work under the supervision of Prof. Alan R. Katritzky. I am most grateful for his thoughtful suggestions and untiring persuasion which ultimately led to fruitful results. I thank him, especially, for believing in me and for the independence I received during the later phases of my research. I was rather fortunate to carry the blessings of Prof. Merle A. Battiste, a person whom I could approach both as a friend and as a teacher and from whom I learned most of my chemistry. Prof. W. M. Jones, Prof. W. R. Dolbier and Prof. J. F. Helling also extended their helping hands whenever needed and I thank them for their support. The generosity of Prof. K. S. Schanze and Prof. J. A. Deyrup is deeply appreciated as well. I am very much indebted to Dr. Steve Cato who took great pains in reading through the entire manuscript and correcting my "Indian English." Dr. Ramiah Murugan and Sutha deserve special recognition for providing confidence, easing many of my anxieties and for making me feel at home. Jamshed Lam gave me much needed company both in and out of the lab i i i

PAGE 4

and I thank him for all his help. I also thank all my fellow group members and our secretary Ms. Dawn Sullivan for being on my side all these years and especially Dr. Rick Offerman and Mike Walker for creating such a wonderful working atmosphere in CRB 320. As a friend in need, Dr. Radi Awartani went out of his way to help me on numerous occasions and I am very much thankful to him. Without the wholehearted support of my mother and my sister I wouldn't have been here in the first place. They gave me all that one could ask for. Words cannot describe the sacrifice they undertook and that, too, during a very difficult phase of their lives. I am grateful, yet proud of them and only wish that may I be born in such a loving family again. IV

PAGE 5

TABLE OF CONTENTS page ACKNOWLEDGEMENTS m ABSTRACT v CHAPTERS I. GENERAL INTRODUCTION 1 1.1. Modification of Reactivity by Heteroatoms ... 1 1.2. Amines • 2 1.3. Alcohols 4 1.4. Vinylic Amines 5 1.5. Preview of the Work 8 II. a-AM I NOCARBAN I ON S 2.1. Introduction 2.1.1. Indirect Routes to a-Aminocarbanions . 2.1.2. Protection-Activation Route to a-Aminocarbanions 2.1.3. Carbon Dioxide as a Protecting-Activating Group 2.1.4. Aims of the Work 2.2. Results and Discussion 2.2.1. a-Metallation Studies on Lithium 1,2,3, 4-Tetrahydroquinolinyl Carbamte . 2.2.2. a-Metallation Studies on Lithium Indolinyl Carbamate 2.2.3. Role of Carbon Dioxide in the a-Metallation of Lithium Carbamates . . 2.3. Conclusions 2.4. Experimental III. a-HYDROXYCARBANIONS 3.1. Introduction 45 3.1.1. Indirect Routes to a-Hydroxycarbanions 46 3.1.2. Protection-Activation Route to a-Hydroxycarbanions 47 3.1.3. Carbon Dioxide as a Protecting-Activating Group 53 3.1.4. Aims of the Work 57 9 10 11 16 2 3 24 24 32 35 37 38 v

PAGE 6

3.2. Results and Discussion 59 3.2.1. a-Lithiation Studies on Lithium Methyl Carbonate 59 3.2.2. a-Lithiation Studies on Lithium Trimethylsilylmethyl Carbonate 60 3.2.3. a-Lithiation Studies on Lithium Cinnamyl Carbonate 68 3.3. Conclusions 69 3.4. Experimental 70 IV. THERMAL BEHAVIOR OF N-VINYL-1 , 2-DIHYDROPYRIDINES . 76 4.1. Introduction 76 4.1.1. 1 , 2-Dihydropyr idines 76 4.1.2. Synthetic Routes to 1 . 2Dihydropyr idines 77 4. 1.2.1. Reaction of Nucleophiles with Pyridines 77 4. 1.2. 2. Reaction of Nucleophiles with Pyridinium Salts 79 4.1.3. Reactivity of 1 , 2-Dihydropyr idines ... 80 4. 1.3.1. Cycloaddition Reactions 81 4. 1.3. 2. Electrocyclic Reactions 82 4.1.4. Effects of N-Substi tuents on the Thermal Behavior of 1 . 2Dihydropyr idnes 84 4. 1.4.1. N-Alkyl and N-Aryl Substituents 84 4.1. 4.2. N-Vinyl Substituents 85 4.1.5. Aims of the work 87 4.2. Results and Discussion 89 4.2.1. Synthetic Strategy 89 4.2.2. Preparation of N-(Cycloalken-3-on-l-yl) pyridinium Salts 90 4.2.3. Preparation of N-(Cycloalken-3-on-l-yl)-l,2dihydropyr idines 92 4.2.4. Pyrolysis of N( Cycloalken-3-on-l-yl ) -4-phenyl 1 . 2dihydropy r idine 95 4.2.5. Preparation of N( Cycloalken-l-yl ) pyridinium Salts 97 4.2.6. Preparation of N( Cycloalken-l-yl ) 1 . 2dihydropyr idines 101 4.2.7. Pyrolysis of N-(Cycloalken-l-yl)4-phenyl-l , 2-dihydropyr idine 109 vi

PAGE 7

4.3. Conclusions 4.4. Experimental 109 111 V. THERMAL BEHAVIOR OF 1 , 2-DIHYDROPYRIDINES WITH AN UNSATURATED 4-SUBSTITUENT 120 5.1. Introduction 120 5.1.1. Aim of the Work 121 5.2. Results and Discussion 123 5.2.1. Synthetic Strategy 123 5.2.2. Preparation of 4-Ethoxycarbonylaminopyridinium Salt 125 5.2.3. Attempted Sodium Borohydride Reduction of 4-Ethoxycarbonylaminopyridinium Salt 125 5.2.4. Preparation of 4 -Me thoxyca rbony lmethyl-1-me thylpyridinium Salt 126 5.2.5. Sodium Borohydride Reduction of 4 -Me thoxyca rbony lme thy 1-1 -me thy 1pyridinium Salt 126 5.2.6. Preparation of l-Methoxycarbonyl-4me t hoxy car bony lmethyl1 , 2-dihydropyr idine 128 5.2.7. Thermolysis of l-Methoxycarbonyl-4me thoxy car bony lme thy 11 , 2-dihydropyr idine 129 5.3. Conclusions 130 5.4. Experimental 130 VI. SUMMARY 133 BIBLIOGRAPHY 136 BIOGRAPHICAL SKETCH 149 vi i

PAGE 8

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy HETEROATOM MEDIATED MODIFICATION OF REACTIVITY IN CARBON SKELETONS BY SAUMITRA SENGUPTA DECEMBER, 1987 Chairman: Prof. Alan R. Katritzky Major Department: Chemistry The generation of a-aminocarbanions from secondary amines is a matter of long-standing interest. The usual way to generate such species involves a three-step sequence: protection of the amino proton, a-lithiation and electrophilic substitution followed by deprotection. As a superior alternative, carbon dioxide was investigated as a transient protecting group, which enabled electrophilic substitution at the a-position of tetrahydroquinol ine and indoline in a one-pot sequence. However, the results show that the a-activating effect of lithium carbamates is small and requires strongly basic reagents for a-deprotonation . Carbon dioxide has also been investigated as a protecting group for generating a-hydroxy carbanions. The a-activating power of lithium carbonates was found to be vi 1 1

PAGE 9

indeed very small, even in cinnamyl system. In view of this, a trimethylsilyl group has been used as an easily removable a-activating auxilliary. Thus, 1-tr imethylsilylmethanol via its lithium carbonate was successfully exploited to give a novel methanol dianion synthon in a one-pot sequence. There are a few examples of thermal electrocycl i c ring-opening of 1 , 2-dihydropyr idines to the 1-azatr ienes , a process considered uphill by thermodynamic considerations. It was of interest whether such ring-opening would be favored in N-vinyl-1 , 2-dihydropyr idines . In the latter case, it was argued that the incipient 3-azatetraene may be induced into further non-degenarate reactions, thus driving the ring-opening equilibrium toward the right. Some novel N-cycloalkenyl-1 , 2-dihydropyr idines were synthesized for this purpose and their thermal behavior studied. However, they preferably underwent oxidative fragmentation to the corresponding pyridine bases and no products arising out of their ring-opening was observed. Similar failures were also recorded for a 4-vinyl-l , 2-dihydropyridine precursor. 1 x

PAGE 10

CHAPTER I GENERAL INTRODUCTION 1 • 1 • Modification of Reactivity by Heteroatoms In general, saturated hydrocarbons are very inert molecules. They show little tendency to take part in most types of chemical reaction. Thus they do not behave as an electrophile nor can they be used as nucleophiles in a reaction. However, introduction of a heteroatom in an all-carbon framework dramatically alters the situation. By including a nitrogen atom (as in amines) or an oxygen atom (as in alcohols) many transformations can now be achieved at the carbon atoms as well as on the heteroatom. This is mainly due to the large electronegativity difference between carbon and the heteroatoms which induces considerable polarization in such carbon-heteroatom bonds. Unsaturated carbon skeletons like alkenes and alkynes are more reactive than their saturated counterparts. By 1

PAGE 11

2 virtue of their loosely bound n-electrons they undergo reactions with electrophiles and again, introduction of a heteroatom like nitrogen next to a n-bond greatly enhances their electrophilicity . The effect of the lone-pair on nitrogen is con jugatively transmitted to the n-electrons thus increasing their nucleophilic character which is well documented in enamine chemistry [82T1975, 82T3363]. 1.2. Amines The reactivity of the a-carbons in amines is dominated by nucleophilic attack at that center [68MI1] which shows that amines normally prefer a electropositive center next to the nitrogen. However, being more electronegative than carbon, the nitrogen atom in amines polarizes the C-N bonds to induce a partial positive character on the adjacent carbon, thus rendering the a-hydrogens somewhat acidic. This, in turn, opens up the possibility of generating a carbanionic center next to the nitrogen and hence to the prospect of electrophilic attack at that center. This type of reactivity umpolung at the a-carbon of amines was forseen by Seebach [ 79AG( E ) 239 , 77S357 ] and is most desirable since a-functionalization by nucleophilic attack has many limitations. For example, it would take a multistep sequence to construct a g-hydroxy amine skeleton which, instead, can be easily prepared in one step via the reaction of an a-amino carbanion with a suitable electrophile (Scheme 1.1).

PAGE 12

3 \ R / \ N-CH--CH 2 I / N-CH 2 + RCHO Scheme 1.1 In practice, generation of a-amino carbanions proved to be a rather difficult task, especially for primary and secondary amines. In these cases the amino groups contain more acidic hydrogens and require protection during the a-deprotonation step. Here the principal problem lies in the choice of a suitable protecting group which should be easily introduced and subsequently removed as well as be inert to the deprotonating agent. However, with time, it became well recognized that the a-protons in amines are not acidic enough to be abstracted, even by strong bases and therefore needed additional activation, namely a-activation . This has been achieved in various ways, most commonly through the choice of an a-activating protecting group [84CRV471], Yet another problem associated with a-amino carbanions is that, once generated, they face strong repulsion from the adjacent lone-pair on the nitrogen and their stability is greatly reduced. As a result, many a-amino carbanions show tendencies to undergo rearrangements whereby the negative charge is tranferrerd to the nitrogen atom, leading to a thermodynamically more stable situation.

PAGE 13

4 Since a-functionalization in amines has been a matter of long-standing interest, research in the generation and utilization of a-amino carbanions continues. Now that generation of a-amino carbanions has been successfully demonstrated, greater stress has been applied to find novel yet simpler routes to these species. 1.3. Alcohols Since oxygen is more electronegative than carbon, the oxygen atom enjoys the greater share of the bonding electrons in a C-0 bond. Hence the a-carbon atom in an alcohol attains some electropositive character. Being more electronegative than nitrogen, the degree of polarization in hydroxy derivatives is much more than in amines and nucleophilic attack on the a-carbon atom is much more facile in oxygenated compounds [ 7 1MI 1 ] . On the other hand, inductive electron-withdrawal by oxygen makes the a-hydrogens acidic so that generation of an a-oxy carbanion is feasible. However, the generation of a-oxy carbanions faces similar problems as encountered with a-amino carbanions. Thus, the more acidic hydroxy proton needs prior protection and although the kinetic acidities of a-hydroxy protons are slightly higher than in amines, the process of a-activation is again necessary. The stability of a-oxy carbanions is even lower and they are very prone to 1,2-anionic migrations, most notably the Wittig rearrangement [ 70AG( E ) 763 ] .

PAGE 14

5 As a demonstration of reactivity umpolung at the a-carbon of alcohols, a-oxy carbanions has been successfully generated and utilized to construct key segments which otherwise require lengthy procedures. However, in contrast to the amines, reactivity umpolung at the a-carbon of hydroxy compounds has found less prominence since several other synthetic sequences are available which formally implement the concept. Nevertheless, a-oxy carbanions are synthetically important intermediates and are attractive targets for synthetic chemists. 1.4. Vinylic Amines Pericyclic reactions are one of the most interesting classes of reactions associated with alkenes and dienes; of these, the thermal electrocyclic reactions of polyenes have created much interest [80MI1] both from the mechanistic and the synthetic viewpoints. Cyclic dienes have attracted considerable attention as substrates for thermal electrocyclic reactions. Suitable 1 , 3-cyclohexadiene moieties undergo 6 n cycloreversion to the corresponding 1 , 3 , 5-hexatriene system (Scheme 1.2). Although this transformation is considered an equilibrium, the forward reaction can rarely be used preparatively due to unfavorable energetics [ 80MI2 ] .

PAGE 15

6 R Scheme 1.2 The formation of 1 , 3-cyclohexadiene from cis-1 , 3 , 5-hexatriene has an activation enthalpy of 29.0 Kcal/mole [64JCS3080]. The heats of formation of 1 , 3-cyclohexadiene and cis -1, 3, 5-hexatriene are reported as 40.6 and 25.4 Kcal/mole, respectively [69CC833]. These figures establish the AH for the ring opening of cyclohexadi ene at 44.2 Kcal/mole which means that very high temperatures would be necessary to achieve ring— opening of 1 , 3-cyclohexadiene type systems. Thus, in contrast to numerous instances of elect rocycl i zations of 1 , 3 , 5-hexa t r i ene s , one finds only a few examples of cycloreversion in 1 , 3-cyclohexadiene systems. In 1 , 3-cyclohexadiene , substitution of a methylene group by a heteroatom such as nitrogen transforms it to a 1 , 2-dihydropyr idine which can be looked upon as a dienamine. Both enamines and dienamines are more polarized than their parent alkenes. Although they participate in pericyclic reactions, the reactions are mostly polar in nature and quite often occur in stepwise fashion. Thus it can be

PAGE 16

7 expected that the diene moiety in 1 , 2-dihydropyr idinies would be perturbed by the nitrogen so as to aid in its cycloreversion. Experimental evidence shows this not to be of considerable extent and 1 , 2-dihydropyr idines are much more stable than the 1-azatrienes [80MI3]. However, in contrast to 1 , 3-cyclohexadienes , examples of ring-opening of 1 , 2-dihydropyr idines are more numerous and the process is indeed facile in those cases where the resulting azatrienes enjoy extensive charge delocalization. This has been elegantly shown by Maas [85AG(E)511] (Scheme 1.3). HC (CN) 2 5 0 C , 2 hr 2 2 OTf " N CH(CN) OTf' Me 2 N^+ NMe 2 Me Me -/ CN 'CN 2 N Scheme 1 . 3

PAGE 17

8 The above reaction, together with several examples of the ANRORC (Addition of Nucleophile — Ring Opening — Ring Closure) type reactions of pyridinium salts [80S589] show that, given the proper scenerio, 1 , 2-dihydropyr idines are much easier to ring-open than 1 , 3-cyclohexadienes . 1.5. Preview of the Work In Chapter II of this dissertation, a novel methodology for the generation of a-carbanions from secondary amines is described. In that section, stress has been given to the role of the protecting group; also, the concept of transient a-activation via anionic protection has been introduced. The same concept was next applied to the generation of a-hydroxv v-arbanions (Chapter III) which in the process also saw the utilization of the trimethylsilyl group as an easily removable a-activating auxilliary. In Chapter IV, thermal electrocyclic reactions of a cyclic dienamine, a 1 , 2-dihydropyr idine were evaluated. The synthesis and thermal behavior of some novel N-cycloalkenyl-1 , 2-dihydropyr idines have been investigated and their results are presented.

PAGE 18

CHAPTER II a-AMINOCARBANIONS 2.1. Introduction In amines, the inductive electron withdrawal by nitrogen generates a small positive charge on the a-carbon, rendering the a-hydrogen ( s ) acidic. As a result, the generation of a-aminocarbanions becomes feasible and leads to the prospect of electrophilic substitution at the a-carbon of amines. However, the generation of a-aminocarbanions has several associated problems. Primary and secondary amines possess more acidic hydrogen(s) on the heteroatom and, in these cases, generation of a-aminocarbanions would require prior protection of the amino group. Furthermore, the a-aminocarbanions, once generated, experience strong repulsion from the adjacent lone-pairs on nitrogen, thus destabilizing them. 9

PAGE 19

10 Chemical transformations at the a-carbon of an amine is a matter of long-standing interest and in this regard, a-aminocarbanions were expected to play a powerful role. Thus in the last decade much research has been devoted to the generation of a-aminocarbanions, as a result of which several a-aminocarbanion synthons are now available. These synthons have been critically assesed and several reviews have appeared on these and related topics [ 7 5AG ( E ) 1 5 , 77S 3 57 , 7 8CRV2 7 5,7 9 AG (E)239,80T2531, 8 4CRV4 7 1 , 8 5MI 1 ] . 2.1.1. Indirect Routes to a-Aminocarbanions Some a-aminocarbanions can be generated as their lithio salts without using an a-deprotonati on step in the sequence. These methods are termed here as "indirect methods," which utilize either trans-metallation (usually tin-lithium exchange) or lithium-metalloid exchanges [ 80T2531 ] . In the latter class, the name "metalloid" implies elements such as S, Se, and the heavier members of the halogen family. Tin-lithium exchange has been successfully utilized to generate a-lithio-tertiary amines. Thus the readily-available dialkylaminomethylstannanes (2.1), on treatment with n-butyl 1 i thium , gave rise to the di alkyl am inomethyl lithium reagents (2.2). The latter upon reaction with different electrophiles produced the expected

PAGE 20

11 a-alkylated amines (2.3) (Scheme 2.1) [86TL2361, 84MI1, 71JA4027]. That this type of t ransmetal lati on produces a free organoli thium (e.g. 2.2), and not any pentavalent "tin-ate" complex, can be nicely shown by 1 '*'^Sn-NMR studies [ 8 5TL1 1 4 1 , 8 6 JA2 10 2] . BuLi E + R 2 N-CH 2 -SnBu 3 > R 2 N-CH 2 _Li > R 2 N-CH 2~ E (21 ) (2.2) (2.3) Scheme 2.1 Although quite rewarding in its own merit, this transmetallation route suffers from a few drawbacks. The process fails completely when the a-carbon is either secondary or tertiary in nature [84MI1]. Similar transmetallations for a-stannyl-pr imary and -secondary amines are yet to be reported [85TL1141], Lithium-metalloid exchange may not be a wise choice for generating a-lithioamines . The starting materials (e.g. a-haloamines ) needed for this purpose are rather unstable and are often difficult to prepare. 2*1-2. Protection-Activation Route to a— Aminocarbanions This section deals with the more direct approach toward a— ami noca rbani ons , formed as their lithio salts. We would focus our attention on the a-lithio secondary amines where this methodology has found considerable success. In the

PAGE 21

12 protection-activation" route, the a-lithioamines are generated by direct abstraction of an a-hydrogen using a strong lithiating base, in order to do so, the more acidic amino hydrogen needs protection. The protecting groups are so chosen that they are capable of performing two functions simultaneously. Firstly, the more acidic amino proton is protected from reacting with the lithiating base. Secondly, they play a major role in activating the a— position toward hydrogen abstraction. This process of a-activation is necessary since the a-hydrogens of amines are not acidic enough to be abstracted by strong bases. If "G" be such a hypothetical protecting-acti vating group, a-alkylation of a secondary amine can be visualized to occur as depicted in Scheme 2.2. + G-X R"Li + > RN CH-R' I \ G* * 'Li H H G ( 2 . 4 ) ( 2 . 5 ) ( 2 . 6 ) E + G RN CH-R' I I I I H E ( 2 . 8 ) < + RN CH-R' + H G E (2.7) Scheme 2.2

PAGE 22

13 The key step in the sequence in Scheme 2.2 is the formation of the a-lithio species (2.6) from 2.5, since it involves the process of a-activation . There are several possibilities by which the protecting group (in 2.5) could activate the a-position. According to Beak and Meyers [86ACR356], prior complexation of lithium (from the lithiating base) with "G" would activate the lithiating agent as well as direct the latter on to the a-carbon via a rigid chelate. In doing so, a partial dipole might also be created resulting in increased kinetic acidities of the a-protons (Figure 2.1). RN CH-R' RN + CH-R' Figure 2.1. a-Activation of Secondary Amines In all probability the processes of prior complexation and dipolar activation would work in unison and complement each other. The a-lithio species once formed owes its stability to a great extent on the co-ordinating power of the protecting group so as to (partly) satisfy the co-ordination sites for lithium. For an all round success,

PAGE 23

14 the protecting group itself should not contain any acidic hydrogen, should be inert to the raetallating agent and, of course, be easily introduced and removed. Based on these concepts several protecting-activating groups have been successfully employed in generating a-lithio secondary amines. Some of the versatile a-lithio secondary amine precursors are shown in Figure 2.2. A comprehensive picture can be obtained from Beak's recent review on this subject [84CRV471]. -N-CH^R' 1 2 R-N-CH-R' 1 z R-N-CH 0 R N=0 C=0 1 1 ^ C=0 (2.9) C(Et) 3 C(Ph) 3 (2.10) (2.11) “ j N 2 k-n-ch 2 R' P ( NM e 2 ) 2 HC 0 N N-CMe 3 ( 2 12 ) (2.13) Figure 2.2. g-Lithio Secondary Amine Precursors For the N-nitrosoamines (2.9) [ 7 5AG ( E ) 1 5 ] , resonance stabilization of the a-carbanion is thought to be the chief source of a-activation [78CRV275]. Lithium di i sopropylamide (LDA) has been used as the lithiating agent. The remaining examples in Figure 2.2 operate on the "prior complexation-

PAGE 24

15 activation" concept [86ACR356] and stronger metallating agents like alkylli thiums were necessary for their success. However, alkyllithiums are strong nucleophiles as well and the use of such strongly nucleophilic agents warrants that the protected center be sufficiently hindered so as to prevent attack at that position. Thus the sterically crowded triethylacet amides (2.10) [84JA1010], and triphenylacetamides (2.11) [83T1963] have been chosen to protect their carbonyl centers from any nucleophilic attack. By comparison, phosphonamides (2.12) [83T1963] and formamidines (2.13) [85MI1] are relatively inert to attacks by organolithiums . At present, the N-nitrosoamines (2.9) and the formamidines (2.13) stand out as the most useful precurssors to a-lithio secondary amines. Both such systems encourage . . , . 0 0 0 a-lithiation on 1 , 2 and 3 carbons and alkylation has been achieved with a wide variety of electrophiles. However, the N-nitrosoamines are notoriously toxic and hazardous. The formamidines, on the other hand, are easy to prepare, nontoxic and can be deprotected (by hydrazinolysis or LiAlH^ reduction) to liberate the a-alkylated secondary amines. In addition, chiral formamidines derived from secondary amines served as substrates for asymmetric a-alkylation and have

PAGE 25

16 been utilized for the synthesis of various alkaloids in high (90-98%) enantiomeric excesses [87JA1263, 87JA1265, 86JOC3108, 86JOC3076]. The a-aminocarbanion precursors 2.102.12 require very harsh deprotection steps and are thus not of much practical use. 2.1.3. Carbon Dioxide as a Protecting-Activating Group As just described, a-alkylation of secondary amines can be achieved in a three-step sequence: protection, a-lithiation and electrophile addition followed by deprotection (Scheme 2.2); the principal focus being in the a-lithiation step. In order to accomplish this, special protecting groups have been employed. In some cases (e.g. 2.10 and 2.11), preparation of the protecting group itself involved multistep synthesis while in others, deprotection needed extreme conditions (e.g. 2.11 requires dissolving metal reduction) thereby limiting their synthetic utility. An additional problem encountered in the generation of a-aminocarbanions is the 1,2-migration of the N-protecting group. This type of anionic migration is extremely facile since it involves an entropically favored thr ee-membe r ed transition state. The equilibrium in Scheme 2.3 demonstrates such a rearrangement [76CC339].

PAGE 26

17 Ph \ yC N-R Ph VC=0 I OEt Ph N-R /l Ph C=0 I OEt Scheme 2 . 3 Another example of such 1,2-anionic migration is found in the case of N-ni t roso ( 2-me thylamino ) acetoni t r i le ( 2 . 14 ) (Scheme 2.4). In basic methanol, 2.14 gives rise to the a-carbanion (2.15) which triggers off a facile nitroso-transfer , ultimately leading to the amidoxime (2.17) [62HCA2426, 64HCA33, 78JOC3617]. Similar nitroso-transfer has also been observed during the a-lithiation of N( t-butyl )-N-( nit roso ) trimethyl si lylmethyl amine [ 77CB18 52 ] . base NC— CH.,— N— Me 1 I N=0 (2.14) > NC-CH-N-Me ^ |S» MeOH ^*N=0 (2.15) NC-C-NH-Me < N-OH (2.17) NC-CH-N-Me I “ N=0 (2.16) Scheme 2 . 4

PAGE 27

18 This type of migration of the N-protecting group is aided by the fact that the nitrogen, in most cases, is attached to an electrophilic center (either a carbonyl for the amides or a nitroso for the N-ni trosoamines ) . On the other hand, such electrophilic centers on nitrogen is essential both for a-activation and easy deprotection. This requirement withholding, a decrease in the electrophilici ty of the N-bound center appeared to be a viable solution to the 1,2-migration problem. An anionically protected amine, e.g. (2.18) was thus considered as the remedy (Figure 2.3). ^N-C ^ 0 Ns X _ Li + (2.18) Figure 2.3. Anionically Protected Secondary Amine It was envisaged that the negative charge at the protected end of 2.18 would decrease the electrophilicity of the carbonyl group and hence, the N — > C migration could be minimized. Furthermore, the N-bound center in 2.18 remains electronically protected from nucleophilic attack by the lithiating agent. The process of ex-activation could still be operative through prior complexation of the lithiating agent

PAGE 28

19 either to "X" or to the carbonyl oxygen. The biggest advantage of using 2.18 lies in the prospect of its facile deprotection. This should occur readily since the protecting group would be lost as a neutral molecule (0=C=X). The proposed synthon (2.18) could possibly be made via the reaction of an N-lithio amine with a suitable cumulene of general stucture 0=C=X (Scheme 2.5). In choosing the appropriate cumulene, preference should be given to those in which X is a heteroatom so that the negative charge in 2.18 is better accomodated. Also, a heteroatom at that position may be neccessary for the process of a-activation through prior complexation with the lithiating agent. R \ N-Li CH~ / 2 R ' 0 II C II X > R \ / N — C S 'x _ Li + .CH-, (2.18) Scheme 2 . 5 One such cumulene which meets all these requirements is carbon dioxide. It is readily available, relatively non-toxic and has the proven ability as an electrophile. A thorough search of the literature revealed that various N-lithio derivatives of secondary amines react with C0 o to

PAGE 29

20 form the corresponding lithium carbamates [87H1333], These salts, although stable in the absence of moisture, readily loose CC >2 on mild hydrolysis or simple heating [87H1333 ]. This latter property, we thought, could be efficiently exploited to achieve a very mild deprotection step. Given the hope that a-lithiation of lithium carbamates could be achieved, a novel synthon for o-aminocarbanions was in the offering . In a preliminary study carried out by previous workers in this group, CO 2 has shown considerable promise as a protecting-acti vating group for a-lithiation of a primary amine. Benzylamine (2.19) was investigated as a model compound which formed its lithium carbamate (2.20) on being treated with n-butyllithium (1 equivalent) followed by CC >2 . Treatment of 2.20 with two equivalents of t-butyllithium gave rise to the a-lithio species (2.21) which was subsequently trapped with various electrophiles (D-,0, CO-,, Mel, PhC0 2 Et, Bu t NCO, PhNCO) providing the a-substi tuted carbamates (2.22) (Scheme 2.6). The latter, without the need of isolation, was readily deprotected under acidic work-up to give the a-substi tuted benzylamines (2.23) in good yields [ 8 7 S 4 1 5 ] . As an added advantage, the whole set of operations can be executed in the same reaction vessel.

PAGE 30

21 1) Bu n Li 2 Bu fc Li PhCH 2 NH 2 > PhCH 2 NHC0 2 Li 2) C0 2 THF/— 2 0 0 C ( 2 19 ) ( 2 . 20 ) H + /H 2 0 PhCH(E)NH 2 < co 2 ( 2 . 23 ) Scheme 2.6 The directing power of the lithium carbamate ( 2 . 20 ) can be well appreciated if one considers that monoand disubstituted benzylamines undergo exclusive ortho-lithiation in the phenyl ring [ 71JOC1607 , 63JOC663 , 63JOC3461 , 67JOC1479 ] (Scheme 2.7). For ortho -lithiation to occur in 2.20, a seven-membe red chelate would be neccesary; instead a more rigid f ive-membered chelate in 2.21 directs the lithiation at the benzylic carbon. PhCH-N=C' -OLi Li ( 2 . 21 ) \ OLi PhCH-N=C' .OLi \ ( 2 . 22 ) OLi 2 BuLi Scheme 2 . 7

PAGE 31

22 Previous workers have also studied the lithium carbamate route for a-lithiation of a secondary amine, namely tetrahydroisoquinoline (2.24). Most grati fyingly , the derived lithium carbamate (2.25) underwent facile lithiation at the 1-position to produce the a-lithio species (2.26). The latter upon reaction with different electrophiles followed by acidic work-up directly afforded the 1-substi tuted tetrahydroisoquinolines (2.27) (Scheme 2.8) [ 86T2571 ] . Again, the whole sequence can easily be carried out in the same reaction vessel. (2.24) (2.25) Bu THF/-2 0 0 C (2.27) (2.26) Scheme 2.8

PAGE 32

23 With these preliminary results in hand, a detailed study of this novel a-lithiation process was most desirable. 2.1.4. Aims of the Work The aim of the present investigation was to explore various other lithium carbamates as a— aminocarbanion synthons, in particular, those derived from unactivated secondary amines. a-Lithiation of benzylamine and tetrahydroisoquinoline via their lithium carbamates can be considered as examples where the respective a-hydrogens are fairly activated, being benzylic in nature. Therefore, the role of C0 2 in the a-activation step could not be critically assessed. Hence it was of interest to study the a-lithiation of those lithium carbamates where any kind of inductive activation of the a-hydrogens is absent. The secondary amines chosen for the study were 1 , 2 , 3 , 4-tetrahydroquinol ine and indoline. a— ki thi a t i on of tetrahydroquinoline and indoline has previously been achieved only via their formamidines [81TL5119]. In that study, the respective a-lithio species were trapped with alkyl halides (Mel, Bui) and benzaldehyde to give the a-alkylated formamidines in good yields. However, deprotection of the latter required rather harsh

PAGE 33

24 conditions such as refluxing with aqueous KOH or hydrazine, especially for the benzaldehyde adducts which could only be deprotected by LiAlH^ in refluxing THF . In view of the mild deprotection procedure for the lithium carbamates, succesful a-lithiation of the tetrahydroquinoline and indoline via their lithium carbamates appeared as a potentially superior method for the a-functionalization of these amines. 2.2. Results and Discussion 2.2.1. g-Metallation Studies on Lithium 1,2,3, 4-Tetrahydroquinolinyl Carbamate 1 , 2 , 3 , 4-Tetrahydroquinoline ( 2 . 28 ) was easily converted to its lithium carbamate ( 2 . 29 ) by sequential addition of n-butyllithium followed by CC> 2 gas (Scheme 2.9). In order to achieve the desired a-lithiation, 2.29 was exposed to t-butyllithium in tetrahydrofuran . After 2 hrs at -20°C, £-tolualdehyde was added as the electrophile. However, after acidic work— up with 2N HCl, formation of the desired a-adduct 2.31 was not observed; the starting materials were recovered unchanged. Thus it was evident that the a-lithio species 2.30 was not generated under these conditions.

PAGE 34

25 (2.28) 1) Bu n Li 2) C0 2 (2.29) Me 1) £-MeCgH 4 CHO < 2) H + /H 2 0 Bu t Li // I C0 2 Li (2.30) Scheme 2.9 In order to achieve the desired a-lithiation, 2.29 was subjected to various other reaction conditions. It was found that solvent had no effect on the lithiation process. Thus the reaction failed when carried out in ether as well as in hexane (heterogeneous system). Complexing agents [86TL331] like hexamethyl phosphor amide ( HMPA ) or tetramethylethylenediamine (TMEDA) also had no effect on the reaction and in each case, the starting materials were recovered unchanged. These failures, however, shed some light on the role of C0 2 in the a-lithiation step: the lithium carbamate 2.29 had no perceptible activation toward

PAGE 35

26 a-lithiation. At this stage it was felt that the source of a-a ctivation for lithium carbamates lies somewhere else, probably in the use of a stronger metallating agent. This received some support from the recent studies by Schlosser [84TL741] on the superior metallating power of alkyllithiums in an admix with potassium alkoxides. More pertinent to the present study was the report by Ahlbrecht [84TL1353] who described the successful a-metallation of N-methyl tertiary amines using a mixture of potassium t-butoxide and s^-butyllithium, otherwise a poor yielding process with alkyllithiums alone. On the basis of these reports, we proceeded to reinvestigate the a-metallation of the lithium carbamate 2.29 using Schlosser's mixed base system. The lithium carbamate 2.29 was generated as described before in Scheme 2.9. When 2.29 was treated at -78°C with an equimolar mixture of t-butyl 1 i thium and potassium t-butoxide in tetrahydrofuran, a wine-red solution was immediately formed. Addition of benzaldehyde discharged this color and after stirring overnight at room temperature, the reaction was worked-up with 2N HCl . This time the crude product mixture afforded the desired a-adduct 2.33 in 55% yield as a mixture of diastereomers (Scheme 2.10).

PAGE 36

27 C0 2 Li (2.29) Bu t Li/KOBu t > THF , 50 °C I C0 2 Li (2.32) 1 ) PhCHO 2) 2N HCl V (2.33) (55%) ( erythro : threo = 40:60) Scheme 2.10 Optimization of the reaction conditions gave the best results when t-butyllithium was used as the alkyllithium component and the metallation being carried out at -50°C for 2 hrs. The reaction worked best in tet rahydrof uran since the lithium carbamate 2.29 was found to be insoluble in ether solvents such as hexane or ether. The diastereomers of 2.33 were separated by careful column chromatography (silica gel) to give a 40:60 ratio of

PAGE 37

28 e rythro : threo isomers and were characterized individually by 1 13 their H and C-NMR spectra. Their stereochemical assignments were based, tentatively, on the nature of the a-hydroxy proton in the 1 H-NMR spectra (200 MHz) (Fig. 2.4) and were in agreement with the trend which is observed for similar a-adducts of tetrahydroquinol ine [81TL5119] and piperidine [79TL771, 85JOC1019]. According to these reports the a-hydroxy proton of the erythro -adducts usually appears at a more downfield region and with a lower coupling constant than that of the threo -i some r s . Thus, in our hands, the product-isomer whose a-hydroxy proton appeared at 6 4.97 (d, J 6 Hz) was assigned to erythro -2 . 33 ; the corresponding proton for the other isomer was found at 6 4.97 (d, J 10 Hz) and was assigned to threo -2 . 33 . In their C-NMR spectra the a-hydroxy carbons were found at 6 75.1 and 76.0 wheras the a-amino carbons appeared at 6 41.6 and 43.0, respectively, for threo and e ry thr o -2 . 33 . It should, however, be noted that both th r eo -2 . 3 3 (m.pt. 147°-149°C, from benzene pet. ether) and erythro -2 . 33 (m.pt. 105°-107°C, from ether pet. ether) were isolated as crystalline solids. This is in contrast to the only previous report [81TL5119] of their preparation where they were described as the hydrochloride salt (m.pt. 138°-141°C, from ether ethanol) of the diastereomer ic mixture.

PAGE 38

29 I Figure 2.4. 1 H-NMR Spectra of the threo and erythro-2.33 A = threo ; B = e r y t h r o

PAGE 39

30 The a-metallo species 2 . 32 , generated as in Scheme 2.10, reacted with other carbonyl electrophiles such as cyclohexanone and benzophenone to give the corresponding adducts 2.34 and 2.35 in 52 and 55% yield, respectively (Scheme 2.11). These novel adducts were each characterized by their 1 H and 13 C-NMR spectra as well as elemental (C,H,N) analyses . Scheme 2.11

PAGE 40

31 Our attempts at the a-functionalization of t e t r ahyd r oqui nol i ne via 2.32 had its share of failures also. Of particular significance are the failures when diphenyl diselenide or dimethyl formamide were used as electrophiles. These electrophiles are known to be highly lithiophilic and the failure in their reaction with 2.32 speaks in favor of formation of an a-potassio species, rather than an oc-lithio species. Reaction of 2.32 with benzoyl chloride, methyl ch 1 o r o f o r ma t e or ethyl jo— toluate produced complex product mixtures from which only a small amount of the desired products could be detected in their 1 H-NMR spectra. Interestingly, when diphenylcarbamoyl chloride was used as the electrophile, it produced 2.36 (35%), the product of ca rbamoy 1 a t i on . This latter result indicates partial cleavage of the N-protecting group in 2.29 during the a-metallation step which might account for the above failures as well as for the modest yields of the products 2.33-2.35 obtained so far. 2.36

PAGE 41

32 Similar results were obtained when methyl iodide was used as the electrophile. Although in the ^"H-NMR spectrum of the crude product the desired 2-methyltetrahydroquinoline could be observed, it was inseparable from the starting material due to very close R ^ values. What could be separated by chromatography was shown to be N-methyltetrahydroquinoline (22%), the N-alkylation product. 2.2.2 . g-Metal lation Studies on Lithium Indolinyl Carbamate Similar to the case of tetrahydroquinoline , the lithium carbamate ( 2 . 38 ) of indoline was prepared by sequential addition of n-butyl 1 i thium to the amine followed by CC^ gas. As was expected, 2.38 did not undergo a-lithiation with t-butyllithium alone under a variety of reaction conditions. However, when 2.38 was treated at -78°C with an equimolar mixture of t-butyllithium and potassium t-butoxide in tetrahydrofuran , formation of the characteristic wine-red solution of the a-metallo species 2.39 was observed. The color was discharged upon addition of jotolualdehyde and acidic work-up (with 2N HCl ) after 12 hrs gave the a-adduct 2.40 in 52% yield (Scheme 2.12). The incipient a-metallo species 2.39 could also be trapped with cyclohexanone to give the adduct 2.41 in 57% yield (Scheme 2.12).

PAGE 42

33 (2.38) Bu t Li/KOBu t THF , -50 °C V (2.40) (52%) ( erythro : threo = 55:45) R p Me 1 ) RCHO < 2) 2N HCl 1 ) 2 ) Cf° 2N HCl Scheme 2.12 The a-adduct 2.40 was obtained as a mixture of diastereomers . They were separated by careful column chromatography over silica-gel to give a 55:45 ratio of erythro : threo isomers. The stereochemical assignments were again based on the chemical shift and coupling constant of

PAGE 43

34 their a-hydroxy proton in the ''H-NMR spectra. In accordance to the literature data [81TL5119], the isomer whose a-hydroxy proton appeared at S 4.85 (d, J 6 Hz) was assigned to erythro -2 . 40 . Accordingly, the a-hydroxy proton for threo -2 . 40 appeared at a relatively upfield position, 6 4.6 and with a slightly larger coupling constant (J 8 Hz). In 1 3 their C-NMR spectra, both the isomers showed the requisite number of carbons, the a-hydroxy carbons being found at S 75.2 and 75.6, whereas the a-amino carbons appeared at S 48.6 and 49.3, respectively, for threo and e rythro -2 . 40 . Again it should be mentioned that the benzaldehyde a-adduct of indoline was previously reported [81TL5119] only as the hydrochloride salt of the diastereomer ic mixture. The novel cyclohexanone a-adduct 2.41 (m.pt. 108°-110°C, from ether hexane) was obtained in 57% yield after column chromatography of the crude product mixture. It was characterized by its 1 H and 13 C-NMR spectra as well as by elemental (C,H,N) analysis. Just like its tetrahydroquinoline counterpart, the a-metallo species 2.39 was also found to be inert to diphenyl diselenide and dimethyl formamide , whereas reactions with ethyl g-toluate and methyl chlorof ormate gave complex product mixtures.

PAGE 44

35 2.2.3. Role of Carbon Dioxide in the g-Metallation of Lithium Carbamates " In the previous two sections, a novel one— pot sequence for a-metallation of tetrahydroquinoline and indoline via their lithium carbamates was described. However, the principal question which remains unanswered is that concerning the role of lithium carbamates toward a-activation. In this section, we attempt to clarify this through some experimental observations. We begin with Ahlbrecht's results [84TL1353] where the N— methyl groups of some tertiary amines were successfully metallated with Schlosser's mixed base system. In these cases, the substrate did not posses any a— activating group on the nitrogen. However, it should be noted that the proton in question was a primary hydrogen which would be expected to be more acidic than those encountered in our study, i.e. secondary hydrogens. Unfortunately, Ahlbrecht's report [84TL1353] provides no information on attempted a-metallation of secondary carbons in tertiary amines. Unable to find a direct comparison, we set out to study the a-metallation of protected secondary amines where the protecting groups do not provide any kind of a— activation whatsoever. The substrate chosen for this purpose was the Nt r ime thy 1 s i lyl i ndol ine (2.42). The latter was prepared in s * *u f rom indoline ( n-butyll i thium chlorotrimethylsilane )

PAGE 45

36 and treated first with a mixture of t-butyllithium and potassium t^-butoxide followed by benzaldehyde as the electrophile (Scheme 2.13). However, after acidic work-up (with 2N HCl ) , none of the desired a-adduct (2.44) could be detected. Indoline and benzaldehyde were recovered unchanged . 1) Bu n Li 2) Me 3 SiCl (2.44) (2.43) Scheme 2.13 The above result, indirectly shows that lithium carbamates indeed have an a-activating effect. We believe that dipole creation (as in 2.45) is the major factor behind

PAGE 46

37 such a— activation resulting in increased acidity of the a— protons. However, the extent to which this happens must be small since it required a very strong base like Schlosser's mixed base system for proton abstraction. 2.45 2.3. Conclusions Tetrahydroquinoline and indoline, via their lithium carbamates underwent successful a-metallation when treated with a mixture of t-butyll i thium and potassium t-butoxide. The resulting a-metallo species were trapped with aldehydes and ketones to give the respective a-alkylated products in good yields, although alkyl halides and acyl chlorides failed in this respect. The nature of "M" in the a-metallo species 2.32 and 2.39 could not be ascertained at this stage. However, on the basis of Schlosser's [84TL741] and .Ahlbrecht's [84TL1353] accounts, it was tentatively assigned

PAGE 47

38 as potassium, although the results of attempted t r ansme tal la tion to the a— lithio species remain inconclusive; thus treatment of 2.32 with lithium, bromide , prior to the addition of benzaldehyde , led to much poorer yields of a-alkylation . The advantages of the present methodology include a very mild deprotection step and the convenience of a one-pot operation. These score over the only reported procedure [81TL5119] for a-alkylation of tetrahydroquinol ine and indoline which involved a conventional three-step sequence as well as very harsh deprotection conditions. 2.4. Experimental Melting points were determined on a Bristoline hot-stage microscope and are uncorrected. H and C-NMR spectra were recorded on a Varian XL200 spectrometer. Chemical shifts for the protons are reported in 5 (ppm) downfield from Me^Si whereas those for the carbons were determined with reference to the solvent peaks (CDC1 3 , 6 77.0 or DMSO-dg , 6 39.5). The nature of the carbon signals in the 13 C-NMR spectra were determined by the Attached Proton Test [81CC150, 82JMR535]. Mass spectra were recorded on a AEI MS instrument attached to a DS 55 database. Elemental analyses (C,H,N) were performed by Dr. Roy W. King of this department. All

PAGE 48

39 manipulations involving organolithium compounds were carried out in oven-dried (120°C, overnight) apparatus under a slight positive pressure of dry argon. Transferr ring operations were done either by syringe techniques or via canula . Tetrahydrofuran ( THF ) and ether were freshly distilled from sodium-benzophenone ketyl. Chlorotrimethylsilane , hexamethylphosphoramide (HMPA) and tetramethylethylenediamine (TMEDA) were distilled from calcium hydride under reduced pressure and stored over molecular sieves. Petroleum ether (36°-60°C), hexane and triethylamine were distilled from calcium hydride and stored over potassium hydroxide. Column chromatography was carried out on MCB silica gel (230-400 mesh) and the solvents used for chromatography were routinely distilled prior to use. General Procedure for a-Alkylation of Indoline an d Tetrahydroquinoline n-Butyll i thium (2.5 M in hexanes, 2 ml, 5 mmole) was added dropwise at -78°C, under argon, to a solution of the amine (5 mmole) in dry THF (35 ml) in a 250 ml Schlenk type reactor. It was then slowly warmed to room temperature and dry CC >2 gas rapidly bubbled through it. The solution turned colorless (a colorless solution at this stage is vital for ultimate success; otherwise, poor yields were obtained). All volatile contents were then removed under 0.5 mm leaving a white powdery residue of the lithium carbamates. A preheated

PAGE 49

40 (140°C, 0.5 mm, 2 hrs) sample of potassium t-butoxide (0.57 g, 5 mmole) was then quickly added, under argon, to the reactor. After redissolving in dry THF (40 ml), t.— butyl 1 i thium (1.7 M in pentane, 3 ml, 5 mmole) was slowly added at -78°C. The resulting wine-red solution was stirred for 2 hrs at -50°C and the appropriate electrophile was then added at -78°C. Stirring was continued at room temperature for 12 hrs; the solvent was then removed at 0.5 mm and the residue treated with hydrochloric acid (2N, 10 ml). After gas evolution ceased, it was basified with solid sodium carbonate and was extracted with ethyl acetate or methylene chloride ( 3 x 20 ml). The combined organic layer was washed with brine and dried (Na 2 SO ( j). Removal of solvent followed by column chromatography (silica-gel, 2-15% gradient of ethyl acetate in hexane containing 10-15% tr iethylamine ) gave the products which were characterized according to the individual cases described below: 1-Phenyl-1-(1,2,3, 4-tetrahydroquinolin-2-yl )carbinol ( 2 . 33 ). Yield 55%; e rythro : threo = 4 0:60. threo ( 2 . 33 ) : R ^ 0.5 (silica-gel, 10% ethyl acetate in hexane containing 10% tr iethylamine ) ; m.p. 147-149°C (needles from benzene-pet . ethe r ) ; Found: C, 80.52, H, 7.02, N, 5.80. C 16 H 17 NO requires C, 80.30, H, 7.16, N, 5.85%. § H

PAGE 50

41 (CDCl 3 /DMS0-d 6 /D 2 0) 1.6(2H, m), 3.1 (3H, m), 4.65 (1H, d, J 10 Hz), 6. 4-6. 6 ( 2H , m) , 7.0 (1H, m) and 7. 2-7. 5 (6H, m) . 22.0 (CH 2 ) , 36.6 (CH 2 ), 41.6 (CH-N), 75.1 (CH-OH), 112.4 (aryl CH), 113.7 (aryl CH), 119.6 (aryl C) , 125.6 (aryl CH ) , 125.7 (aryl CH ) , 125.8 (aryl CH ) , 126.7 (aryl CH ) , 130.1 (aryl CH ) , 142.8 (aryl C) and 143.6 (aryl C) . e rythro (2.33) : R ^ 0.4 (silica-gel, 10% ethyl acetate in hexane containing 10% tr iethylamine ) ; ra.p. 105-107°C (needles from ether-pet . ether ) ; Found: C, 80.30, H, 7.31, N, 5.55. C lg H 17 NO requires C, 80.30, H, 7.16, N, 5.85%. S R (CDC1 3 /D 2 0) 1.65 (1H, m), 2.2 (1H, m) , 3.0 (1H, m) , 3.2 (1H, m), 3.4 (1H, m), 4.97 (1H, d, J 6 Hz), 6. 4-6. 7 (3H, m), 6.9 (1H, m) and 7. 2-7. 3 (5H, m) . 21.7 (CH 2 ), 39.2 (CHj), 43.0 (CH-N), 76.0 (CH-OH), 114.3 (aryl CH ) , 116.7 (aryl CH), 120.4 (aryl C ), 126.3 (aryl CH ) , 127.1 (aryl CH ) , 127.4 (aryl CH ) , 128.1 (aryl CH ) , 129.2 (aryl CH ) , 142.5 (aryl C) and 145.6 ( aryl C ) . 1( 1 , 2 , 3 , 4-Tetrahydroquinolin-2-yl ) cyclohexanol (2.34). Rj 0.5 (silica-gel, 12% ethyl acetate in hexane containing 10% triethylamine ) ; yield 52%; viscous oil; M + (HRMS) found: 231.1632. C^H^NO requires 231.1624. § H (CDC1 3 /D 2 0) 1.4-1. 7 ( 1 1 H , m), 2.2-2. 3 (1H, m), 2.7 (1H, m), 3.3 (1H, m), 3.55 (1H, m), 6. 4-6.6 (2H, m) and 6. 9-7.0 (2H, m). 6 C 22.0 (CH 2 ), 22.2 (CH 2 ), 25.8 (CH 2 ), 34.1 (CH 2 ), 37.6

PAGE 51

42 (CH 2 ), 39.2 (CH 2 ), 44.5 (CH-N), 73.8 (C-OH), 113.9 (aryl CH), 115.6 (aryl CH ) , 119.8 (aryl C) , 127.6 (aryl CH ) , 131.5 (aryl CH ) and 145.3 (aryl C) . l,l-Diphenyl-(l,2,3, 4-tetrahydroquinolin-2-yl )carbinol (2.35) . R f 0.5 (silica-gel, 10% ethyl acetate in hexane containing 10% tr iethylamine ) ; m.p. 128-129°C (plates from ether-hexane); Found: C, 83.81, H, 6.97, N, 4.09. C 22 H 21 NO requires C, 83.78, H, 6.71, N, 4.44%. (CDCl,/D-0) 1.8 (1H, m), 2.1 (1H, m), 3.0 (1H, m), 3.4 (1H, m), 4.05 (1H, m), 6.2-6.55 (3H, m), 6.95 (1H, m) and 7. 1-7. 8 (10H, m). 6 C 24.1 (CH 2 ), 39.6 (CH 2 ), 43.9 (CH-N), 81.5 (C-OH), 114.4 (aryl CH ) , 116.4 (aryl CH ) , 118.9 (aryl C) , 125.8 (aryl CH ) , 126.1 (aryl CH ) , 126.2 (aryl CH ) , 127.8 (aryl CH ) , 128.0 (aryl CH ) , 130.8 (aryl CH ) , 145.3 (aryl C) , 146.3 (aryl C) and 147.8 (aryl C ) . 1( Pi phenyl carbamoyl ) -1 , 2 , 3 , 4tetrahydroqui no line (2.36) 0.6 (silica-gel, 10% ethyl acetate in hexane containing 10% triethylamine ) ; yield 35%; m.p. 120-122°C (needles from ether-hexane); Found: C, 80.50, H, 6.23, N, 8.19. C 22 H 20 N 2 O requires C, 80.46, H, 6.14, N, 8.53%. S R (CDC1 3 ) 1.9 ( 2H , m), 2.5 (2H, m), 3.6 (2H, m) , 6. 8-7. 3 (13H,

PAGE 52

43 m) and 7.4 (lH, d, J 12 Hz). § c (CDClg) 23.4 (CH 2 ), 26.5 (CH 2 ), 45.4 (CH 2 ) , 122.5 (aryl CH), 123.3 (aryl CH ) , 124.8 (aryl CH ) , 125.1 (aryl CH ) , 126.0 (aryl CH ) , 128.1 (aryl CH), 128.7 (aryl CH), 129.8 (aryl C), 138.7 (aryl C) , 143.9 (aryl C) and 159.3 (C=0). 1[ ( 4 ' -methyl )phenyl]-l-(indolin-2-yl)carbinol (2.40). yield 52%; erythro : threo = 55:45. threo (2.40): R ^ 0.4 (silica-gel, 15% ethyl acetate in hexane containing 15% triethylamine ) ; m.p. 130-133°C (reprecipitated from chloroform-hexane); M + ( HRMS ) found: 239.1331. C lg H 17 NO requires 239.1310. & H (CDC1 3 /D 2 0) 2.3 (3H, s), 3. 0-3. 3 (2H, m), 3.55 (lH, m), 4.6 (lH, d, J 8 Hz), 6. 5-6. 7 ( 2H , m) and 7. 0-7. 4 (6H, m). 20.4 (CH 3 ), 48.6 (CH-N), 49.1 (CH 2 ), 75.2 (CH-OH), 108.9 (aryl CH ) , 117.6 (aryl CH ) , 125.4 (aryl CH ) , 126.0 (aryl CH ) , 126.9 (aryl CH), 128.0 (aryl CH ) , 129.9 (aryl C) , 136.0 (aryl C) , 139.9 (aryl C) and 151.3 (aryl C) . erythro (2.40): R f 0.3 (silica-gel, 15% ethyl acetate in hexane containing 15% triethylamine); m . p . 12 3-12 5 °C (reprecipitated from benzene-pet . ethe r ) ; M + ( HRMS ) found: 239.1316. C 16 H 17 NO requires 239.1310. § H (CDC1 3 /D 2 0) 2.35 (3H, m), 3. 5-3. 7 (3H, m), 4.85 (lH, d, J 6 Hz), 6.5-6.65 ( 2H , m) and 7. 0-7. 3 (6H, m) . § c 21.1 (CH 3 ), 49.3 (CH-N),

PAGE 53

44 49.7 (CH 2 ), 75.6 (CH-OH), 110.0 (aryl CH ) , 118.5 (aryl CH ) , 125.0 (aryl CH ) , 126.5 (aryl CH ) , 128.1 (aryl CH ) , 128.6 (aryl C) , 129.0 (aryl CH ) , 137.3 (aryl C) , 139.5 (aryl C) and 152.2 ( aryl C ) . l-(Indolin-2-yl) cyclohexanol (2.41). R f 0.5 (silica-gel, 15% ethyl acetate in hexane containing 15% triethylamine ) ; yield 57%; m.p. 108-110°C (needles from ether-hexane); Found: C, 77.36, H, 9.03, N, 6.14. C 14 H lg NO requires C, 77.38, H, 8.81, N, 6.45%. S R (CDC1 / D 2 0) 1.3-1. 8 (10H, m), 3.25 (1H, t, J 7 Hz), 3.6 (2H, d, J 7 Hz), 6. 6-6. 7 (2H, m) and 7 . 0-7 . 3 (2H, m). S c 21.6 (CH 2 ), 25.7 (CH 2 ), 33.8 (CH 2 ), 34.4 (CH 2 ), 48.8 (CH 2 ), 52.9 (CH-N), 73.6 (C-OH), 109.6 (aryl CH ) , 118.1 (aryl CH ) , 126.3 (aryl CH ) , 127.9 (aryl CH ) , 128.5 (aryl C) and 152.5 (aryl C) .

PAGE 54

CHAPTER III a-HYDROXYCARBANIONS 3.1. Introduction In alcohols, the inductive electron withdrawal by oxygen induces a polarization of the carbon-oxygen bond. As a result, the a-carbon bears a small positive charge, rendering the attached hydrogens acidic. This could allow the generation of an a-hydroxycarbanion , leading to the possibility of electrophilic substitution at the a-carbon atom of the alcohol. In practice, however, a-hydroxy carbanions are very difficult to generate. First, the more acidic hydroxyl proton needs suitable protection. A more severe problem appears to be the instability of carbanions adjacent to oxygen. Such carbanions, once generated, experience strong repulsion from the oxygen lone pairs making them energetically unfavorable. Hence, a-hydroxycarbanions , or their synthons, remain as challenging targets to synthetic chemists. 45

PAGE 55

46 a-Hydroxycarbanions (or their synthons) hold tremendous promise in their synthetic utility. Such synthons offer easy access to 1,3and 1,4-dihydroxy carbon fragments which are key segments in numerous antibiotics and, until now, have only been made by lengthy synthetic operations. 3 • 1 • 1 • Indirect Routes to a-Hydroxy Carbanions The a-hydrogens of an alcohol are not acidic enough to be abstracted by a strong base. Therefore, indirect methods have been adopted to generate a-hydroxy carbanion synthons. The most notable approach is the transmetallation (tinlithium) between tributyl stannylmethanol (3.1) and n-butyllithium which was reported by Seebach [80CB1290], This actually gives rise to a pentavalent "tin-ate" complex ( 3 . 2 ) (and not LiCH 2 OLi ) [80CB1290] which reacts with electrophiles to yield the a-substi tuted methanol derivatives (Scheme 3.1). 2 n-BuLi Bu 3 SnCH 2 -OH ( 3 . 1 ) -^ L \ Bu.Sn ,0 Li V ( 3 . 2 ) -> ECH 2 0H Scheme 3 . 1 Lithium-metalloid exchange (e.g. reductive cleavage of monothioketals [84JA1130] or 1 i thiodehalogenation of a-chloro ethers [64TL1503]) have also been used in generating a-lithio ethers but with limited success.

PAGE 56

47 3.1.2. ProtectionActiva tion Route to a-Hydroxyca rbanions The protection-activation route to a-hydroxyca rbani ons is a direct approach to a-hydroxycarbanion synthons. This involves three distinct steps (Scheme 3.2): i) protection of the more acidic hydroxyl proton ii) a-lithiation and subsequent alkylation with an electrophile, and finally iii) deprotection of the hydroxy group. step (i) step (ii) rch 2 -oh > rch 2 -o-p > + P Li I RCH-O-P E step(iii) RCH-OH < V E I RCH-O-P P P = Protecting Group Scheme 3 . 2 The crucial step in this sequence is the generation of the a-lithio species via lithium--a-hydrogen exchange. Since the a-hydrogen( s ) of alcohols are not sufficiently acidic to be abstracted by strong base the process of a-lithiation requires additional activation. This can be achieved in various ways. The most common practice is to choose a

PAGE 57

48 choose a suitable protecting group which simultaneously protects the hydroxy function and directs the lithiation on to the a-carbon atom. As would be expected, the choice of an appropriate protecting group is vital to the overall success of the process. This has been nicely demonstrated by Beak in his study on the a-lithiation of methanol and ethanol via their aryl esters (3.3) [77JA5213] (Scheme 3.3). In that report, the alcohol was first protected in the form of a highly hindered ester (3.3). a-Lithiation was achieved with s-butyll i thium in presence of tetramethylethylenediamine to form the a-lithio species 3.4. Alkylation of the latter, followed by deprotection with lithium aluminum hydride, liberated the a-alkylated alcohols 3.6. Thus, the a-lithio species (3.4) behaved as an a-hydroxy carbanion equivalent. R = H, CH 3 Scheme 3.3

PAGE 58

49 Several features of this work are worthy of discussion. It should be noted that a highly hindered aryl group was used. This was necessary to prevent any nucleophilic attack on the carbonyl group of the aryl ester (3.3). Ironically, this became a serious problem in the latter stage when deprotection of 3.5 by conventional hydrolysis could not be achieved. A very harsh deprotection step involving LiAlH^ had to be employed which is a serious drawback to this method. Nevertheless, the a-lithiation step (3.3 > 3.4) deserves special attention since it involves the process of a-activation . According to Beak, a-activation is derived through a prior complexation" phenomenon where the carbonyl oxygen of (3.3) first coordinates with the lithiating base ( s-butyllithium) . This evidently directs the delivery of the s-butyl residue on to the a-hydrogens of 3.3. Similar arguments have been forwarded to explain the regioand stereochemistry of various lithiation processes [86ACR356]. a— Li thi a t i on of an alcohol is much easier when an activating (mesomeric) group is present in the a-position. Thus upon treatment with a strong base ( t-butyl 1 i thium ) benzyltrimethylsilyl ether gives rise to the a-lithio species (3.7) [74JA3214]. Allyl silyl ethers can also be metallated to generate the delocalized lithio species (3.8) [ 76JOC3620 ] .

PAGE 59

50 Li Li Ph-CH-OSiMe 3 3 ( 3 . 7 ) ( 3 . 8 ) In recent years, great strides have been made in the utilization of silicon in organic synthesis [79MI1, 82MI1]. Organosilicon chemistry has shown two general trends. A silicon atom usually favors a positive charge "beta" to itself whereas it is capable of stabilizing an a-carbanion. Both these effects are due to the availability of d-orbitals on silicon. Thus the negative charge on an a-silyl carbanion can be effectively delocalized into the vacant d-orbitals of silicon. This has been amply demonstrated by Magnus who led the field in the application of a-silyl carbanions [80MI4] . A plethora of a-silyl carbanions are currently known, many of which also possess an a-hete roatom . Some of these, which have been useful to synthetic chemists, are shown below [ 80MI4 ] . Li Li Li Cl SiMe Ph SiMe 3 3 ( 3 . 9 ) ( 3 . 10 ) ( 3 . 11 )

PAGE 60

51 To us, the most encouraging example was that of ct-lithio tri me thylsilyl methyl methyl ether (3.9) [79CC822, 820M553], since it shows that an a-trimethylsilyl group can also provide enough activation to generate a— oxy carbanions. The a-lithio species (3.9) was shown to react with carbonyl electrophiles (aldehydes and ketones) and the derived products were generally obtained in fair to good yields [820M553] . Returning to the problems associated with a-oxycarbanions , one is faced with the facile 1,2-migration these species undergo via the Wittig rearrangement [ 70AG( E ) 763 ] . Thus, in sharp contrast to the a-li thio-me thyl ether derivative (3.9), the corresponding a-lithio-aryl ether species (3.13) undergoes Wittig rearrangement when kept at room temperature [82JOC5051] (Scheme 3.4). s-BuLi Li ArO-CH 2 SiMe 3 > Ar-O-CH-S iMe 3 (3.12) (3.13) i)H 2 0 l i i ) KF HO-CH 2 -Ar (3.15) -> LiO-CH \ Ar SiMe. (3.14) Scheme 3.4

PAGE 61

52 1,2-Anionic migrations in a-oxycarbanions are also common. Thus, both 1,2-acyl migration (e.g. 3.16 > 3.17) [81JOC2363] and 1 , 2-phosphoryl migration (e.g. 3.18 > 3.19) [80S289, 76TL47] have been reported for the respective a-oxycarbanions (Scheme 3.5). These rearrangements involve the transposition of the oxygen protecting group to the a-carbon. Therefore, the choice of an appropriate protecting group is of utmost importance for generating an a-oxy carbanion, especially via the protection-activation methodology . 0 Li II II 0 > OLi (3.16) (3.17) Li 0 II N 0 OLi (3.18) (3.19) Scheme 3.5

PAGE 62

53 The current status of the synthetic art concerning a-hydroxycarbanions can be summarized as follows: a) In the absence of a suitable activating group, a— hyd r oxy ca r ban i ons are very difficult to generate. b) Once generated, a-oxycarbanions tend to undergo 1,2-migrations either via the Wittig rearrangement or by a 1,2-anionic shift. These limitations are borne out by the experimental facts; until now only two a-hydroxy carbanion synthons, (3.2) and (3.4), have been reported which are derived from unactivated alcohols. ^ ^ ^ • Carbon Dioxide as a Protecting-Activating Group The 1,2-migrations, which are so problematic in the generation of a-oxy carbanions, could be minimized, in principle, if a-lithiation of anionically protected alcohols could be achieved. A proposed sequence is shown in Scheme 3.6. First the alcohol is protected by using a suitable protecting group (preferably a cumelene, Y=X=Z ) so as to generate an anionically protected species like 3.20. Successful a-deprotonation of 3.20 would give rise to the a-lithio species 3.21, in which any 1,2-anionic migration would be electronically disfavored. Alkylation of 3.21 followed by simple hydrolysis would afford the desired a-alkylated alcohols.

PAGE 63

54 i) Base Y R"Li Li Y rch 2 oh ii) Y=X=Z > r-ch 2 o-x <> RZ Z (3.20) (3.21) jii) h 2 o/h + E R-CH-OH Scheme 3 . 6 Earlier in section 2.2.3, carbon dioxide was introduced as a novel protecting group for amines. The resulting lithium carbamates proved to be useful substrates for a-lithiation of amines. Carbon dioxide appeared equally promising as a protecting group for alcohols. The derived hemi ca rbonate salts are very unstable in presence of moisture and rapidly lose carbon dioxide to return to the alcohol. In spite of their facile deprotection, the inherent instability of the hemicarbonates has restricted their synthetic utility as protected alcohols. To us, however, the hemi ca r bona t e salts appeared ideally suited for a— lithiation studies: (a) carbon dioxide is a non— toxic, cheap commercial material, (b) the hemicarbonate salts aptly qualify as anionically protected alcohols and (c) their deprotection would be extremely rapid in the presence of aqueous acid.

PAGE 64

55 The above hypothesis has already been put to test by previous workers in this group. Benzyl alcohol ( 3 . 22 ) was chosen by those workers as the model hydroxy compound: treatment of this alcohol with n-butyll i thium followed by CO^ readily gave the corresponding lithium carbonate ( 3 . 23 ). When exposed to _t-butyll i thium , the latter underwent smooth metallation giving rise to the a-lithio species ( 3 . 24 ) which was trapped with different electrophiles (D 2 0, CH 3 I, CC> 2 , PhCO ? Et, Ph 2 CO). Deprotection occurred spontaneously during acidic work-up and the a-alkylated benzyl alcohols ( 3 . 25 ) were isolated in fair to good yields [87S415] (Scheme 3.7). As an added advantage, the entire sequence was carried out in the same reaction vessel. i ) n-BuLi 0 t-BuLi PhCH 2 OH > PhCH,0-C' z \ > Ph-CH — 0 I \ ii) C0 2 OLi ( 3 . 22 ) ( 3 . 23 ) ( 3 . 24 ) PhCH-OH < I E -CO 2 E ( 3 . 25 ) Scheme 3.7

PAGE 65

56 Successful a-lithiation of lithium benzyl carbonate (3.23) is in sharp contrast to Seebach's study [80CB1304], the latter showing that benzyl alcohol on treatment with two equivalents of n-butyl 1 i thium afforded ortho -lithiation on the phenyl ring (Scheme 3.8). Scheme 3 . 8 In this context, a-lithiation of benzyltrimethylsilyl ether (3.26) is also worthy of discussion. a-Lithiation of 3.26 can be achieved with t-butyl 1 i thium to produce the a-lithio species (3.7), which however, undergoes a rapid 1 . 2si.lyl migration as shown in Scheme 3.9 (74JA3214). Such 1.2migration was completely suppressed during the a-lithiation of lithium benzyl carbonate (3.23).

PAGE 66

57 PhCH 2 -0-SiMe 3 t-BuLi ( 3 . 26 ) ( 3 . 7 ) 1 ^ \| SiMe-, / 3 PhCH ( 3 . 27 ) Scheme 3 . 9 3.1.4. Aims of the Work As discussed above, anionically protected alcohols have been introduced as superior substrates for a-lithiation and initial results with benzylalcohol via its lithium carbonate have shown considerable promise [87S415]. However, benzylalcohol is regarded as an activated system since the a-carbanion is stabilized by the adjacent phenyl group. Therefore, in the present work, it was our aim to probe into the limits of a-lithiation on lithium carbonates and for possible extension of this methodology to other alcohols. The lithium carbonate of methanol was chosen as the primary target, for two main reasons. Firstly, a-lithiation would generate a primary carbanion in this case. Any other alcohol would give rise to either a secondary or a tertiary carbanion which are more difficult propositions, due to electronic as well as steric reasons. Secondly, a-lithiation studies on lithium methyl carbonate would provide a direct comparison with Beak's aryl ester methodology [77JA5213] depicted in Scheme 3.3. The role of the lithium carbonates,

PAGE 67

58 if any, in the process of a-activation could thus be properly evaluated. a-Lithiation of 1-trimethylsilylmethanol (via its lithium carbonate) appeared very promising for two main reasons. Firstly, the trimethylsilyl group has the proven ability to stabilize an a-carbanion. Secondly, a judicious choice of the electrophile would enable the trimethylsilyl auxilliary to be readily removed from the product. The latter prediction has strong literature precedents [74ACR77] in that a silicon atom is extremely labile when adjacent to carbonyl functions. Thus the successful a-lithiation of lithium 1-tri me thylsilyl methyl carbonate would formally constitute the generation of a methanol dianion synthon. a-Li thiation of allylic alcohols has been the subject of numerous synthetic efforts in the past and present [ 84AG( E ) 932 ] . The generated a-carbanions can be effectively delocalized to the y-carbon and, as such, varying degrees of selectivity (a vs. y) are observed in the alkylation of these anions. Several oxygen protecting groups with varying co-ordinating powers have been investigated to control this selectivity [ 8 4 AG ( E ) 9 3 2 ] . In this perspective it was of considerable interest to investigate the directing power of

PAGE 68

59 an allyl lithium carbonate. A strong chelating effect ( f ive-membe red chelate) from the carbonate function would be expected to increase the proportion of a-alkylation over the y-attack. The extent to which this selectivity could be achieved was the principal objective in this case. 3 • 2 . Results and Discussion a-Lithiation Studies on Lithium Methyl Carbonate Lithium methyl carbonate ( 3 . 29 ) was prepared in a manner similar to that of benzyl alcohol. Thus methanol ( 3 . 28 ) on treatment with n-butyll i thium followed by carbon dioxide produced the lithium methyl carbonate ( 3 . 29 ) (Scheme 3.10). g-Lithiation of 3.29 was attempted with t-butyl 1 i thium in tetrahydrofuran at -20°C for 2 hrs. Subsequent addition of benzaldehyde followed by acidic work-up did not afford the desired 1,2-diol ( 3 . 31 ), benzaldehyde being recovered unchanged. A repeat reaction replacing benzaldehyde with ethyl benzoate failed, as well. Addition of HMPA or TMEDA to the reaction mixture also showed no promise; it was inferrred that the desired lithium lithiomethyl carbonate ( 3 . 30 ) vas not formed under these conditions.

PAGE 69

60 i) n-BuLi t-BuLi > CH 3 0C0 2 Li // — > LiCH 2 0C0 2 Li ii) C0 2 (3.29) (3.30) i ) PhCHO ii) H + PhCH-CH-,OH I 2 OH (3.31) Scheme 3.10 A comparison of these results with Beak's aryl ester methodology (£f. Scheme 3.3) shows that the lithium carbonates have no perceptible a-activating effects. Thus it became evident that the hemi ca rbonate route to a-oxy carbanions can be applied only to activated systems. This led to the prospect of utilizing a trimethylsilyl group as an easily removable a-activating auxilliary. 3.2.2. g-Lithiation Studies on Lithium Trimethylsilylmethyl Carbonate Commercially available 1-trimethylsilylmethanol (3.32) was easily converted to its lithium carbonate (3.33) following the standard procedure ( n-butyllithium C0 2 ) . The latter was then treated with s-butyllithium in ch 3 oh (3.28)

PAGE 70

61 tetrahydrofuran at -78°C. After stirring the virtually colorless solution for 2 h at -25°C, ethyl £ -toluate was added as the electrophile and the resulting mixture stirred at room temperature for 8 hrs. Subsequent work-up with 10% aqueous hydrochloric acid afforded the a-hydroxyacetophenone derivative (3.36) in 68% yield (Scheme 3.11). The use of butyll i thium in the metallation step was found to be essential as inferior yields were obtained when n— butyl 1 i thium or butyllithium were used. i) n-BuLi Me 3 SiCH 2 OH — (3.32) ii) C0 2 > Me 3 SiCH 2 OCOLi (3.33) THF/-2 5 0 C 0 S-BuLi II Me 3 SiCH-0‘ (3.34) 0 ./ ^ Ll C-OLi I / ArC0 2 Et ArCO-CH 2 OH < (3.36) COAr (3.35) Ar = pMe -C 6 H 4 Scheme 3.11

PAGE 71

62 Formation of the a-hydroxyketone (3.36) can be rationalized via the intermediacy of the cc-lithio species (3.34). Reaction of 3.34 with the ester electrophile presumably led to the a-silyl ketone derivative (3.35), which on hydrolysis lost CO 2 with concommitant protiodesilylation to give the hydroxymethylketone (3.36) (Scheme 3.11). The facile desilylation of 3.35 is to be expected since a-silyl carbonyl compounds are known to be readily desilylated upon hydrolysis [74ACR77]. Other carbonyl electrophiles reacted in the same way as ethyl £toluate . The results are summarized in Table 3.1. Thus N, N-di methyl be nzamide and benzoyl chloride led to a-hydroxyacetophenone , as expected. The low yield obtained with benzoyl chloride was probably caused by undesired side reaction involving metal-halogen exchange. Benzonitrile was also successfully employed as the electrophile; the intermediate imine adduct hydrolyzed to the ketone on acidic work-up. Quite significant are the results when a, 6— unsaturated esters were used as electrophiles (entries 5 and 6, Table 3.1). in both cases, carbonyl addition was the exclusive pathway giving rise to the hydroxymethyl vinyl ketones. The 1,4-addition mode was not observed.

PAGE 72

63 Table 3.1. Reaction of the a-Lithio Lithium Carbonate (3.34) with Electr ophiles Electrophile Product Yield( % ) 1 . £-Me-CgH 4 C0 2 Et £-Me-CgH 4 COCH 2 OH 68 2. PhCOCl PhCOCH 2 OH 23 3 . PhCONMe 2 PhCOCH 2 OH 63 4 . PhCN PhCOCH 2 OH 65 5. PhCH=CHC0 2 Et PhCH=CHCOCH 2 OH 45 6. Me 2 C=CHC0 2 Et Me 2 C=CHCOCH 2 OH 60 All the hydroxymethyl ketones prepared in this study were characterized by their NMR spectra ( 1 H and 13 C) and through matching melting point data (in case of known compounds). The H-NMR spectra of these compounds showed a characteristic two-proton singlet around 4.8 ppm which was assigned to the methylene protons. Their 13 C-NMR spectra displayed the keto-carbonyl , usually near 198 ppm, whereas the methylene carbon appeared around 67 ppm. Saturated aliphatic esters such as ethyl propionate or ethyl butyrate, however, failed to provide homogenuous products when used as electrophiles. Similar failure was also observed with other potential acylating agents like propionic anhydride and N, N-dimethylpropionamide . Dimethyl carbonate and phenyl isocyanate also failed to give the desired products.

PAGE 73

64 The results obtained thus far (Table 3.1) did not necessarily prove the formation of the a— lithio species (3.34). There are reports that alkyllithiums attack the silicon center in certain Si-C bonds giving rise to the C-Li species [80MI4]. Such an attack on the lithium carbonate (3.33) could potentially give rise to the lithiomethyl lithium carbonate (3.30) (Scheme 3.12). The latter, upon reaction with esters, would afford the same results . R L Li > LiO^C-OCH^Li (3.30) RCOX V RCOCH 2 OH Scheme 3.12 The nature of the a-lithio species was decisively settled by using benzaldehyde as the electrophile. It was envisaged that reaction of benzaldehyde with the s i lyl-a-1 i thi o species (3.34) would initially produce a 3-hydroxysilyl adduct (3.37). The latter, after Peterson olefination [84S384] and subsequent hydrolysis of the incipient styryl carbonate (3.38), should lead to phenylacetaldehyde (3.39) (Scheme 3.13). On the other hand, O * LiO 9 C-OCH 0 -SiMe 0 . 33 )

PAGE 74

65 if a-lithiomethyl lithium carbonate (3.30) was indeed the reactive species, then its reaction with benzaldehyde would produce the diol (3.31). Li I Me,SiCHO-CO,Li (3.34) Me 3 SiCH 2 0-C0 2 Li (3.33) LiCH 7 0-C0,Li (3.30) PhCHO Me.,SiCH0-C0 o Li 3 | 2 LiO^Ph (3.37) PhCHO V PhCH-CH~OH I 2 OH (3.31) 'V / _/”0C0 2 Li Ph-> (3.38) H + /H 2 0 > PhCH ? CHO (3.39) Scheme 3.13

PAGE 75

66 In practice, a-lithiation of the lithium carbonate (3.33) with s-butyllithium (-25°C, 2 hrs) followed by addition of benzaldehyde produced phenylacetaldehyde (3.39) in 70% yield, as the sole product. This proved that the s i lyl-a-1 i thi o species (3.34) is indeed formed and no 1 i thiodesi lylation (to give 3.30) had taken place. This also reinforced the earlier view that with the other carbonyl electrophiles a-si lyl-a-hydroxy ketones are the initial products which protiodesilylate upon hydrolysis. The fact that lithium a-lithiotrimethylsilylmethyl carbonate (3.34) converted benzaldehyde to phenylacetaldehyde immediately opens up a novel homologation process for aldehydes and ketones (Scheme 3.14). These transformations, although very common in multistep synthesis, suffer from the lack any general preparative method [76T1943], The synthon (3.34) promised a one-pot procedure to achieve such transformations. Me„SiCH0C0 2 Li Li (3.34) II 0 R 1 \ ^''OCC^Li CH-CHO < R CO 2 (3.40) Scheme 3.14

PAGE 76

67 In practice, reaction of the a-lithio species (3.34) with cyclohexanone resulted in a complex product mixture. Addition of MgB^ or BF2-Et20 to the reaction mixture did not lead to the desired cyclohexane ca rboxaldehyde . In view of the enolization problems with cyclohexanone, benzophenone was next tried as the electrophile; however, the desired product could not be obtained. Further study with 3.34 as a reductive formyl anion equivalent was therefore abandoned. At this time our attention was drawn to the recent fluoride induced desilylative alkylation of a— si lylbenzyl alcohol [84CL1803]. However , no reaction was observed when benzaldehyde and the lithium carbonate (3.33) were together treated with (Bu)^N + F (in tetrahydrofuran , 25°C, 24 hrs) or with CsF (in hexamethylphosphoramide , 60°C, 6 hrs). In another study, Sakurai has shown that one of the Si-C bonds in bi s ( t r i me thy 1 s i ly 1 ) me thane could be cleaved by sodium methoxide in hexamethylphosphoramide to produce the tri me thylsilyl methyl carbanion [ 73TL4193 ] . A similar treatment of the lithium carbonate (3.33) was not fruitful and returned the starting materials. Generation of the methanol dianion equivalent directly from the lithium carbonate (3.33) was therefore deemed impossible.

PAGE 77

68 ^•2.3. a-Lithiation Studies on Lithium Cinnamyl Carbonate Previous work in this group showed that allyl lithium carbonate could not be a-lithiated with t^-butyllithium [86MI1], Hence, a more activated system such as cinnamyl alcohol ( 3 . 41 ) was chosen for the present work. The required lithium cinnamyl carbonate ( 3 . 42 ) could be prepared very easily as shown in Scheme 3.15. The lithium carbonate ( 3 . 42 ) was next treated with t-butyllithium followed by addition of D 2 0. Hydrolytic work-up followed by 1 H-NMR analysis of the product showed no deuterium incorporation in the molecule. This evidently showed that the desired allyl carbanion ( 3 . 43 ) could not be generated. i ) n-BuLi ( 3 . 41 ) ii) C0 2 t-BuLi ( 3 . 42 ) Li0 2 C_0 \^V^ Ph Li ( 3 . 43 ) Scheme 3.15 Since other methods exist which allows for the generation of a-oxyallyl carbanions [84AG(E)932, 8 lAG ( E ) 1 2 7 ] , further investigation on the allyl hemicarbonates was discontinued.

PAGE 78

69 3.3. Conclusions The advantage of using carbon dioxide as a hydroxyl protecting group has been efficiently demonstrated. It is easily introduced and subsequent deprotection of the resultant hemicarbonates can be achieved under extremely mild conditions. The lithium carbonates are virtually inert to most metallating agents and the process of a-lithiation is devoid of the notorious 1,2-anionic migration. However, the lithium carbonates play no role in the process of a-activation . Presumably, the delocalized negative charge on the carbonate moiety reduces its co-ordinating power and the carbonyl oxygen fails to activate the lithiating agent ( via the prior complexation mechanism [86ACR356]). Lithium carbonates with appropriate a-activating auxilliaries promise very interesting synthetic transformations. As a result of the present investigation, lithium a-lithiotrimethylsilylmethyl carbonate (3.34) has emerged as a novel a-hydroxycarbanion synthon. In spite of its limitations, a novel synthesis of a-hydroxyketones has been developed by proper utilization of the synthon 3.34 and that too in a one-pot sequence starting from a commercially available material. a-Hydroxyketones are traditionally prepared via electrophilic hydroxylation of ketone enolates,

PAGE 79

70 a procedure not without complications. Direct oxygenation often results in complex product mixtures due to overoxidation [ 62J1578 , 68JOC3294 , 68JOC3695 ] . The MoO 5 .Py.HMPA complex has also been used for this purpose but the products from methylketone enolates were found to undergo further condensation with the starting material, leading to poor yields [78JOC188]. However, Rubottom's m-CPBA oxidation of silyl enol ethers [78JOC1599] or Moriarty's method using hypervalent iodine oxidation [84TL691] are quite effective. By contrast, our route to hydroxymethyl ketones utilizes a novel sequence which is conceptually opposite to the existing ones and demonstrates the viability of an umpolung approach to these compounds. 3.4. Expe r imental Melting points were determined using a Bristoline hot-stage microscope and are uncorrected. NMR spectra were recorded on a Varian XL200 instrument using Me^Si as the internal standard for protons and CDCl^ (77.0 ppm) as reference for carbon resonances. The nature of the carbon signals in the "*'^C-NMR spectra were determined by the

PAGE 80

71 Attached Proton Test [81CC150, 82JMR535]. Mass spectra were recorded at 70 ev on a AEI MS instrument attached to a DS 55 database. All reactions were carried out in oven-dried (120°C, overnight) apparatus under a slight positive pressure of dry argon and transferring operations done using syringe techniques or via cannula. Tetrahydrofuran (THF) was freshly distilled from sodium-benzophenone ketyl just prior to use. Hexamethylphosphoramide ( HMPA ) and tetramethylethylenediamine (TMEDA) were distilled over CaH-, under reduced pressure and stored under nitrogen. Flash chromatography [78JOC2923] was carried out over silica gel ( MCB , 230-400 mesh) using pre-distilled solvents. General Procedure for the Preparation and a-Alkylation of Lithium Trimethylsilylmethyl Carbonate (3.33): A 250 ml Schlenk reactor (oven-dried at 120°C overnight) was evacuated under 0.5 mm for 2 hrs and purged with argon. 1Trimethylsilylmethanol (3.32) (Aldrich or Petrarch) (0.52 g, 5.0 mmole) was added followed by dry THF (35 ml). After cooling to -78°C, n-BuLi (2.1 ml, 2.5 M in hexanes, 5.2 mmole) was added dropwise. The reactor was warmed to room temperature and held there for 10 min. It was then recooled

PAGE 81

72 to -25°C and dry CC^ gas bubbled through it. The cooling bath was removed and while CO.-, was still bubbled in, the reactor was allowed to warm to room temperature. After a further 5 min all volatile contents were evacuated at 0.5 mm leaving a white powdery residue of the lithium t r ime thy 1 s i lylme thyl carbonate (3.33). The reactor was again purged with argon and dry THF (45 ml) was added to redissolve the white residue. After cooling to -78 °C, s-BuLi (4.2 ml, 1.3 M in cyclohexane, 5.5 mmole) was added dropwise. Stirring was continued at -25°C for 2 h. To this virtually colorless solution was then added the appropriate electrophile (5.2 mmole) in dry THF (15 ml). The reactor was warmed to room temperature and the contents stirred for 30 min. The THF was removed under reduced pressure (on a rotary evaporator) and the residue treated with 10% aqueous HC1 . After stirring for 15 min the aqueous phase was extracted with ether or ethyl acetate (3 X 15 ml). The combined organic layers were washed sequentially with water (10 ml), saturated NaHCC> 3 solution (10 ml) and brine (10 ml) and dried over anhydrous Na 2 SO^. Removal of solvent under reduced pressure gave the products which were purified as described for the individual cases below.

PAGE 82

73 The following compounds were made according to the above procedure : 2-Hydroxy-( 4 ' -methyl )phenylethanone Ethyl p-toluate (0.82 g) was used as the electrophile. Silica gel chromatography using petroleum ether 25% ethyl acetate afforded the product (0.50 g, 68%); m.p. 86-89°C (lit. [06CB3757] m.p. 89-90°C); § H (CDC1 3 ) 7.9 (2H, d, J 8 Hz), 7.4 ( 2H , d, J 8 Hz), 5.0 (2H, s), 3. 5-3. 2 (1H, broad) and 2.6 (3H, s); S c (CDC1 3 ) 197.9 (C=0), 145.3 (aryl CH), 130.8 (aryl CH ) , 129.6 (aryl C) , 127.8 (aryl C) , 65.2 (CH 2 ) and 21.8 ( CH^ ) . 2-Hydroxyphenylethanone (a) Benzoyl chloride (0.74 g) was used as the electrophile. Silica gel chromatography using petroleum ether 20% ethyl acetate afforded the product (0.16 g, 23%); m.p. 86-88 °C (lit. [ 12LA( 394 ) 42 ] m.p. 89-90°C); 6„ 8.07.9 (2H, m), 7. 7-7. 5 (3H, m) , 4.9 (2H, s) and 3. 5-3. 4 (1H, broad ) .

PAGE 83

(b) N , N-D i me t hyl ben z amide (0.74 g) was used as the electrophile. Silica gel chromatography using petroleum ether 20% ethyl acetate afforded the same product (0.42 g, 63%) identical in melting point and 1 H-NMR data to above. (c) Benzonitrile (0.54 g) was used as electrophile. Silica gel chromatography using petroleum ether 25% ethyl acetate afforded the same product (0.44 g, 65%) identical in melting point and 1 H-NMR data to above. t ran s-1 -Hydroxy4 -phenyl but3 -en2 -one trans -Ethyl cinnamate (0.88 g) was used as the electrophile. Silica gel chromatography using petroleum ether 20% ethyl acetate afforded the product (0.34 g, 42%); m.p. 65-68°C (lit. [84TL691] m.p. 69-70. 5°C); 8 U (CDC1 3 ) 7. 9-7. 5 ( 6H , m), 6.8 (1H, d, J 15 Hz), 4.5 (2H, s) and 2.9 (broad, OH); S c (CDC1 3 ) 198.1 (C-O), 144.1, 133.4, 131.1, 129.0, 128.4, 121.0 and 67.0 (CH-). l-Hydroxy-4-methylpent-3-en-2-one Ethyl 3,3-dimethylacrylate (0.64 g) was used as the electrophile. Silica gel chromatography using petroleum ether 10% ethyl acetate afforded the product (0.34 g, 60%)

PAGE 84

75 as an oil; $ H (CDC1 3 ) 5.9 (1H, s), 4.2 (2H, s), 2.3 (broad, OH), 2.2 (3H, s) and 1.9 (3H, s); 6 C (CDC1 3 ) 198.0 (C=0), 158.9 (C-3), 119.0 ( C-2 ) , 68.6 (CH 2 ), 27.8 (CH 3 ) and 21.4 (CH 3 ); HRMS: M + * found 114.0728, CgH 1Q 0 2 requires 114.0681. Phenylacetaldehyde Benzaldehyde (0.52 g) was used as the electrophile. Filtration through a short column of neutral alumina using hexane-2% ethyl acetate afforded the product (0.52 g, 78%); S H ( CDCl 3 ) 9.74 (1H, t, J 2 Hz), 7. 4-7. 2 (5H, m) and 3.67 ( 2H, d, J 2 Hz ) .

PAGE 85

CHAPTER IV THERMAL BEHAVIOR OF N-VINYL-1 , 2-DIHYDROPYRIDINES 4.1. Introduction 4.1.1. 1 , 2-Dihydropyr i dines Of the five possible classes of dihydropyridine (4. 1-4. 5), the 1,2(4.1) and 1 , 4-dihydropyr idines (4.2) have attracted most attention. The rest have been less studied because of their instability and the lack of good preparative methods. R R (4.1) (4.2) (4.3) (4.4) (4.5) The chemistry of dihydropyr idines has been extensively reviewed [ 72CRV1 , 82CRV223 , 82MI1 ] . As compared to 76

PAGE 86

77 1 , 4-dihydropyr idines , which have been widely exploited as NADH mimics, 1 , 2-dihydropyridines until recently, have found less prominence. This may be partly attributed to their lower stability [72JA5926]; 1 , 2-dihydropyridines are readily oxidized to the corresponding pyridines. In recent years, however, 1 , 2-dihydropyridines have found widespread use in natural product synthesis; these highly reactive compounds have proved to be important building blocks in the synthesis of various alkaloids, especially for the indoloquinol i z idine [ 7 5 ACS ( B ) 6 55 , 7 6 ACS (B)251,80TL2341] , elaeocarpine [71TL4395] and cantharanthine [ 78TL5157 , 79TL2485 , 80JOC3382 ] families. 4.1.2. Synthetic Routes to 1 , 2-Dihydropyr idines Due to their inherent instability there are only few methods of preparing 1 , 2-dihydropyr idines . These are described below. 4. 1.2.1. Reaction of Nucleophiles with Pyridines Strong nucleophiles such as phenyllithium add to pyridine exclusively at the 2-position to form the highly reactive N-lithio-1 , 2-dihydropyr idine ( 4 . 6 ) [71TL4961]. This is currently the only practical way of synthesizing 1 , 2-dihydropyr idines directly from pyridines (Scheme 4.1).

PAGE 87

78 (4.6) Scheme 4.1 Normally pyridines (unless highly deactivated by electron withdrawing groups) are not reduced by sodium borohydride. However, in presence of alkyl chlorof ormates , sodium borohydride reduces the incipient pyridinium salt to give high yields of N-ca rbamoyl-1 , 2-dihydropyridines [72JOC1321] (Scheme 4.2). In an analogous manner, other nucleophiles such as Grignard reagents [ 82 J0C4 31 5 , 8 3TL1 8 01 , 84TL4867] and cyanide ion (to form Reissert compounds) [85CI(L)125] have also been successfully employed. + C1C0 2 R NaBH 4 > CH 3 OH, -7 0 0 C Scheme 4.2

PAGE 88

79 4. 1.2. 2. Reaction of Nucleophiles with Pyridinium Salts Pyridinium salts are more susceptible to nucleophilic attack than the corresponding pyridines. However, depending on the nucleophile, the regioselectivity of the attack (1,2vs . 1,4-) varies and, quite often, mixtures of 1,2and 1 , 4-dihydropyridines are obtained. This problem has been addressed with reference to the "hardness" or "softness" of the nucleophile [65TL4615] and there is evidence that in some cases a mixture of 1,2and 1,4-isomers are formed under kinetic control which subsequently equilibrates to the thermodynamically more stable 1,4-isomer [79JOC1757]. Enhanced 1 , 2-regioselectivity is generally observed in the reaction of pyridinium salts with organometallic reagents. Thus Grignard reagents generally attack at the 2-position to form 1 , 2-dihydropyridines [ 74TL59 , 71JOC772 ] (Scheme 4.3). Good yields of 1 , 2-dihydropyridines are also obtained when phenyllithium or t-butylli thium is used [ 72CZ411 , 74JOC59 ] but simpler alkyl 1 i thiums tend to give very low yields. Organocuprates usually attack pyridinium salts at the 4-position [74CJC3563]. Nucleophilic attack on a 4-substituted pyridinium salt invariably occurs at the 2-position. One exception being benzylic organotin reagents, which, in the presence of chloroformates , always attack the 4-position of a pyridine, even when a 4-substituted pyridine is used [ 86TL211 ] .

PAGE 89

80 PhMgBr 70% EtMgB r 70% PhLi 80% n-PrLi 20% Scheme 4.3 Sodium borohydride reduction of pyridinium salts is difficult to arrest at the dihydropyridine stage and over-reduction to tetrahydropyr idines usually occur [ 53BSF53 , 70JOC2809 , 86AHC( 39 ) 1 ] . To summarize, the most synthetically useful methods for preparing 1 , 2-dihydropyr idines are: i) Reaction of pyridines with sodium borohydride in presence of chlorof ormates , or ii) Reaction of pyridinium salts with Grignard reagents . 4.1.3. Reactivity of 1 , 2-Dihydropyr idines The reactivity of 1 , 2-dihydropyr idines ranges from that of an enamine [ 7 5CJC2 30 5 , 76 JHC789 , 79 JCS ( PI ) 3 082 , 7 5 JOC569 , 78CJC1026 , 79CJC2342 , 79JHC409 ] to that of a cyclic 1,3-diene.

PAGE 90

81 4. 1.3.1. Cycloaddition Reactions Depending on the nature of the N-substi tuent , 1 . 2dihydropyr idines behave as either a 4 it or a 2n component in cycloaddition reactions. When the N-substi tuent is an electron withdrawing group, 1 , 2-dihydropyridines behave as the 4n component in cycloaddition reactions. There are numerous reports of intermolecular Diels-Alder reactions where 1 , 2-dihydropyr idines were effeciently used as the diene component [ 76JHC481 , 80JA6157 , 80CJC2447 , 81JOC4836 , 82TL2527 , 83TL2927 , 85JOC3236 , 85TL2617 ] . The intramolecular version is also known [ 74HCA1204 , 83TL2711 ] . In absence of electron withdrawing N-substi tuents , 1 . 2dihydropyr idines behave more like enamines and hence undergo thermal [2+2] cycloadditions. An example of thermal [2+2] cycloaddition of 1 , 2-dihydropyr idines is shown in Scheme 4.4. Thus 1-phenyl-l , 2-dihydropyr idine ( 4 . 7 ) on heating with dimethyl acetylenedicarboxylate gave rise to the 1 , 2-dihydroazocine derivative ( 4 . 9 ). The formation of the latter was rationalized through an initial [2+2] adduct (4.8) which underwent further cycloreversion to produce the observed product [ 7 4 JCS ( PI ) 2 496 ] .

PAGE 91

82 CO.,Me l 2 I Ph (4.7) C0 2 Me J' -CO^Me N C0 o Me I 2 Ph (4.8) C0 2 Me C0 2 Me Ph (4.9) Scheme 4.4 4. 1.3. 2. Electrocyclic Reactions The ANRORC reaction (Addition of Nucleophile Ring Opening Ring Closure) is probably the most studied electrocyclic reaction involving 1 , 2-dihydropyr idines . A crucial step in this reaction sequence is the electrocyclic ring opening of the intermediate 1 , 2-dihydropyr idine (4.11) to the triene (4.12) (Scheme 4.5). A detailed mechanistic discussion can be found in the recent reviews on ANRORC-type reactions [ 80MI 3 , 80S589 , 81T3423 ] . Similar ring openings of pyridine N-oxides are also known [ 67T2775 , 69T4291 , 71 JOC1705 , 74T4055 , 76TL4717 ] .

PAGE 92

83 (4.10) (4.11) (4.12) ArNH 2 + Ar = 2 , 4-dinitrophenyl Scheme 4 . 5 Thermally-induced electrocyclic reactions are very rare for 1 , 2-dihydropyridines . One such example is the thermal isomerization of 1-ethyl-l , 2-dihydropy r idine (4.13) to the 2 , 3-dihydropyr idine derivative (4.14) (Scheme 4.6) [78JA6696]. The lack of many such examples probably reflects the greater thermodynamic stability of 1 , 2— dihydropyr idines as compared to 1-azatrienes [ 8 OMI 3 ] .

PAGE 93

84 If 1 A j > CH 2 ch 3 ch 2 ( 4 . 13 ) [1,7-H] ,/n^s^CH. N CH^N'^CH. ( 4 . 14 ) Scheme 4.6 4.1.4. Effects of N-Substi tuents on the Thermal Behavior of 1 , 2-Pihydropyridnes 4 . 1 . 4 . 1 . N-Alkyl and N-Aryl Substituents N-Aryl-1 , 2-dihydropyr idines are more stable under pyrolytic conditions than the corresponding N-alkyl derivatives [ 79JCS ( PI ) 442 ] . The latter, on pyrolysis, yield the parent pyridine together with the hydrocarbon derived from the N-substi tuent . This has been effectively exploited

PAGE 94

85 by Katritzky and co-workers for the three-step reductive deamination of primary amines [ 84 JCS ( PI ) 1671 ] (Scheme 4.7). Although usually limited to allyl, benzyl or heterobenzyl amines, in some cases, it can be extended to alkyl amines. Ph Ph Ph ( 4 . 15 ) ( 4 . 16 ) ( 4 . 17 ) ji + R-H r-nh 2 R = allyl, benzyl or heterobenzyl Scheme 4.7 4. 1.4. 2. N-Vinyl Substituents In a manner similar to the N-alkyl-1 , 2-dihydropyr idines , pyrolysis of a N-vinyl-1 , 2-dihydropyr idine might be expected to yield an olefin corresponding to the N-vinyl substituent; however, this was not the case. When the dihydropyridine ( 4 . 18 ) was pyrolyzed for 2 hrs (180°C, 0.1 mm),

PAGE 95

86 m-chlorostyrene and 2 , 4 , 6-tr iphenylpyr idine (4.17) were formed [82JOC492] (Scheme 4.8). Ph Ph ( 4 . 18 ) Ar = 3-chlorophenyl Scheme 4.8 Apparently the result of a rearrangement, the above transformation was rationalized as shown in Scheme 4.9. After initial electrocyclic ring opening of the 1 , 2-dihydropyr idine ( 4 . 18 ), the resultant 3-azatetraene ( 4 . 19 ) could lead to the bicyclic intermediate ( 4 . 21 ) via two consecutive electrocyclizations . The intermediate ( 4 . 21 ) underwent fragmentation and consequent aromatization to afford the observed products. Pyrolysis of some other N-styryl-1 , 2-dihydropyr idines followed the same fragmentation pattern [82JOC492].

PAGE 96

87 Ph ( 4 . 17 ) 'T Ph Ar Scheme 4.9 4.1.5. Aims of the work As discussed above, pyrolysis of N-vinyl-1 , 2-dihydropyr idines led to an interesting f ragmentative rearrangement. The aim of the present project was to extend this observation to 1 , 2-dihydropyridines having an N-cycloalkenyl substituent. Thus the latter class

PAGE 97

88 of dihydropyr idines (e.g. 4.22), on pyrolysis, would be expected to give pyridines 4.23, having long side chains ( Scheme 4.10) . fl h N 1 A (4.22) CH < CH 2>n 2 n CH, (4.23) Scheme 4.10 As a further extension, by capping the dihydropyridine nucleus with cycloalkyl groups as in 4.24, medium-ring fused pyridines such as 4.25 might be obtained (Scheme 4.11). Scheme 4.11

PAGE 98

89 4.2. Results and Discussion 4.2.1. Synthetic Strategy We intended to prepare the required N-cycloalkenyl-1 , 2-dihydropy ridines by sodium borohydride reduction of the corresponding pyridinium salts. Thus our primary synthetic targets were the N-cycloalkenylpyr idinium salts. At the time of this work, no such N-vinylpyr idinium salts were known where the vinyl grouping was part of a carbocycle. Our strategy for the preparation of such salts comprised of two basic pathways as illustrated in Scheme 4.12. In Path A, it was envisioned that a suitable pyridine (4.26) would displace the halogen from 6-chlorocycloalkenones (4.27), thus leading to N-cycloalkenonylpy r idinium salts (4.28). Path B, on the other hand, requires the preformation of a suitable pyridinium salt (4.29) with a ^-substituted cycloalkyl group as the N-substi tuent . The latter, on elimination of HX, might afford the desired N-cycloalkenylpyr idinium salts (4.30). The latter pathway is quite similar to the one generally followed for the preparation of acyclic N-vinylpyr idinium salts [83JOC4017].

PAGE 99

90 Path A: (4.26) Path B: (4.29) (4.30) X = Cl , Br , OH Scheme 4.12 4.2.2. Preparation of N-( Cycloalken-3-on-l-yl ) pyridinium Salts To test the viability of Path A in Scheme 4.12, the reaction of pyridine and g-chlorocyclohexenone [76JOC636] was attempted. The reaction was monitored by 1 H-NMR (CDCl^, 55°C) and was found to be quite slow. It was only after

PAGE 100

91 48 hrs that the a-protons of the pyridinium salt (4.31a) appeared in the spectrum (at ~ 10.1 ppm ). However, with more nucleophilic pyridines such as £-dimethylaminopyr idine ( DMAP ) the reaction was much faster and the derived pyridinium salt (4.31b) could be isolated in very good yield (94%) (Scheme 4.13). 4-Phenylpyr idine could also be used which gave the corresponding pyridinium salt (4.31c) in 57% yield; 4-picoline, however, gave an intractable product mixture. In these reactions, the best yields were obtained when benzene was used as the solvent; in chloroform or acetonitrile the yields were much lower. 0 (4.31) a, R = H; b, R = NMe 2 ; c, R = Ph Scheme 4.13 When the tricyclic pyridine (4.32) [ 71JPS( A-I ) ( 9 ) 1807 ] was reacted with g-chlorocyclohexenone , the reaction resulted in a complete failure, even when the components were heated in a sealed tube at 200°C for 24 hrs; starting

PAGE 101

92 materials were recovered in each case. Steric inhibition on the part of the pyridine (4.32) was thought to be responsible for this failure. (4.32) 4.2.3. Preparation of N-(Cycloalken-3-on-l-yl)-l,2dinydropyr idines To prepare the N-cycloalkenonyl-1 , 2-dihydropyr idines , the pyridinium salts 4.31b and 4.31c were subjected to sodium borohydride reduction. The salt 4.31b showed a high degree of resistance toward borohydride reduction, even in refluxing methanol, presumably due to the strong electron-releasing effect of the £-dimethylamino group. On the other hand, the p-phenylpy r idinium salt 4.31c was smoothly reduced to the corresponding 1 , 2-dihydropyr idine (4.33) in excellent yield (91%) (Scheme 4.14). (4.31c) (4.33) Scheme 4.14

PAGE 102

93 The structure of 4.33 was confirmed by its and 13 1 C-NMR spectra. The H-NMR chemical shifts of 4.33 are shown in Table 4.1. The two H-2 hydrogens showed the expected doublet (J = 4 Hz) at 6 4.4. The H-6 hydrogen appeared as a more downfield doublet (J = 8 Hz) at S 6.6 whereas the H-3 and H-5 hydrogens formed a complex multiplet at 6 5. 4-5. 8. These chemical shifts and the multiplicities are in good accord with the characteristic pattern found in the reported ^H-NMR spectra of 1 , 2-dihydropyr idines [72CRV1]. In addition, the one-proton singlet at & 5.3 was assigned to the vinyl proton of the N-cyclohexenone substituent . The 13 C-NMR spectrum of the dihydropyridine 4.33 showed the carbonyl carbon at 197.0 ppm and three downfield methine signals at S 114.6, 104.7 and 102.9 which were assigned to C-3, C-5 and the C-2 ' carbons, respectively. These assignments are only tentative and for C-5 and C-2' they may be interchanged. However, the C-2 methylene carbon could be assigned with certainity, being the most deshielded aliphatic signal at 47.3 ppm. The three remaining cyclohexenone carbons appeared at 36.0, 25.9 and 21.9 ppm.

PAGE 103

94 (4.33) Table 4. 1 1 H-NMR Data for Dihydropyridine (4.33) a Chemical Shift Values N-Substi tuent Di hydropyridine Ring Phenyl Vinyl Others 2 3 & 5 6 Ring 5 . 3 ( s ) b 1.8-2. 5(m) 4. 4 ( d ) c 5 . 4-5 . 8 ( m) 6 . 6 ( d ) d 7 . 4-7 . 6 ( m ) chemical shifts in ppm downfield from Me^Si. b multiplicities are given in parentheses : s = singlet, d = doublet, m = unresolved multiplet. C J = 4 Hz. d J = 8 Hz.

PAGE 104

95 4.2.4. Pyrolysis ofN(Cycloalken3-on-l-yl)-4-pheny 1-1,2dihydropyr idine As described above, pyrolysis of N-cycloalkenyl-1 , 2-dihydropy r idines was our main objective. We embarked upon this objective with the dihydropyridine 4.33 in hand. According to the mechanism illustrated in Scheme 4.10, it was expected that the pyridine derivative 4.34 would result from the pyrolysis of 4.33 (Scheme 4.15). Ph (4.33) (4.34) Scheme 4.15 In practice, the pyrolysis of 4.33 was conducted in a Kugelrohr apparatus at 180°C at 0.1 mm pressure. After 6 hrs, a white crystalline sublimate was collected. The ^ H— NMR analysis of this solid showed no aliphatic hydrogens and indicated it to be 4-phenylpyr idine . Elemental analysis and melting point data further confirmed this contention. However, the yield was a modest 25%. One possible mechanistic explanation is that in the canonical form 4.33a,

PAGE 105

96 the dihydropyridine might have induced a 1,7-hydride shift to form the pyridinium zwitterion intermediate 4 . 35 , which then underwent a reverse-Michael process to eliminate 4-phenylpyr idine (Scheme 4.16). Ph Ph Scheme 4.16 Partial support for this reasoning came from the mass spectral fragmentation data of 4.33 ( M + = 251). The M + -l fragment ion at m/e 250 (65%) indicates facile oxidation to the pyridinium cation through the loss of a hydrogen. Further degradation to the 4-phenylpyr idine fragment (m/e 155, base peak) was also observed together with the protonated form of 4-phenylpyr idine (m/e 156, 40%). However, the cyclohexenonyl fragment (m/e 95) showed only a 4% intensity.

PAGE 106

97 The vulnerability toward oxidation to the pyridinium cation shown by the dihydropyridine 4.33 would certainly be expected to discourage electrocyclic ring-opening of the dihydropyridine ring. Thus it became essential at this point to assess the role of the keto-function of 4.33 in its pyrolytic behavior. It was thought that the absence of such an electron sink would minimize the delocalization of the nitrogen lone pair and thus stabilize the dihydropyridine against oxidation and further fragmentation. This might allow sufficient lifetime for the electrocyclic ring-opening to occur. To test this hypothesis, we needed to prepare a 1 , 2-dihydropyr idine with a simple N-cycloalkenyl substituent . 4.2.5. Preparation of N( Cycloalken-l-yl ) pyridinium Salts A general outline for the preparation of N-cycloalken-l-ylpyr idinium salts was shown before as Path B in Scheme 4.12. This required the preformation of N-cyclohexylpyr idinium salts suitably substituted with a leaving group at the (3-position of the cyclohexyl ring. We thus started with the known [56JA2527] N( e-hydroxycyclohexyl ) pyridinium perchlorate (4.36) which was readily prepared by the addition of pyridine to

PAGE 107

98 cyclohexene oxide in the presence of perchloric acid (Scheme 4.17). However, the hydroxyl group in 4.36 showed some unusual resistance towards thionyl chloride [83JOC4017] in attempts to convert it to the chloride. Dehydration of 4.36 with perchloric acid, p-toluenesulfonic acid (refluxing toluene with azeotropic removal of water) or with methylt r iphenoxyphosphonium iodide [ 72 JOC4190 , 8 4 JCS ( PI ) 1 9 3 3 ] also failed. When the dehydration was attempted in boiling acetic anhydride, only the corresponding acetate (4.37) was obtained . (4.37) Scheme 4.17

PAGE 108

99 Since the hydroxy group of 4.36 could not be transformed to the chloride, we resolved to prepare (3-bromocyclohexylpy r idinium salts, taking advantage of the fact that bromination of cyclohexene in presence of excess pyridine affords 6-bromocyclohexylpyridinium bromide [ 50 JA4524 ] . Although the yield of this reaction was reportedly low (20-40%), the ready availability of pyridines and the convenience of large scale preparation made it an attractive alternative. Thus, following this procedure, the novel pyridinium salts (4.38b-c) were easily prepared in yields ranging from 21 to 47% (Scheme 4.18). The stereochemistry of cyclohexane ring-substituents in 4.38 were tentatively assigned as "trans" [50JA4524]. Scheme 4.18

PAGE 109

100 After an initial failure to dehydrobrominate the pyridinium salts (4.38) with NaOH [83JOC4017], it was found that treament with pyridine in refluxing nitromethane effects the dehydrobromination to yield the N-cyclohexenylpyridinium salts (4.39) (Scheme 4.19). Rather harsh conditions were necessary as the a-hydrogen and the g-bromine in the salts 4.38 are "syn" to each other and elimination via an E 2 transition state would be disfavored. We believe that, here the elimination of HBr proceeded through an E^CB mechanism. (4.38) (4.39) a; R 1 ii so OJ = K, r 2 = CH b; R 1 = R 3 = ch 3 , r 2 c ; R 1 = R 3 = H, R 2 = Ph Scheme 4.19 Formation of the N-cyclohexenyl linkage was supported by the 1 H-NMR spectra of the salts 4.39. Each of these salts showed a characteristic one-proton multiplet around 8 6.6,

PAGE 110

101 which was assigned to the vinyl proton of the cyclohexene ring. The set of multiplets at 6 4. 8-5. 2, characteristic of the a and 8 protons of the cyclohexyl ring of the starting pyridinium salts (4.38), disappeared. The N-vinyl salts (4.39a and 4.39b) were highly hygroscopic. Anion exchange to the more stable tetraf luroborate (with HBF^ , AgBF^ , NaBF^ ) or to the perchlorate (with NaClO^ ) was unsuccessful. The bromide 4.39c was comparatively stable and easier to handle. The dicyclopentanoid pyridine (4.32) failed again to react with cyclohexene oxide or with cyclohexene-bromine mixture in the desired fashion. Further investigations toward a N-cycloalkenylpyr idinium salt incorporating a tricyclic pyridine skeleton was thus abandoned. 4.2.6. Preparation of N( Cycloalken-l-yl ) -1 , 2dinydropyr idines It is very well documented that borohydride reductions of pyridinium salts are usually accompanied by over reduction to the tetrahydropyridines [ 8 6AHC (39)1] . This evidently occurs via the formation of an intermediate iminium salt (4.40) which is produced by the uptake of a proton (from the solvent) by the incipient 1 , 2-dihydropyr idine (Scheme 4.20).

PAGE 111

102 (4.40) NaBH . 4 < I R Scheme 4.20 Apart from the problem of overreduction, the question regarding the 1,2vs . 1,4-attack of the hydride on the pyridinium ring needs to be adressed, also. By selecting a 4-picoline moiety, as in 4.39a, the hydride attack could be directed exclusively at the 2-position; the same argument holds true for the 4-phenylpyr idinium salt 4.39c. The 3 , 5-lutidine system (4.39b) was chosen so that, in the corresponding 1 , 2-dihydropyridine , the methyl groups would offer some resistance towards protonation of the enamine system. Thus, formation of an iminium ion ( cf . 4.40) would be retarded, as would be further reduction. To test these

PAGE 112

103 hypotheses and arrive at an optimum system, sodium borohydride reduction of the salts 4.39(a-c) was undertaken. An extensive survey of the literature also revealed that certain experimental conditions, when properly adopted, can minimize the problem of over-reduction. The method chosen for this study was the "biphasic" reduction mode developed by Kutney and co-workers [77H593], By this method, the reduction is carried out in a biphasic mixture of benzene and aqueous alkali. Once the dihydropyr idines were formed they were immediately extracted into the organic phase, thus minimizing further reduction. When the pyridinium salts (4.39a-c) were reduced with one equivalent of NaBH^ under a biphasic condition [^OtpH 10)-ether], the preferential formation of the 1 , 2-dihydropyridines was observed. The "'‘H-NMR spectrum of the crude product derived from 4.39a showed characteristic peaks for the 1 , 2-dihydropyr idine (4.41) (Scheme 4.21). It showed a 1H doublet (J = 6 Hz) at 6 6.5 (for H-6 ) , a 1H 2 multiplet at 6 5.1 (cyclohexene sp hydrogen) and a narrow 2H doublet (J = 4 Hz) at 6 4.2 (for H-2 ) . The H-3 and H-5 hydrogens of the dihydropyridine ring formed a multiplet at 6 4. 7-5.0 while the saturated hydrogens of the cyclohexene ring clustered at S 1.5-2. 5.

PAGE 113

104 (4.39a) (4.41) Scheme 4.21 Under similar reaction conditions, the pyridinium salt 4.39b was also reduced, but gave a mixture of 1,2and 1 , 4-dihydropyr idines (4.42 and 4.43) (Scheme 4.22). CH. Br NaBH 4 , 2 5 0 C > ether-water + Scheme 4.22

PAGE 114

105 The ^H-NMR of the reaction product from 4.39b showed one-proton singlets at 6 6.3 and 4.5, a 1H multiplet at 5 5.7 together with a 2H singlet at 6 4.0. These signals were assigned to the 1 , 2-dihydropyr idine (4.42). Also present were a singlet at 6 6.1 (2H), a multiplet at 6 5.1 (1H) and another singlet at 6 2.8 (2H). These latter peaks indicated the presence of the 1 , 4-dihydropyr idine (4.43). From the spectral integral, the ratio of 4.42 to 4.43 was found to be 6 : 1. However, these dihydropy r idines (4.41-4.43) were extremely unstable and attempts at their purification (flash chromatography, neutral alumina) met with failures. On the other hand, the 4-phenylpy r idinium salt 4.39c was reduced by sodium borohydride to give a more stable 1 , 2-dihydropyridine (4.44) (Scheme 4.23). In this case, the reaction did not require the "biphasic" conditions and may be carried out in methanol-acetonitrile mixture at 0°C; the dihydropyridine (4.44), once formed, precipitated out of the medium and thus escaped further reduction. NaBH 4 , 0°C > CH 3 OH-CH 3 CN (4.44) Ph (4.39c) Scheme 4.22

PAGE 115

106 The structure of the dihydropyridine 4.44 was confirmed by its H and C-NMR spectra. The 1 H-NMR spectrum of 4.44 is illustrated in Figure 4.1. The one proton doublet (J = 8 Hz) at 6 6.5 is due to the H-6 hydrogen whereas the complex multiplet at 6 5.5 was assigned to the H-3 hydrogen. The H-5 proton gave rise to the doublet of doublets (ABX pattern) at 5.1 ppm: the higher coupling ( J = 8 Hz) is due to the H-6 hydrogen while H-3 splits the H-5 proton with a smaller coupling constant (J = 2 Hz). The remaining doublet (J = 4 Hz) at 6 4.3 was assigned to the two H-2 protons and the multiplet at 6 4.6 to the vinyl proton of the N-cyclohexenyl r ing . The 13 C-NMR spectrum of 4.44 is shown in Figure 4.2A. Five relatively upfield resonances were observed, of which the most downfield signal at 6 46.8 was assigned to the C-2 of the 1 , 2-dihydropyridine nucleus. Figure 4.2B shows the C-NMR spectra of the same dihydropyridine ( 4 . 44 ) run under the APT (Attached Proton Test) pulse sequence [81CC150, 82JMR535]. The peaks above the baseline signify the CH 2 and the quarternary carbons while those below are due to the CH^ s and CH's. Assignments of the dihydropyridine ring-carbons was difficult, not only due to lack of model spectra but also because of the overlapping signals from the 4-phenyl group. However, the spectra can clearly account for the requisite number of carbons present in 4.44.

PAGE 116

107 c*. | 1 H-NMR Spectrum of the 1 , 2-Dihydropyridine 4.44. Figure 4.1.

PAGE 117

108 < * (i s i !/ i I 5 I * i Hi ! i "i I I I ‘ * ’ ' 140 \20 «* raa. 1 I 1 1 ' I 100 M I I 60 40 "P" ?0 * Figure 4.2. Proton Decoupled C-NMR Spectra of the 1 , 2-Dihydropyridine 4.44. A) Normal Spectrum B) Under APT Mode

PAGE 118

109 4.2.7. Pyrolysis of N(Cycloalken-l-yl)-4phenyl-1, 2dihydropyr idine Pyrolysis of the more stable dihydropyridine ( 4 . 44 ) was carried out at 180°C under 0.1 mm. After 4 hrs a white crystalline sublimate was collected. Unfortunately, it again turned out to be 4-phenylpyr idine (matching NMR and melting point data), albeit in low yield (20%). The mass spectral fragmentation pattern for 4.44 is supportive of radical fragmentation. Initial loss of hydrogen gives rise to the pyridinium fragment (m/e 236) which constitutes the base peak. Other prominent peaks are at m/e 156 (16%) and 155 (14%) from the protonated 4-phenylpyr idine and free base fragments, respectively. 4.3. Conclusions The thermolytic rearrangement shown by acyclic N-vinyl-1 , 2-dihydropy r idines could not be extended to their cyclic counterparts. Considering that cyclic N-vinyldihydropyridines should obey the same ring opening-ring contraction sequence as proposed for the acyclic ones, the following set of transformations (Scheme 4.24) should hold true for 4.44.

PAGE 119

110 Scheme 4.24 In Scheme 4.24, the intermediate 4.47 appears to have a highly strained sp^ carbon. Thus, even if the tetraene 4.45 is formed, it probably reverts back to the dihydropyridine, only to be oxidized to the corresponding pyridine. In fact, the mass spectra of the dihydropyr idines (4.33) and (4.44) provided strong indications for this. Unfortunately, the mass spectral fragmentation of the acyclic 1 , 2-dihydropyridines (4.18) were not available for direct comparision .

PAGE 120

Ill 4.4. Experimental Melting points were determined on a Bristoline hot-stage microscope and are uncorrected. ‘'"H-NMR spectra were recorded either on a Varian EM 360L (60 MHz) or a Varian XL-200 (200 MHz) instrument. Chemical shifts are reported in S (ppm) values downfield from Me^Si unless specified otherwise. 13 C-NMR spectra were taken either on a JEOL FX-100 (25 MHz) or a Varian XL-200 (50 MHz) spectrometer with CDCl^ (77.0 ppm) or (CD^^SO (39.5 ppm) as reference. IR spectra were recorded on a Perkin Elmer 283B instrument. Mass spectra were obtained at 70 ev on a AEI MS 30 mass spectrometer connected to a DS 55 data system. Elemental analyses (C, H, N, F) were carried out by Dr. R. w. King of this department. Removal of solvent under reduced pressure refers to solvent evaporation on a rotary evaporator connected to a water aspirator. All solvents were distilled over appropriate drying agents under nitrogen prior to use. The following compounds were prepared according to the known literature procedures: 3-Chlorocyclohex-2-ene-l-one , b.p. 62-63 °C/4 mm (lit. [76JOC636] b.p. 63°C/4 mm); l , 2 , 3 , 5 , 6 , 7-Hexahydrodicyclopentanoid [ b,e] pyridine (4.32), m. p. 8 5-87 0 C (lit. [ 7 1 JPS ( A1 ) ( 9 ) 1 8 0 7 ] m.p. 87°C); 1( 2-Hydroxycyclohexyl ) pyridinium perchlorate (4.36), m.p. 118-119 °C (lit. [56JA2527 ] m.p. 122°C);

PAGE 121

112 l-[3-(Cyclohex-2-ene-l-on)-yl j-4-dimethylaminopyridinium chloride ( 4 . 31b) . 4-Dimethylaminopyr idine (0.94 g, 7.7 mmole) and 6-chlorocyclohexenone (1 g, 7 mmole) in dry benzene (10 ml) was heated under reflux for 2 hrs. The precipitate formed was filtered, washed with ether and dried. Recrystallization from ethanol-ether gave the pyridinium chloride (4.31b) as white microcrystals (1.7 g, 94%), m.p. 120-122°C (Found: C, 61.60, H, 6.90, N, 11.42. C 13 H 17 ClN 2 0 requires C, 61.77, H, 6.78, N, 11.08%); 6^ (D 2 0, DSS as internal standard) 8.5 (2H, d, J 7 Hz), 7.2 (2H, d, J 7 Hz), 6.5 (1H, s), 3.5 (6H, s ) , and 3 . 2-2 .2 ( 6H, m) . 1~ [ (Cyclohex-2-ene-l-on)-3-yl ] -4-phenyl pyridinium chloride (4.31c). 4-Phenyl pyridine (4.7 g, 30 mmole) and 0-chlorocyclohexenone (4 g, 30 mmole) in dry benzene (25 ml) was heated under reflux for 36 hrs. The precipitate formed was filtered, washed with ether and dried (4.8 g, 57%), m.p. 191-192 °C . 5 h ( DMSO-dg ) 9.2 (2H, d, J 7 Hz), 8.5 (2H, d, J 7 Hz) 8. 2-7. 6 (5H, m), 6.7 (1H, s), 3. 3-3. 2 (2H, m) , and 3.02.3 (4H, m). For analytical purpose, the tetraf luoroborate salt (prepared by anion exchange with a slight excess of sodium tetraf luoroborate in water) was used. Recrystallized

PAGE 122

113 from ethanol as needles, m.p. 223-225°C (Found: C, 60.82, H, 4.62, N, 4.02. C^H^gBF^NO requires C, 60.56, H, 4.78, N, 4.15%) . (^Cyclohex-2-ene-l-on)-3-yl ]-4-phenyl-l, 2-dihydropy r idine l-[ (cyclohex-2-ene-l-on)-3-yl] -4-phenylpyr idinium chloride (4.31c) (2 g, 7 mmole) dissolved in acetonitrile-methanol (2:1, 25 ml) was cooled to 0°C. Sodium borohydride (0.27 g, 7 mmole) in water (5 ml) made al’kaline (pH 10) with 10% sodium hydroxide was added dropwise and the mixture stirred for 2 hrs. Water (50 ml) was then added to give a yellow precipitate which was washed with cold water (2 X 10 ml) and dried under vacuum over phosphorus pentoxide (1.7 g, 91%); m.p. 135-139°C; HRMS : M + * found, 251.1287. C 17 H 17 NO rec 3 uires 251.1310 a.m.u.; & H (CDCl 3 ) 7.6-7. 4 (5H, m) 6.6 (1H, d, J 8 Hz), 5. 8-5. 4 (2H, m), 5.3 (1H, s), 4.4 (2H, d, J 4 Hz), and 2. 5-1. 8 (6H, m) ; (CDC1 3 ) 197.0 (C=0), 158.9, 137.9, 132.9, 128.4, 127.6, 126.7, 125.0, 114.6 ( C— 3 ) , 104.7 (C-5), 102.9 ( C-2 '), 47.3 (C-2), 36.0, 25.9, and 21.9. Pyrolysis of 1[ ( Cyc lohex-2-ene-l-on ) -3-vl ] -4 -phenyl -1 , 2di hydropyridine (4.33). ~ “ L * The 1 , 2-dihydropyridine (4.33) (0.5 g, 2 mmole) was heated for 6 hrs at 180°C under 0.1 mm of Hg in a Kughelrohr

PAGE 123

114 apparatus. The white sublimate formed proved to be 4-phenyl pyridine (0.08 g, 25%); m.p. 68-70°C (lit., m.p. 6 9 7 0 0 C ) ; S H (CDC1 3 ) 8.8 (2H, d, J 8 Hz) and 7. 8-7. 4 (7H, m) . N( 2 -Ace toxy cyclohexyl ) pyridinium perchlorate (4.37). N( 2-Hydr oxy cyclohexyl ) py r idinium perchlorate (4.36) (2.0 g, 7.2 mmole) in acetic anhydride (10 ml) was heated under reflux for 24 hrs. After cooling the solution to room temperature, acetone (20 ml) was added and the whole was added dropwise to dry ether (100 ml). The precipitate which formed was filtered, washed with ether and was recrystallized from ethanol (2.3 g, 98%); m.p. 198-200°C (Found C, 48.83, H, 5.72, N, 4.20. C^H^gClNOg requires C, 48.83, H, 5.67, N, 4.38%); 6 H (DMSO-dg) 9.3 (2H, d, J 8 Hz), 9. 0-8. 6 (1H, m), 8. 5-8. 2 (1H, m) , 5. 5-4. 6 (2H, m) and 2.51.2 ( 11H, m) . General Method for the Preparation of N( 2-B r omocyclohexyl ) Pyridinium Salts (4.38). Bromine (9.65 g, 60 mmole) was added dropwise to an ice cold mixture of cyclohexene (5.0 g, 60 mmole) and the appropriate pyridine (250 mmole) (in 25 ml of CHClg when 4-phenyl pyridine was used) over a period of 30 min. After stirring at room temperature for a further 30 min the

PAGE 124

115 reaction mixture was triturated with ether (200 ml). The precipitate formed was filtered, washed with ether (2 X 25 ml), dried and crystallized from acetone-methanol to give white needles. The following pyridinium salts were made by this procedure: l-( 2-Bromocyclohexyl ) -4-methylpyr idinium Bromide (4.38a). 4-Methylpyridine was used as the pyridine component. Yield : 7.8 g (40%); m.p. 190-192 °C ( Found : C , 42.70, H, 5.02, N, 4.04. C 12 H 17 B r 2 N requires C , 43.01, H, 5.11, N, 4.18%); 6 h ( DMSO-dg ) 9.2 ( 2H, d, J 8 Hz), 8.3 (2H, d, 00 ^1 Hz) 5. 1-4. 8 (2H, m), 2.9 ( 3H , s ) and 2. 8-1. 5 (8H, m). 1( 2-Bromocyclohexyl )3, 5-dimethyl pyridinium Bromide (4.38b) . 3 , 5-Dimethylpyr idine was used as the pyridine component. Yield : 7.6 g (37%); m.p. 186-188°C (Found: C, 44.50, H, 5.46, N, 3.85. C^H^B^N requires C, 44.72 , H, 5.48, N, 4.01%); (D 2 <0, DSS as internal standard) 9.1 (2H, s), 8.7 (1H, s), 5. 2-4. 9 ( 2H , m) and 2. 8-1. 6 (14H, m). 1( 2-Bromocyclohexyl )-4-phenyl pyridinium Bromide (4.38c). 4-Phenylpyr idine was used as the pyridine component. Yield : 5.0 g (21%); m.p. 209-211°C (Found: C, 51.34, H, 4.85, N, 3.39. C 17 H ig Br 2 N requires C, 51.41, H, 4.82, N,

PAGE 125

116 3.52%); S H ( DMSO-dg ) 9.6 (2H, d, J 8 Hz), 8.9 (2H, d, J 8 Hz), 8. 6-8. 3 (2H, m), 8. 0-7. 7 (3H, m), 5. 4-5.1 (2H, m ) and 2. 8-1. 6 (8H, m). General Procedure for the Preparation of N( Cyclohex-l-enyl ) Pyridinium Salts (4.39). The appropriate 1( 2-Bromocyclohexyl ) py r idinium salt (4.38) (7 mmole) and pyridine (3 ml) in nitromethane (25 ml) were heated under reflux for 32 hrs. After removing all the volatile matter under reduced pressure, the residue was dissolved in CH 2 CI 2 (25 ml) and treated with NH^ gas for 5 min. The precipitate formed was centrifuged and ethyl acetate (100 ml) added to the CH 2 CI 2 layer forming a precipitate. The supernatent liquid was carefully decanted and the residual solid extracted with boiling acetone (30 ml). Addition of ether (100 ml) to the acetone extract reprecipitated the desired products which were filtered, washed thoroughly with ether and dried (60°C at 15 mm of Hg). The following pyridinium salts were made by this procedure : 1(Cyclohexen-l-yl ) -4-methylpy r idinium bromide (4.39a). From 1( 2-B romocyclohexyl ) -4-methylpy r idinium bromide (4.38a). Yield : 1.25 g (70%); 5^ (D 2 O, DSS as internal standard) 9.1 (2H, d, J 7 Hz), 8.3 (2H, d, J 7 Hz), 6.6 (1H, m), 2.8 (3H, s) and 2. 7-1. 6 (8H, m).

PAGE 126

117 l-(Cyclohexen-l-yl)-3,5-dimethylpyridinium bromide (4.39b). From 1( 2 -Bromo cyclohexyl )-3,5-dimethylpyridinium bromide (4.38b). Yield : 1.3 g (70%); (D 2 0, DSS as internal standard) 8.9 (2H, s), 8.6 (1H, s), 6.5 (1H, m) and 2. 8-1. 6 (14H, m). l-(Cyclohexen-l-yl)-4-phenylpyridinium bromide (4.39c). From l-( 2-Bromocyclohexyl ) -4-phenylpy r idinium bromide (4.38c). Yield : 2.2 g (90%); § H ( DMSO-dg ) 9.5 (2H, d, J 8 Hz), 8.8 (2H, d, J 8 Hz), 8. 7-8. 3 (2H, m), 8. 1-7. 8 (3H, m), 6.7 (1H, m) and 2. 9-1. 6 (8H, m). Sodium Borohydride Reduction of N( Cyclohehexen-l-yl ) pyridinium Salts (4.39). A solution of NaBH^ (0.02 g, 0.5 mmole) in water (5 ml) (adjusted to pH 10 with 10% aqueous NaOH) was added dropwise to a suspension of the appropriate N-cyclohexenylpyr idinium bromide (4.39) (2 mmole) in ether (15 ml) at 0°C. After 3 hr at room temperature, water (10 ml) was added and the ether layer separated. The aqueous layer was extracted with ether (10 ml). The combined ether fractions was washed with brine (10 ml) and dried (Na 2 SO^). Removal of solvent under reduced pressure gave the crude N-cyclohexenyldihydropyr idines which were characterized through their ^H-NMR spectra. The following dihydropyr idines were prepared according to this general procedure:

PAGE 127

118 l-(Cyclohexen-l-yl)-4-methyl-l > 2-di hydropyridine (4.41). Prepared from 1( Cyclohex-l-yl ) -4-methylpyr idinium bromide (4.39a) as a yellow oil. Crude yield : 0.25 g (71%); S H (CDC1 3 ) 6.5 (1H, d, J 8 Hz), 5.1 (1H, m) , 4. 5-4. 9 (2H, m), 4.2 ( 2H , d, j 3 Hz) and 2. 5-1. 4 (11H, m). 1( Cyclohexen-l-yl ) -3 , 5-dime thy 1-1 , 2-dihydropyr idine (4.42) and^l-(Cyclohexen-l-yl)-3,5-dimethyl-l, 4-dihydropyr idine From 1( Cyclohexen-l-yl ) -3 , 5-dimethylpyr idinium bromide (4.39b) as a yellow oil. Crude yield of the mixture : 0.28 g (70%); characteristic S R (CDC1 3 ) for (4.33) : 6.3 (1H, s), 5.7 (1H, m) , 4.6 (1H, m) and 3.9 (2H, s); characteristic S„ H (CDC1 3 ) for (4.34) : 6.1 (1H, s), 5.1 (1H, m) and 2.7 (2H, s ) . Preparation of 1(Cyclohexen-l-yl ) -4 -phenyl1 , 2dihydropyr idine (4.44). NaBH 4 (0.02 g, 0.5 mmole) in water (3 ml) (adjusted to pH 10 with 10% aqueous NaOH) was added dropwise at 0°C to a suspension of 1( Cyclohexen-l-yl ) -4-phenylpyr idinium bromide (4.39c) (0.67 g, 2 mmole) in CH^CN (5 ml). After 2 hrs at room temperature, water (15 ml) was added and the yellow precipitate was filtered, washed with water and crystallized from CH^CN to give yellow needles (0.42 g, 85%), m.p. 138-140°C (decomp.) (Found: C, 86.12, H, 8.15, N, 5.83. C 17 H 19 N rec J uires c > 86.03, H, 8.07, N, 5.90%). & H (DMSO-dg)

PAGE 128

119 7. 5-7. 2 (5H, m) , 6.5 (1H, d, J 8 Hz), 5.5 (1H, m) , 5.1 (1H, dd, J 8 Hz, J 2 Hz), 4.6 (1H, m), 4.2 (2H, d, J 4 Hz) and 2. 4-1. 4 (8H, m). S_ (DMSO-d c ) 138.9, 138.1, 134.2, 129.3, 127.7, 126.5, 124.7, 109.6 (C-3), 100.4 (C-5), 95.5 (C-2'), 46.8 (C-2), 25.6, 23.9, 22.5 and 21.8. Pyrolysis of 1( Cyclohexen-l-yl ) -4-phenyl-l , 2dihydropyr idine ( 4 . 44 ). The 1 , 2-dihydropyr idine ( 4 . 44 ) (0.5 g, 2 mmole) was heated at 160°C in a Kughelrohr apparatus under 0.1 mm of Hg. After 4 hrs the white sublimate was collected which proved to be 4-phenylpyr idine (0.06 g, 20%); m.p. 69-71°C (lit. m.p. 69-70 0 C ) ; (CDCl 3 ) 8.7 (2H, d, J 8 Hz) and 7.87.4 ( 7H , m ) .

PAGE 129

CHAPTER V THERMAL BEHAVIOR OF 1 , 2-DIHYDROPYRIDINES WITH UNSATURATED 4-SUBSTITUENTS 5.1. Introduction The thermal behavior of 1 , 2-dihydropyr idines was discussed in chapter IV. A striking feature was their el ec t r ocycl i c ring opening to 1-azatrienes . We were interested in the potential utilization of these highly reactive 1-azatrienes in non-degenerate electrocyclic processes. It was envisaged that thermal electroreversion of a 4-vinyl-l , 2-dihydropyridine would produce a cross-conjugated tetraene (Scheme 5.1). The latter showed promise of undergoing interesting electrocyclic processes. Scheme 5.1 120

PAGE 130

121 5.1.1. Aim of the work Thermal ring opening of a 4-vinyl-l , 2-dihydropyr idine (5.1) would lead to the cross-conjugated tetraene (5.2a). Electrocyclization of tetraene 5.2a in its isomeric form 5.2b would afford a new dihydropyridine (5.3) in which the elements X and Y of the original 4-vinyl substituent are now incorporated in the ring (Scheme 5.2). However (5.3) is itself a 4-vinyl-l , 2-dihydropyridine derivative and its thermal behavior would be expected to be very similar to (5.1) and the sequence shown in Scheme 5.2 can therefore be considered as an equilibrium process. R (5.1) (5.2a) (5.2b) l < R (5.3) Scheme 5.2 Such an equilibrium may be driven forward if sufficient stability be imparted to 5.3 so as to prevent its cycloreversion. The stability of 5.3 would be highly dependent on the 4-substituent in 5.1 since the elements of the latter become part of the ring-skeleton in 5.3. Thus the

PAGE 131

122 success of the proposed transformation rested on the proper choice of 4-substituents in 5.1. A cummulative double bond at the 4-position of 5.1 appeared ideally suited for this purpose. On this basis, a model dihydropyridine (5.4) can be drawn which shows the essential features required for our starting material. The cummulated function could either be an isocyanate group (as in 5.5) or a ketene function (as in 5.7) which would eventually lead to the 2-pyr imidinone (5.6) and the 2-pyridone (5.8) derivatives, respectively (Scheme 5.3). Both 2-pyr imidinone and 2-pyridone posess certain degree of aromaticity and are more stable than 1 , 2-dihydropyr idines . Thus the systems (5.5) and (5.7) were chosen to study the proposed set of thermal transformations. R (5.4)

PAGE 132

123 (5.7) (5.8) Scheme 5.3 5.2. Results and Discussion 5.2.1. Synthetic Strategy As a possible route to the desired 1 , 2-dihydropyridines , NaBH 4 reduction of the corresponding pyridinium salts was considered. It was well anticipated that a selective reduction of the pyridinium ring in 4-i socyanatopyr idinium salts would be far from simple. Therefore, a suitable isocyanate precursor was needed which would be inert to

PAGE 133

124 NaBH^ reduction. A logical choice was the carbamate function which is one of the known precursors to isocyanates and is also inert to NaBH^ . This allowed for a convenient retrosynthesis of the 4-i socyanato-1 , 2-dihydropyr idine (5.5) as shown in Scheme 5.4. (5.5) Scheme 5.4 Based on similar arguments, a methyl acetate group was chosen as a possible precursor to the ketene function for 4-keteno-l , 2-dihydropyridine (5.7). A possible retrosyntheti c route for the latter is shown in Scheme 5.5. (5.7) Scheme 5 . 5

PAGE 134

125 5.2.2. Preparation of 4-Ethoxycarbonylaminopyridinium Salt We first studied the synthetic route to the 4-i socyanatopyr idinium salt. We started with ethyl 4-pyr idylcarbamate (5.9) which was easily prepared from 4-aminopyr idine according to literature procedure [ 58CA( 52 ) 18475g , 62JCS2379]. The corresponding methiodide (5.10) was prepared in 98% yield by treatment with excess Mel in refluxing ethanol (Scheme 5.6) ClC0 2 Et > Et 3 N Mel > NHCC> 2 Et "VV I + Me (5.9) (5.10) Scheme 5.6 5.2.3. Attempted Sodium Borohydride Reduction of 4 -Ethoxy ca r bony 1 ami nopyr idinium Salt Even under strenuous conditions (refluxing MeOH ) , NaBH^ failed to reduce the pyridinium salt (5.10). This may be attributed to the electron releasing effect of the amino function at the 4-position of the pyridinium ring. In order to diminish this effect, a second electron withdrawing group on the 4-amino function was considered. However the desired 4( bis -ethoxy carbonyl) amino pyridine could not be prepared as the sodio-salt of 5.9 failed to react with diethyl carbonate or ethyl chloroformate .

PAGE 135

126 5.2.4. Preparation of 4-Methoxycarbonylmethyl-l-methyl pyrldinium Salt Turning to the 4-keteno series, we started with methyl 4-pyr idylacetate (5.11) [55JCS2586] which was prepared by the acid catalyzed esterification of the commercially available 4-pyridylacetic acid. Treatment with Mel in refluxing ethanol produced the corresponding methiodide (5.12) [64AJC455] in 70% yield (Scheme 5.7). ch 2 co 2 h N ch 3 oh H Mel (5.11) Me (5.12) Scheme 5.7 5.2.5. Sodium Borohydride Reduction of 4-Methoxycarbonylmethyl-l-methylpyr idinium Salt Sodium borohydride reduction of the pyridinium salt (5.12) was attempted in order to prepare the desired 1 , 2-dihydropyridine (5.13). Reduction in methanol, at -78°C, afforded an oil whose ^H-NMR spectrum gave no indication for the presence of a 1 , 2-dihydropyridine moeity. Reduction under biphasic conditions (CH 2 C1 2 -H 2 0) [77H593] also gave

PAGE 136

127 the same result. Based on its 1 H-NMR data, a tetrahydropyridine structure (5.14) was assigned to this product (Scheme 5.8). Me (5.12) NaBH 4 !-//—> (5.13) V Me (5.14) Scheme 5.8 The H-NMR of (5.14) showed a one proton multiplet at 6 6.1 assigned to the vinylic hydrogen while a complex multiplet (4H) at 2.8 3.1 ppm was assigned to the C-5 methylene and the exocyclic methylene protons. The methyl ester produced a singlet at 3.6 ppm (3H); an unresolved multiplet at 5 2.1 2.7 accounted for the remaining aliphatic hydrogens and the N-CH 3 .

PAGE 137

128 5.2.6. Preparation of l-Methoxycarbonyl-4metnoxycar bony lme thy 1-1, 2-dihydropyr idine Sodium borohydride reduction of N-alkylpyr idinium salts, most often, are accompanied by over-reduction to the tetrahydropyridines [ 8 6AHC (39)1]. Since no easy solution exists, we turned our attention to Fowler's method [72JOC1321] of preparing 1 , 2-dihydropyr idines directly from the corresponding pyridines. Thus the previously described 4-me thoxycar bony lme thylpyr idine ( 5 . 11 ) was treated with NaBH^ (~65°C, MeOH) in presence of methyl chloroformate which produced the novel 1 , 2-dihydropyr idine ( 5 . 15 ) in 87% yield (Scheme 5.9). for for ch 2 co 2 ch 3 CH 2 C0 2 CH 3 NaBH 4 / CH 3 OH fi "k N > ClC0 2 Me, -65 °C 1 1 C0 2 ( 5 . 11 ) (5 . 15 ) Scheme 5.9 The H-NMR spectrum of (5.15) showed two doublets, one the H-2 protons at 6 4.4 (2H, J = 4 Hz) and the other the H-6 proton at 6 6.8 (J = 8 Hz). The H-3 and H-5

PAGE 138

129 protons formed a multiplet around 6 5. 1-5. 5. The methyl ester and the methyl carbamate produced a pair of singlets at S 3.7 and 3.8 whereas a broad singlet at 6 3.0 (2H) was assigned to the exocyclic methylene protons. This novel 1 , 2-dihydropy r idine (5.15) may be purified by flash chromatography (neutral alumina, CHCl^ eluent) and is stable upto 24 hrs at 0°C under nitrogen. 5.2.7. Thermolysis of l-Methoxycarbonyl-4methoxycarbonyl methyl-1, 2-dihydropyr idine Thermolysis of the 1 , 2-dihydropy r idine (5.15) was performed in chlorobenzene. After heating at 160°C in a sealed tube for 12 hrs, the ^H-NMR spectrum still showed the presence of unreacted dihydropyridine; the desired 2-pyridone derivative (5.16) could not be observed even after prolonged heating (24 hrs) at that temperature (Scheme 5.10). When the thermolysis was carried out at a higher temperature (200°C for 12 hrs), it only led to extensive decomposition. (5.15) (5.16) Scheme 5.10

PAGE 139

130 5.3. Conclusions The dihydropyridine 5.15 failed to undergo the proposed rearrangement. This was attributed to its inability to undergo ring-opening to the cross-conjugated aza-tetraene . No further variations were attempted. 5.4. Experimental Melting points were determined on a Bristoline hot-stage microscope and are uncorrected. ^H-NMR spectra were recorded on a Varian EM 360L instrument. Chemical shifts are reported in 6 (ppm) values downfield from Me^Si. Elemental analyses (C, H, N, F) were carried out by Dr. R. W. King of this department. Removal of solvent under reduced pressure refers to solvent evaporation on a rotary evaporater connected to an water aspirator. The following compounds were prepared according to known literature procedures: 4-Ethoxycarbonylaminopyridine (5.9), m.p. 127-129°C (lit. [ 58CA( 52)18475g, 62JCS2379] m.p. 129°C); 4-Me thoxyca rbonylme thylpy r idine (5.11), b.p. 72-73°C/0.5 mm (lit. [55JCS2586] b.p. 85-87°C/1.5 mm); 4-Methoxy carbonyl me thy 1-1-methylpyridinium iodide (5.12), m.p. 136-137°C (lit. [64AJC455] m.p. 141-142°C);

PAGE 140

131 4-Ethoxycarbonylamino-l-methylpyridinium iodide (5.10). 4-Ethoxy carbonyl amino pyridine (5.9) (5.0 g, 30 mmole) and methyl iodide (12.7 g, 90 mmole) in ethanol (15 ml) was heated under reflux for 12 hrs. On cooling to room temperature the methiodide (5.10) crystallized out which was filtered and washed with cold ethanol. Recrystallization from ethanol gave white plates (9.1 g, 98%); m.p. 189-190°C (Found: C, 34.83, H, 4.18, N, 8.81. CgH^I^C^ rec 3 u i res C, 35.08, H, 4.25, N, 9.09%); § H (DMSO-dg) 11.6 (1H, broad), 8.9 ( 2H , d, J 7 Hz), 8.0 (2H, d, J 7 Hz), 4. 5-4.1 (5H, m) and 1.3 ( 3H, t, J 9 Hz ) . Sodium Borohydride Reduction of 4-Methoxycarbonylmethyl-lmethylpyridinium iodide (5.12). At 0°C, a solution of NaBH^ (0.06 g, 2.5 mmole) in methanol (2 ml) was added to a solution of the methiodide (5.12) (2.93 g, 10 mmole) in methanol (10 ml). After 1.5 hrs at 0°C, water (20 ml) was added and the reaction extracted with ether (3 X 15 ml). The combined organic layer was dried ( Na 2 S 04 ) and removal of solvent under reduced pressure gave an oil (1.3 g, 80%); § H (CDC1 3 ) 5.7 (1H, m) , 3.7 (3H, s), 3. 1-2. 8 ( 4 H , m) and 2. 6-2.0 (7H, m).

PAGE 141

132 Methyl ( 1-methoxycarbonyl-l , 2-dihydropyr idin-4-yl ) acetate ( 5 . 15 ) . To a mixture of NaBH^ (0.38 g, 10 mmole) and methyl 4-py r idylacetate ( 5 . 11 ) (1.51 g, 10 mmole) in dry methanol was added slowly a solution of methyl chlorof ormate (0.94 g, 10 mmole) in dry ether (10 ml) at -78°C. After 3 hrs at -65°C it was added to crushed ice and extracted with (3 X 15 ml). The combined organic layer was washed with saturated NH^Cl solution (10 ml) followed by brine (10 ml) and dried (Na 2 SO^). Removal of solvent under reduced pressure gave an oil. This was purified by flash chromatography over a short column of neutral alumina (CHCl^ as eluent). The yellow oil thus obtained (1.83 g, 87%) was stored under nitrogen at 0°C ; § H (CDC1 3 ) 6.8 (1H, d, J 8 Hz), 5. 5-5.1 ( 2H , m), 4.4 (2H, d, J 4 Hz), 3.8 (3H, s), 3.7 ( 3H, s ) and 3.0 ( 2H, s ) .

PAGE 142

CHAPTER VI SUMMARY The effect of heteroatoms on an all-carbon skeleton was outlined in chapter I. The reactivity patterns of the a-carbons in amines and alcohols was discussed, with particular reference to the generation of carbanions at those centers. In addition, thermal electrocyclic ring-opening of 1 , 3-cyclohexadiene and its heteroatom modified derivative, 1 , 2-dihydropyr idine was compared. A preview of the present investigation was also provided. In chapter II, the concept of transient activation of an ot-sp^ center toward proton loss in secondary amines was discussed. In a novel approach, carbon dioxide was utilized as a protecting group for the a-metallation of indoline and tetrahydroquinoline . The lithium carbamates derived from these two amines, however, did not offer much a-activation ; nevertheless, a-metallation was achieved with a combination of potassium t-butoxide and t-butyl 1 i thium . The a-metallo species were trapped with aldehydes and ketones (other electrophiles gave poor results) and afforded the 133

PAGE 143

134 a-substituted amines directly after acidic work-up. The advantages of this procedure over the existing three step sequence are that the protecting group was easily introduced and deprotection occurred under very mild conditions (during acidic work-up); also, the whole set of operations was conveniently carried out in the same reaction vessel without the need to isolate any intermediates. Carbon dioxide was next applied as a hydroxyl protecting group for the generation of a-oxycarbanions which was described in chapter III. Although its protecting ability was successfully demonstrated, the a-activating power of the derived lithium carbonates was extremely small, if any. This called for the utilization of an easily removable a-activating auxilliary, a role which was aptly fulfilled by the trimethylsilyl group. Thus it was shown that the commercially available 1-trimethylsilylmethanol via its lithium carbonate can be successfully a-lithiated with s-butyllithium. The resulting a-lithio species reacted with esters (including a, P-unsaturated derivatives), amides, acid chlorides to give the corresponding a-hydroxy ketones after acidic work-up. Thus the combination of carbon dioxide protecting group and the trimethylsilyl auxilliary led to the development of a methanol dianion synthon, a species difficult to generate.

PAGE 144

135 In chapters IV and V, the thermal electrocyclic ring-opening of 1 , 2-dihydropyr idines was investigated. It was envisaged that the derived 1-azatrienes from N-vinylor 4-vinyl-l , 2-dihydropyr idines would participate in further electrocyclic processes to give interesting products. In order to study this, two novel N-cyclohexenenyl1 . 2dihydropyr idines were synthesized via the borohydride reduction of the corresponding pyridinium salts as was described in chapter IV. However the pyrolysis of these dihydropyridines did not afford any products resulting from their ring-opened form and instead, oxidation to the pyridine bases took place. For the 4-vinyl-series, the thermal behavior of the 4-methoxycarbonylmethyl1 . 2dihydropyr idine was studied which was discussed in chapter V. In this case also, the dihydropyridine was found to be highly resistant toward its thermal ring-opening and under drastic conditions, extensive decomposition occurred.

PAGE 145

BIBLIOGRAPHY The system adopted for references is the one used by Katritzky and Rees in "Comprehensive Heterocyclic Chemistry", Pergamon Press, New York, 1984, vol . 4, p. 1085. References are designated by a number-letter code of which the first two digits (or the first four digits for references before 1900) denote the year of publication, the next one or two letters the journal, and the final digits the page number. Books and all other sources are coded MI (miscellaneous) and are listed under the relevant year of publication . Letter Codes for Journal Titles Code Journal Abbreviation ACR Acc. Chem. Res. ACS ( B ) Acta Chim. Scand., Ser. B. AG ( E ) Angew. Chem., Int. Ed. Engl. AHC Adv . Heterocycl. Chem. 136

PAGE 146

137 AJC Austr. J. Chem. BSF Bull. Soc. Chim. France. CA Chem. Abstr. CB Chem. Ber. CC J. Chem. Soc., Chem. Commun. Cl ( L ) Chem. Ind . ( Lond . ) CJC Can. J. Chem. CL Chem. Lett. CRV Chem. Rev. CZ Chem. -Ztg. H Heterocycles HCA Helv. Chim. Acta. JA J. Am. Chem. Soc. JCS J. Chem. Soc. JCS( PI ) J. Chem. Soc., Perkin Trans. 1. JHC J. Heterocycl. chem. JMR J. Mag. Reson. JOC J. Org. Chem. JPS J. Polym. Sci. LA Liebigs Ann. MI Miscellaneous OM Or ganome tallies PAC Pure Appl . Chem. S Synthesis T Tetrahedron TL Tetrahedron Lett.

PAGE 147

138 REFERENCES 06CB3757 von Auwers, K. Chem. Ber. 1906, 39, 3757. 12LA( 394 ) 42 Wolff, L. Liebigs Ann. 1912, 394, 42. 50JA4524 Hayes, F. N.; Suzuki, H. K. and Peterson, D. E. J. Am. Chem. Soc. 1950, 72, 4524. 53BSF53 Panouse, J. J. Bull. Soc. Chim. France 1953, 53. 55JCS2586 Katritzky, A. R. J. Chem. Soc. 1955, 2586 . 56JA2527 Hayes, F. N.; King, L. C. and Peterson, D. E. J. Am. Chem. Soc. 1956, 78, 2527. 58CA( 52 )18475g Richter, C. and Sieber, P. Swiss Patent 324,439, 1957; Chem. Abstr. 1958, 52, P18 47 5g . 62HCA2426 Daeniker, H. U. and Druey, J. Helv. Chim. Acta 1962, 45, 2426. 62JCS1578 Baily, E. J.; Barton, D. H. R.; Elks, J. and Templeton, J. F. J. Chem. Soc. 1962, 1578 . 62JCS2379 Clark-Lewis, J. W. and Singh, R. P. J. Chem. Soc. 1962, 2379. 63JOC663 Jones, F. N. ; Zinn, M. F. and Hauser, C. R. J. Org. Chem. 1963, 28, 663. 6 3 JOC3461 Jones, F. N. ; Vaulx, R. L. and Hauser, C. R. J. Org. Chem. 1963, 28, 3461. 64AJC455 Katritzky, A. R. and Jones, R. A. Austr. J. Chem. 1964, 455. 64HCA33 Daeniker, H. U. Helv. Chim. Acta 1964, 47, 33.

PAGE 148

139 64 JCS3080 Lewis, K. E. and Steiner, H. J. Chem. Soc. 1964, 3080. 64TL1503 Schollkopf, H. and Kuppers, H. Tetrahedron Lett. 1964, 1503. 65TL4615 Lyle, R. E. and Gauthier, G. J. Tetrahedron Lett. 1965, 4615. 67JOC1479 Klein, K. P. and Hauser, C. R. J. Org. Chem. 1967, 32, 1479. 67PAC19 Corey, E. J. Pure Appl . Chem. 1967, 14, 19. 67T2775 Eisenthal, R.; Katritzky, A. R. and Lunt, E. Tetrahedron 1967, 23, 2775. 68 JOC3294 Gardner, J. N.; Carlon, F. E. and Gno j , 0. J. Org. Chem. 1968, 33, 3294. 68JOC3695 Gardner, J. N. ; Popper, T. L.; Carlon, F. E.; Gno j , 0. and Herzog, H. L. J . Org. Chem. 1968, 33, 3695. 68MI1 "The Chemistry of the Amino Group", ed. Patai, S., Interscience, London, 1968. 69CC833 Schwartz, J. J. Chem. Soc., Chem. Commun. 1969, 833. 69T4291 Katritzky, A. R. and Lunt, E. Tetrahedron 1969, 25, 4291. 7 0AG ( E ) 7 6 3 Schollkopf, U. Angew. Chem., Int. Ed. Engl. 1970, 9, 763. 70JOC2809 Fry, E. M. and Beisler, J. A. J. Org. Chem. 1970, 35, 2809. 71JA4027 Peterson, D. J. J. Am. Chem. Soc. 1971, 93, 4027. 71 JOC772 Lyle, R. E. and White, V, E. J. Org. Chem. 1971, 36, 772.

PAGE 149

140 71 JOC1607 Ludt, R. E. and Hauser, C. R. J. Orq. Chem. 1971, 36, 1607. 71JOC1705 van Bergen, T. J. and Kellog, R. M. J. Org. Chem. 1971, 36, 1705. 7 1 JPS ( AI ) ( 9)1807 Peebles, Jr., L. H. and Huffman, M. W. J. Polym. Sci., Part AI 1971, 9(7), 1807 . 71MI1 "The Chemistry of the Hydroxyl Group", ed. Patai, S., Inte r science , London, 1971. 71TL4395 Onaka, T. Tetrahedron Lett. 1971, 4395. 71TL4961 Giam, C. S. and Knaus, E. E. Tetrahedron Lett. 1971, 4961. 72CRV1 Eisner, U. and Kuthan, J. Chem. Rev. 1972, 12, 1. 72CZ411 Bestmann, H. J. and Ruppert, D. Chem. — Z tq . 1972 , 96, 411. 72JA5926 Fowler, F. W. J. Am. Chem. Soc. 1972, 99, 5926. 72JOC1321 Fowler, F. W. J. Org. Chem. 1972, 37, 1321. 72JOC4190 Hutchins, R. 0.; Hutchins, M. G. and Milewski, C. A. J. Org. Chem. 1972, 37, 4190. 73TL4193 Sakurai, H.; Nishiwaki, K. -I. and Kira, M. Tetrahedron Lett. 1973, 4193. 74ACR77 Brook, A. G. Acc. Chem. Res. 1974, 7, 77. 74ACR147 Evans, D. A. and Andrews, G. C. Acc. Chem. Res. 1977, 7, 147. 74CJC3563 Piers, E. and Soucy, M. Can. J. Chem. 1974, 52, 3563.

PAGE 150

141 74HCA1204 74JA3214 74JCS(PI )2496 74JOC59 74T4055 74TL59 7 5ACS ( B ) 6 5 5 7 5AG ( E ) 1 5 75CJC2305 75 JOC569 76ACS ( B ) 251 76CC339 76 JHC481 76JHC789 Greuter, H. and Schmid, H. Helv. Chim. Acta 1974, 57, 1204. Wright, A. and West, R. J. Am. Chem. Soc. 1974, 96, 3214. Acheson, R . ; Paglietti, G. and Tasker, P. J. Chem. Soc., Perkin Trans. 1 1974, 249V. Francis, R.; Davis, W. and Wisener, J. J. Org. Chem. 1974 , 39 _, 59. Schnekenbur ge r , J. and Heber, D. Tetrahedron 1974, _30.' 4055 . Thiessen, L. M . ; Lepoivre, J. A. and Alderwei reldt , F. C. Tetrahedron Lett. 1974, 59. Lounasmaa, M. and Johansson, C. Acta Chem. Scand . , Ser. B 1975 , 2_9, 6 55 . Seebach, D. and Enders, D. Angew. Chem., Int , Ed. Engl. 1975, 14 , 15. Giam, C.; Knaus, E.; Lockhart, R. and Keener, I. Can. J. Chem. 1975, 5_3, 2305. Finch, N. and Gemenden, C. J. Org. Chem. 1975, _40, 569. Lounasmaa, M. ; Johansson, C. and Svensson, J. Acta Chem. Scand., Ser. B 1976, 30, 251. Smith, J. G. and Sheepy, J. M. J. Chem. Soc., Chem. Commun. 1976, 339. Knaus, E. E.; Pasutto, F. M.; Giam, C. S. and Swinyeard, E. A. J. Heterocycl. Chem. 1976, 13, 481. Knaus, E.; Ondrus, T. and Giam, C. J . Heterocycl. Chem. 1976, .T3, 7 89.

PAGE 151

142 76JOC636 Clark, R. D. and Heathcock, C. H. J. Org. Chem. 1976, 41, 636. 76JOC3620 Still, W. C. and MacDonald, T. L. J. Org. Chem. 1976, 41, 3620. 76T1943 Lever, Jr. 0. W. Tetrahedron 1976, 1943 76TL47 Sturtz, G.; Corbel, B. and Paugam, J. P. Tetrahedron Lett. 1976, 4776TL4717 Sliwa, H. and Tartar, A. Tetrahedron Lett. 1976, 4717. 77CB1852 Seebach, D.; Enders, D. and Renger, B. Chem. Ber. 1977, 110, 1852. 77H593 Kutney, J. P. Heterocycles 1977, 7, 593 77JA5213 Beak, P. and McKinnie, B. G. J. Am. Chem. Soc. 1977, 99, 5213. 77S357 Grobel, B. and Seebach, D. Synthesis 1977, 357. 78CJC1026 Ondrus, T. A.; Knaus, E. E. and Giam, C S. Can. J. Chem. 1978, 56, 1026. 78CRV275 Beak, P. and Reitz, D. B. Chem. Rev. 1978, 78, 275. 78JA6696 Hasan, I. and Fowler, F.W. J. Am. Chem. Soc. 1978, 100, 6696. 78 JOC188 Vedejs, E.;. Engler, D. A. and Telschow, J. E. J. Org. Chem. 1978, 43, 188. 78JOC1599 Rubottom, G. M. and Gruber, J. M. J. Org. Chem. 1978, 43, 1599. 78JOC2923 Still, W. C.; Kahn, M. and Mitra, A. J. Org. Chem. 1978, 43, 2923. 78 JOC3617 Vohra, S. K.; Harrington, G. W. and Swern, D. J. Org. Chem. 1978, 43, 3617.

PAGE 152

143 78TL5157 Sundberg, R. and Bloom, J. Tetrahedron Lett. 1978, 5157. 79AG( E ) 239 Seebach, D. Angew. Chem. , Int. Ed. Engl 1979, 18, 239. 79CC822 Magnus, P. and Roy, G. J. Chem. Soc. , Chem. Commun. 1979, 822. 79CJC2342 Ondrus, T. A.; Knaus, E. E. and Giam, C S. Can. J. Chem. 1979, 57, 2342. 79JCS( PI ) 442 Katritzky, A. R.; Lewis, J. and Nie, P.L. J. Chem. Soc., Perkin Trans. 1 1979, 442. 79 JCS ( PI ) 3082 Giam, C.; Goodwin, T.; Rion, K. and Abbot, S. J. Chem. Soc., Perkin Trans. 1979, 3082. 79 JHC409 Ondrus, T. A.; Knaus, E. E. and Giam, C S. J. Heterocycl. Chem. 1979, 16, 409. 79JOC1757 Dam j i , S. and Fyfe, C. A. J. Org. Chem. 1979, 44, 1757. 79MI1 Fleming, I. in "Comprehensive Organic Chemistry", eds. Barton, D. H. R. and Ollis, W. D., Pergamon Press, Oxford, 1979 , vol . 3 , p . 541 . 79TL2485 Wender, P.; Schaus, J. and To.rney, D. Tetrahedron Lett. 1979, 2485. 79TL771 Stork, G.; Jacobson, R. M. and Levitz, R. Tetrahedron Lett. 1979, 771. 80CB1290 Meyer, N. and Seebach, D. Chem. Ber. 1980, 113, 1290. 80CB1304 Meyer, N. and Seebach, D. Chem. Ber. 1980, 113, 1304. 80CJC2447 Knaus, E. E.; Avasthi , K. and Giam, C. S. Can. J. Chem. 1980, 58, 2447.

PAGE 153

144 80JA6157 Wender, P. A.; Schause, J. M. and White, A. W. J. Am. Chem. Soc. 1980, 102, 6157. 80 JOC3382 Sundberg, R. and Bloom, J. J. Org. Chem. 1980, £5, 3382. 80MI1 Marvel, E. N. "Thermal Electrocyclic Reactions", Academic Press, New York, 1980 . 80MI2 Marvel, E. N. "Thermal Electrocyclic Reactions", Academic Press, New York, 1980, p. 266. 80MI3 Marvel, E. N. "Thermal Electrocyclic Reactions", Academic Press, New York, 1980, p. 323. 80MI4 Magnus, P. Aldrichim. Acta 1980, 13, 43. 80S289 Strutz, G.; Yaouanc, J. -J . ; Krausz, F. and Labeeuw, B. Synthesis 1980, 289. 80S589 Becher, J. Synthesis 1980, 589. 80T2531 Krief, A. Tetrahedron 1980, 36, 2531. 80TL2341 Ashcroft, W. and Joule, J. Tetrahedron Lett. 1980, 2341. 8lAG( E ) 127 Hanko, R. and Hoppe, D. Angew. Chem., Int. Ed. Engl. 1981, 20, 127. 81CC150 Le Cocq, C. and Lallemand, J. -Y. J. Chem. Soc., Chem. Commun. 1981, 150. 81JOC2363 Beak, P. and Carter, L. G. J. Org. Chem. 1981, 46, 2363. 81 JOC48 36 Sundberg, R. J. and Bloom, J. D. J. Org. Chem. 1981, 46, 4836. 81T3423 Kost, A. N. ; Gromov, S. P. and Sagitullin, R. S. Tetrahedron 1981, 37, 3423 .

PAGE 154

145 81TL5119 Meyers, A. I. and Hellring, S. Tetrahedron Lett. 1981, 5119. 82CRV223 Stout, D. M. and Meyers, A. I. Chem. 'Rev. 1982, 82, 223. 82JMR535 Patt, S. L. and Shoolery, J. N. J. Mag. Reson. 1982, 46, 535. 82JOC492 Katritzky, A. R. ; Chemprapai, A.; Patel R. C. and Ta r raga-Tomas , A. J. Org. Chem. 1982, 47, 492. 82JOC4315 Comins, D. L. and Abdullah, A. H. J. Org. Chem. 1982, 47, 4315. 82JOC5051 Eisch, J. J.; Galle, J. E.; Piotrowski, A. and Tsai, M. -R. J. Org. Chem. 1982, 47, 5051. 82MI1 Magnus, P. D.; Sarkar, T. and Djuric, S in "Comprehensive Organometallic Chemistry", eds. Wilkinson, G.; Stone, F. G. A. and Abel, E. W. , Pergamon Press, Oxford, 1982, vol . 7, p. 515. 82MI2 Kuthan, J. and Kurfurst, A. Ind. Eng. Chem. Prod. Res. Dev. 1982, 21, 191. 820M553 Magnus, P. and Roy, G. Organometallics 1982, 1, 553. 82T1975 Hickmott, P. W. Tetrahedron 1982, 38, 1975. 82T3363 Hickmott, P. W. Tetrahedron 1982, 38, 3363 . 82TL2527 Krow, G. R.; Carey, J. T. ; Cannon, K. C and Henz, K. J. tetrahedron Lett. 1982, 2527 . 83JOC4017 Katritzky, A. R. and Rubio, 0. J. Org. Chem. 1983, 48, 4017.

PAGE 155

146 83T1963 Seebach, D.; Lohman, J. J.; Syfrig, M. A. and Yoshifuji, M. Tetrahedron 1983, 39, 1963. 83TL1801 Yamaguchi , R.; Nakazono, Y. and Kawanisi, M. Tetrahedron Lett. 1983, 1801 . 83TL2711 Comins, D. L.; Abdullah, A. H. and Smith, R. K. Tetrahedron Lett. 1983, 2711 . 83TL2927 Raucher, S. and Lawrence, R. F. Tetrahedron Lett. 1983, 2927. 84AG( E) 932 Hoppe, D. Angew. Chem. , Int. Ed. Engl. 1984, 23 , 932. 84CL1803 Tsuge, 0.; Kanemasa, S.; Nagahama, H. and Tanaka, J. Chem. Lett. 1984, 1803. 84CRV471 Beak, P.; Zajdel, W. J. and Reitz, D. B Chem. Rev. 1984, 84, 471. 84 JA1010 Beak, P. and Zajdel, W. J. J. Am. Chem. Soc. 1984, 106, 1010. 84JA1130 Cohen, T. and Lin, M. -T. J. Am. Chem. Soc. 1984, 106, 1130. 84JCS( PI ) 1671 Katritzky, A.R.; Bravo-Borja, S.; El-Mowafy, A. M. and Lopez-Rodriguez , M L. J. Chem. Soc., Perkin Trans. 1 1984, 16717“ 84JCS( PI ) 1933 Casey, M. ; Moody, C. J. and Rees, C. W. J. Chem. Soc., Perkin Trans. 1 1984, 1933 . 84MI1 Pereyre, M.; Elissondo, B. and Quintard J. -P. in "Selectivity A Goal for Synthetic Efficiency", eds. Bartmann, W and Trost, B. M., Verlag Chemie, Weinheim, 1984, p. 191. 84S384 Ager, D. J. Synthesis 1984, 384.

PAGE 156

147 84TL691 Moriarty, R. M. and Hou , K. -C. Tetrahedron Lett. 1984, 691. 84TL741 Schlosser, M. and Strunk, S. Tetrahedron Lett. 1984, 741 and references cited therein . 84TL1353 Ahlbrecht, H. and Dollinger, H. Tetrahedron Lett. 1984, 1353. 84TL4867 Comins , D . L . ; Abdullah, A. H. and Mantlo, N. B. Tetrahedron Lett. 1984, 4867 . 85AG( E) 511 Maas, G. and Feith, B. Angew. Chem. , Int. Ed. Enql. 1985, 24, 511. 85CI ( L ) 125 Kant, J. and Popp, F. D. Chem. Ind. (Lond.) 1985, 125. 85JOC1019 Meyers, A. I.; Edwards, P. D.; Bailey, T. R. and Jagdmann, Jr., G. E. J. Orq. Chem. 1985, 50, 1019. 85JOC3236 Raucher, S. and Bray, B. L. J. Org. Chem. 1985, 50, 3236. 85MI1 Meyers, A. I. Aldrichim. Acta 1985, 18, 59 . — 85TL1141 Newmans-Evans, R. H. and Carpenter, B. K. Tetrahedron Lett. 1985, 1141. 85TL2617 Krow, G. R. ; Lee, Y. B.; Szczepanski, S. W. and Raghavachari , R. Tetrahedron Lett. 1985, 2617. 86ACR356 Meyers, A. I. and Beak, P. Acc. Chem. Res. 1986, 19, 356. 86AHC( 39 ) 1 Keay, J. G. Adv . Heterocycl. Chem. 1986, 29, 1 . 86JA2102 Reich, H. J. and Phillips, N. H. J. Am. Chem. Soc. 1986, 108, 2102.

PAGE 157

148 86JOC3076 Gawley, R. E.; Hart, G.; GoicoecheaPappas, M. and Smith, A. L. J. Org. Chem. 1986, 51, 3076. 86 JOC3108 Meyers, A. I.; Sohda, T. and Loewe , M. F. J. Org. Chem. 1986, 51, 3108. 86MI1 Katritzky, A . R . and Fan, W. -Q. 1986, unpublished results. 86T2571 Katritzky, A. R. and Akutagawa, K. Tetrahedron 1986, 42, 2571. 86TL211 Yamaguchi, R.; Moriyasu, M. and Kawanisi, M. Tetrahedron Lett. 1986, 211 . 86TL331 Fraser, R. R. and Mansour, T. A. Tetrahedron Lett. 1986, 331. 86TL2361 Verlhac, J. -B. and Quintard, J. -P. Tetrahedron Lett. 1986, 2361. 87H1333 Katritzky, A. R.; Faid-Allah, H. and Marson, C. M. Heterocycles 1987, 26, 1333 . 87JA1263 Meyers, A. I. and Dickman, D. A. J. Am. Chem. Soc. 1987, 109, 1263. 87JA1265 Gawley, R. E. J. Am. Chem. Soc. 1987, 109, 1265. 87S415 Katritzky, A. R.; Fan, W. -Q . and Akutagawa, K. Synthesis 1987, 415.

PAGE 158

BIOGRAPHICAL SKETCH Saumitra Sengupta was born on the Christmas Eve of 1957 in Calcutta, India. After receiving his primary education in three different cities while travelling with his family, he graduated with B.Sc. (Hons.) and M.Sc. (specialization in organic chemistry) degrees in chemistry from Jadavpur University at Calcutta. After a brief period of research at Jadavpur, he joined the University of Florida as a graduate student in the Fall of 1982. He is a recipient of the National Merit Scholarship from the Government of India and is a member of the American Chemical Society, Indian Chemical Society and the Indian Association for the Cultivation of Science.

PAGE 159

I certify that I have read this study and that in my opinion it confirms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Alan R. Katritzky, Chairmc Kenan Professor of Organic Chemistry I certify that I have read this study and that in my opinion it confirms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. William Prof essor '"of Jones Chemistry I certify that I have read this study and that in my opinion it confirms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. I certify that I have read this study and that in my opinion it confirms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. IU s. Kirk S. Schanze Assistant Professor of Chemistry

PAGE 160

I certify that I have read this study and that in my opinion it confirms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. This dissertation was submitted to the Graduate Faculty of the Department of Chemistry in the College of Liberal Arts and Sciences and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. December, 1987 Nicholas S. Bodor Graduate Research Professor of Medicinal Chemistry Dean, Graduate School