Organic Synthesis Strategy and Control
Organic Synthesis
Organic Synthesis:
Strategy and Control
Paul Wyatt
Senior Lecturer and Director of Undergraduate Studies, School of Chemistry,
University of Bristol, UK
and
Stuart Warren
Reader in Organic Chemistry, Department of Chemistry,
University of Cambridge, UK
Copyright © 2007
John Wiley & Sons Ltd,
The Atrium, Southern Gate, Chichester,
West Sussex PO19 8SQ, England
Telephone (+44) 1243 779777
Email (for orders and customer service enquiries): [email protected]
Visit our Home Page on www.wiley.com
All Rights Reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any
means, electronic, mechanical, photocopying, recording, scanning or otherwise, except under the terms of the Copyright, Designs
and Patents Act 1988 or under the terms of a licence issued by the Copyright Licensing Agency Ltd, 90 Tottenham Court Road,
London WIT 4LP, UK, without the permission in writing of the Publisher. Requests to the Publisher should be addressed to the
Permissions Department, John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex PO19 8SQ, England, or
emailed to [email protected], or faxed to (+44) 1243 770620.
Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product names
used in this book are trade names, service marks, trademarks or registered trademarks of their respective owners. The Publisher is
not associated with any product or vendor mentioned in this book.
This publication is designed to provide accurate and authoritative information in regard to the subject matter covered. It is sold on the
understanding that the Publisher is not engaged in rendering professional services. If professional advice or other expert assistance is
required, the services of a competent professional should be sought.
Other Wiley Editorial Offices
John Wiley & Sons Inc., 111 River Street, Hoboken, NJ 07030, USA
Jossey-Bass, 989 Market Street, San Francisco, CA 94103-1741, USA
Wiley-VCH Verlag GmbH, Boschstr. 12, D-69469 Weinheim, Germany
John Wiley & Sons Australia Ltd, 42 McDougall Street, Milton, Queensland 4064, Australia
John Wiley & Sons (Asia) Pte Ltd, 2 Clementi Loop #02-01, Jin Xing Distripark, Singapore 129809
John Wiley & Sons Canada Ltd, 6045 Freemont Blvd, Mississauga, ONT, L5R 4J3, Canada
Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in
electronic books.
Anniversary Logo Design: Richard J. Pacifico
Library of Congress Cataloging-in-Publication Data
Wyatt, Paul.
Organic synthesis: strategy and control / Paul Wyatt and Stuart Warren.
p. cm.
Includes bibliographical references.
ISBN: 978-0-470-48940-5
ISBN: 978-0-471-92963-5
1. Organic compounds – Synthesis. 2. Stereochemistry. I. Warren, Stuart
G. II. Title.
QD262.W89 2007
547´.2–dc22
2006034932
British Library Cataloguing in Publication Data
A catalogue record for this book is available from the British Library
ISBN: 978-0-471-48940-5 (HB)
ISBN: 978-0-471-92963-5 (PB)
Typeset in 10/12pt Times by Thomson Digital
Printed and bound in Great Britain by Antony Rowe Ltd, Chippenham, Wiltshire
This book is printed on acid-free paper responsibly manufactured from sustainable forestry
in which at least two trees are planted for each one used for paper production.
Contents
Preface
vii
A: Introduction: Selectivity
1. Planning Organic Syntheses: Tactics, Strategy and Control
2. Chemoselectivity
3. Regioselectivity: Controlled Aldol Reactions
4. Stereoselectivity: Stereoselective Aldol Reactions
5. Alternative Strategies for Enone Synthesis
6. Choosing a Strategy: The Synthesis of Cyclopentenones
3
9
27
43
55
71
B: Making Carbon–Carbon Bonds
7. The Ortho Strategy for Aromatic Compounds
8. σ-Complexes of Metals
9. Controlling the Michael Reaction
10. Specific Enol Equivalents
11. Extended Enolates
12. Allyl Anions
13. Homoenolates
14. Acyl Anion Equivalents
91
113
127
139
155
173
189
203
C: Carbon–Carbon Double Bonds
15. Synthesis of Double Bonds of Defined Stereochemistry
16. Stereo-Controlled Vinyl Anion Equivalents
17. Electrophilic Attack on Alkenes
18. Vinyl Cations: Palladium-Catalysed C–C Coupling
19. Allyl Alcohols: Allyl Cation Equivalents (and More)
223
255
277
307
339
D: Stereochemistry
20. Control of Stereochemistry – Introduction
21. Controlling Relative Stereochemistry
22. Resolution
23. The Chiral Pool — Asymmetric Synthesis with Natural Products as
Starting Materials —
24. Asymmetric Induction I Reagent-Based Strategy
25. Asymmetric Induction II Asymmetric Catalysis: Formation of C–O
and C–N Bonds
26. Asymmetric Induction III Asymmetric Catalysis: Formation of C–H
and C–C Bonds
27. Asymmetric Induction IV Substrate-Based Strategy
371
399
435
465
505
527
567
599
vi
Contents
28. Kinetic Resolution
29. Enzymes: Biological Methods in Asymmetric Synthesis
30. New Chiral Centres from Old — Enantiomerically Pure
Compounds & Sophisticated Syntheses —
31. Strategy of Asymmetric Synthesis
E: Functional Group Strategy
32. Functionalisation of Pyridine
33. Oxidation of Aromatic Compounds, Enols and Enolates
34. Functionality and Pericyclic Reactions: Nitrogen Heterocycles by
Cycloadditions and Sigmatropic Rearrangements
35. Synthesis and Chemistry of Azoles and other Heterocycles with Two
or more Heteroatoms
36. Tandem Organic Reactions
627
651
681
717
749
777
809
835
863
General References
893
Index
895
Preface
We would like to thank those who have had the greatest influence on this book, namely the undergraduates at the Universities of Bristol and Cambridge. But, particularly we would like to thank
the organic chemists at Organon (Oss), AstraZeneca (Alderley Park, Avlon Works, Mölndal and
Macclesfield), Lilly (Windlesham), Solvay (Weesp) and Novartis (Basel) who contributed to the
way the book was written more than they might realise. These chemists will recognise material
from our courses on The Disconnection Approach, Advanced Heterocyclic Chemistry, New Synthetic Methods and Asymmetric Synthesis. Additionally we would like to thank the participants
at the SCI courses organised by the Young Chemists Panel. All these industrial chemists participated in our courses and allowed us to find the best way to explain concepts that are difficult to
grasp. This book has changed greatly over the ten years it was being written as we became more
informed over what was really needed. The book is intended for that very audience – final year
undergraduates, graduate students and professional chemists in industry.
PJW
SGW
July 2006
Section A:
Introduction: Selectivity
1. Planning Organic Syntheses: Tactics, Strategy and Control ............................................. 3
2. Chemoselectivity ..................................................................................................................... 9
3. Regioselectivity: Controlled Aldol Reactions ..................................................................... 27
4. Stereoselectivity: Stereoselective Aldol Reactions ............................................................. 43
5. Alternative Strategies for Enone Synthesis ........................................................................ 55
6. Choosing a Strategy: The Synthesis of Cyclopentenones .................................................. 71
1
Planning Organic Syntheses:
Tactics, Strategy and Control
The roll of honour inscribed with successful modern organic syntheses is remarkable for the
number, size, and complexity of the molecules made in the last few decades. Woodward and
Eschenmoser’s vitamin B12 synthesis,1 completed in the 1970s, is rightly regarded as a pinnacle of
achievement, but since then Kishi2 has completed the even more complex palytoxin. The smaller
erythromycin and its precursors the erythronolides3 1, and the remarkably economical syntheses of
the possible stereoisomers of the cockroach pheromones 2 by Still4 deal with a greater concentration
of problems.
O
O
O
HO
1
erythronolide A
OH
OH
2
periplanone-B
O
O
OH
O
OH
Less applauded, but equally significant, is the general advance in synthetic methods and their
industrial applications. AstraZeneca confess that it took them nearly a century to bring Victor
Grignard’s methods into use, but are proud that Corey’s sulfur ylid chemistry made it in a decade.
Both are used in the manufacture of the fungicide flutriafol5 3.
Me
F
S
CH2
O
F
O
F
F
Me
F
N
RCO3H
F
N
N N
N
N
base
F
F
F
Br2
F
X
O
MgI
OH
F
F
AlCl3
+
H2O
F
3
Flutriazole
fungicide
OH
Cl
COCl
Cl
Optically active and biodegradable deltamethrin6 4 has taken a large share of the insecticide
market, and asymmetric hydrogenation is used in the commercial synthesis of DOPA 5 used to
treat Parkinson’s disease.7 These achievements depend both on the development of new methods
and on strategic planning:8 the twin themes of this book.
Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren
Copyright © 2007 John Wiley & Sons, Ltd
4
1 Planning Organic Syntheses: Tactics, Strategy and Control
Br
HO
R R
O S
Br
O
NH2
O
H
CO2H
HO
CN
4
deltamethrin
5
(S)-L-(–)-DOPA
To make any progress in this advanced area, we have to assume that you have mastered the
basics of planning organic synthesis by the disconnection approach, roughly the material covered
in our previous books.9 There, inspecting the target molecule, identifying the functional groups,
and counting up the relationships between them usually gave reliable guidelines for a logical
synthesis. All enones were tackled by some version of the aldol reaction; thus 6 would require the
attack of enolate 7 on acetone. We hope you already have the critical judgement to recognise that
this would need chemoselectivity in enolising 7 rather than acetone or 6, and regioselectivity in
enolising 7 on the correct side.
O
O
aldol
O
+
O
6
7
In this book we shall explore two new approaches to such a problem. We shall see how to make
specific enol equivalents for just about any enolate you might need, and we shall see that alternative disconnections such as 6a, the acylation of a vinyl anion 8, can be put into practice. Another
way to express the twin themes of this book is strategy and control: we solve problems either
by finding an alternative strategy or by controlling any given strategy to make it work. This will
require the introduction of many new methods - a whole chapter will be devoted to reagents for
vinyl anions such as 8, and this will mean exploring modern organometallic chemistry.
O
O
+
6a
X
8
9
We shall also extend the scope of established reactions. We hope you would recognise the aldol
disconnection in TM 10, but the necessary stereochemical control might defeat you. An early
section of this book describes how to control every aspect of the aldol reaction: how to select
which partner, i.e. 11 or 12, becomes an enolate (chemoselectivity), how to control which enolate
of the ketone 12 is formed (regioselectivity), and how to control the stereochemistry of the product
10 (stereoselectivity). As we develop strategy, we shall repeatedly examine these three aspects of
control.
OH
O
O
R2
R1
?
10
O
aldol
R1
H
11
+
O
R2
R2
?
12
The target molecules we shall tackle in this book are undoubtedly more difficult in several ways
than this simple example 10. They are more complex quantitatively in that they combine functional
1 Planning Organic Syntheses: Tactics, Strategy and Control
5
groups, rings, double bonds, and chiral centres in the same target, and qualitatively in that they
may have features like large rings, double bonds of fixed configuration, or relationships between
functional groups or chiral centres which no standard chemistry seems to produce. Molecules 1 to
5 are examples: a quite different one is flexibilene 13, a compound from Indonesian soft coral. It
has a fifteen-membered ring, one di- and three tri-substituted double bonds, all E but none
conjugated, and a quaternary centre. Mercifully there are no functional groups or chiral centres.
How on earth would you tackle its synthesis? One published synthesis is by McMurry.10
OMe
1. 2 x BuLi
H
2.
Cp2ZrHCl
1. PCC
HO
O
14
15; 78% yield
2. H ,
CH(OMe)3
17
MeO
16; 83% yield
OMe
MeO
17
O
ZrCp2Cl
18
Pd(OCOCF3)2
TiCl3
OHC
O
Zn/Cu
13; flexibilene, 52% yield
19; 76% yield from 16 + 18
This short synthesis uses seven metals (Li, Cr, Zr, Pd, Ti, Zn, and Cu), only one protecting
group, achieves total control over double bond geometry, remarkable regioselectivity in the Zr-Pd
coupling reaction, and a very satisfactory large ring synthesis. The yield in the final step (52%)
may not look very good, but this is a price worth paying for such a short synthesis. Only the first
two steps use chemistry from the previous books: all the other methods were unknown only ten
years before this synthesis was carried out but we shall meet them all in this book.
An important reason for studying alternative strategies (other than just making the compound!)
is the need to find short cheap large scale routes in the development of research lab methods into
production. All possible routes must be explored, at least on paper, to find the best production
method and for patent coverage. Many molecules suffer this exhaustive process each year, and
some sophisticated molecules, such as Merck’s HIV protease inhibitor 20, a vital drug in the fight
against AIDS, are in current production on a large scale because a good synthesis was found by
this process.11
N
OH
Ph
H
N
N
N
t-BuNH
O
OH
O
20; Crixivan
You might think that, say organometallic chemistry using Zr or Pd would never be used in
manufacture. This is far from true as many of these methods are catalytic and the development
of polymer-supported reagents for flow systems means that organo-metallic reagents or enzymes
may be better than conventional organic reagents in solution with all the problems of by-product
disposal and solvent recovery. We shall explore the chemistry of B, Si, P, S, and Se, and of metals
6
1 Planning Organic Syntheses: Tactics, Strategy and Control
such as Fe, Co, Ni, Pd, Cu, Ti, Sn, Ru and Zr because of the unique contribution each makes to
synthetic methods.
In the twenty years since McMurry’s flexibilene synthesis major developments have changed
the face of organic synthesis. Chiral drugs must now be used as optically pure compounds and
catalytic asymmetric reactions (chapters 25 and 26) have come to dominate this area, an achievement crowned by the award of the 2001 Nobel prize for Chemistry to Sharpless, Noyori and
Knowles. Olefin metathesis (chapter 15) is superseding the Wittig reaction. Palladium-catalysed
coupling of aromatic rings to other aromatic rings, to alkenes, and to heteroatoms (chapter 18)
makes previously impossible disconnections highly favourable. These and many more important
new methods make a profound impact on the strategic planning of a modern synthesis and find
their place in this book.
A Modern Synthesis: Fostriecin (CI-920)
The anti-cancer compound Fostriecin 21 was discovered in 1983 and its stereochemistry elucidated in 1997. Not until 2001 was it synthesised and then by two separate groups.12 Fostriecin is
very different from flexibilene. It still has alkene geometry but it has the more challenging threedimensional chirality as well. It has plenty of functionality including a delicate monophosphate
salt. A successful synthesis must get the structure right, the geometry of the alkenes right, the
relative stereochemistry right, and it must be made as a single enantiomer.
O
HO
O
O
P
Na
OH
OH
O
O
OH
21; Fostriecin (CI-920)
The brief report of Jacobsen’s total synthesis starts with a detailed retrosynthetic analysis.
The compound was broken into four pieces 21a after removal of the phosphate. The unsaturated
lactone 24 (M is a metal) could be made by an asymmetric oxo-Diels-Alder reaction from diene 22
and ynal 23. The epoxide 25 provides a second source of asymmetry. One cis alkene comes from
an alkyne 26 and the rest from a dienyl tin derivative 27.
O O
P
HO
O
21a
Disconnection of
Fostriecin (CI-920)
O
Na
OH
OH
O
OH
OR
DielsAlder
RO
O
O
22
O
23
SiMe3
24
S
SnBu3
H
M
O
S
O
25
26
SiMe3
27
The synthesis is a catalogue of modern asymmetric catalytic methods. The epoxide 25 was
resolved by a hydrolytic kinetic resolution (chapter 28) using a synthetic asymmetric cobalt complex. The asymmetric Diels-Alder reaction (chapter 26) was catalysed by a synthetic chromium
1 References
7
complex. The vinyl metal derivative 24 was made by hydrozirconation of an alkyne (this at least
is similar to the flexibilene synthesis) and the secondary alcohol chiral centre was derived from
the dithian 26 by hydrolysis to a ketone and asymmetric reduction with a synthetic ruthenium
complex (chapter 24). The dienyl tin unit 27 was coupled to the rest of the molecule using catalytic
palladium chemistry (chapter 18). Almost none of these catalytic methods was available in 1983
when flexibilene was made and such methods are a prominent feature of this book. Organic synthesis nowadays can tackle almost any problem.13
Please do not imagine that we are abandoning the systematic approach or the simpler reagents
of the previous books. They are more essential than ever as new strategy can be seen for what
it is only in the context of what it replaces. Anyway, no-one in his or her right mind would use
an expensive, toxic, or unstable reagent unless a friendlier one fails. Who would use pyrophoric
tertiary butyl-lithium in strictly dry conditions when aqueous sodium hydroxide works just as
well? In most cases we shall consider the simple strategy first to see how it must be modified. The
McMurry flexibilene synthesis is unusual in deploying exotic reagents in almost every step. A
more common situation is a synthesis with one exotic reagent and six familiar ones. The logic of
the previous books is always our point of departure.
The organisation of the book
The book has five sections:
A:
B:
C:
D:
E:
Introduction, selectivity, and strategy
Making Carbon-Carbon bonds
Carbon-Carbon double bonds
Stereochemistry
Functional Group Strategy
The introductory section uses aldol chemistry to present the main themes in more detail and
gives an account of the three types of selectivity: chemo-, regio-, and stereo-selectivity. We shall
explore alternative strategies using enones as our targets, and discuss how to choose a good route
using cyclopentenones as a special case among enones. Each chapter develops strategy, new
reagents, and control side-by-side. To keep the book as short as possible (like a good synthesis),
each chapter in the book has a corresponding chapter in the workbook with further examples,
problems, and answers. You may find that you learn more efficiently if you solve some problems
as you go along.
References
General references are given on page 893
1. R. B. Woodward, Pure Appl. Chem., 1973, 33, 145; A. Eschenmoser and C. E. Wintner, Science, 1977,
196, 1410; A. Eschenmoser, Angew. Chem., Int. Ed. Engl., 1988, 27, 5.
2. Y. Kishi, Tetrahedron, 2002, 58, 6239.
3. E. J. Corey, K. C. Nicolaou, and L. S. Melvin, J. Am. Chem. Soc., 1975, 97, 654; G. Stork and S. D.
Rychnovsky, J. Am. Chem. Soc., 1987, 109, 1565; Pure Appl. Chem., 1987, 59, 345; A. F. Sviridov, M. S.
Ermolenko, D. V. Yashunsky, V. S. Borodkin and N. K. Kochetkov, Tetrahedron Lett., 1987, 28, 3835,
and references therein.
4. W. C. Still, J. Am. Chem. Soc., 1979, 101, 2493. See also S. L. Schreiber and C. Santini, J. Am. Chem.
Soc., 1984, 106, 4038; T. Takahashi, Y. Kanda, H. Nemoto, K. Kitamura, J. Tsuji and Y. Fukazawa,
J. Org. Chem., 1984, 51, 3393; H. Hauptmann, G. Mühlbauer and N. P. C. Walker, Tetrahedron Lett.,
1986, 27, 1315; T. Kitahara, M. Mori and K. Mori, Tetrahedron, 1987, 43, 2689.
8
1 Planning Organic Syntheses: Tactics, Strategy and Control
5. P. A. Worthington, ACS Symposium 355, Synthesis and Chemistry of Agrochemicals, eds D. R. Baker,
J. G. Fenyes, W. K. Moberg, and B. Cross, ACS, Washington, 1987, p 302.
6. M. Elliott, A. W. Farnham, N. F. Janes, P. H. Needham, and D. A. Pullman, Nature, 1974, 248, 710;
M. Elliott, Pestic. Sci., 1980, 11, 119.
7. J. Halpern, H. B. Kagan, and K. E. Koenig, Morrison, vol 5, pp 1–101.
8. Corey, Logic; Nicolaou and Sorensen.
9. Designing Syntheses, Disconnection Textbook, and Disconnection Workbook.
10. J. McMurry, Acc. Chem. Res., 1983, 16, 405.
11. D. Askin, K. K. Eng, K. Rossen, R. M. Purick, K. M. Wells, R. P. Volante and P. J. Reider, Tetrahedron
Lett., 1994, 35, 673; B. D. Dorsey, R. B. Levin, S. L. McDaniel, J. P. Vacca, J. P. Guare, P. L. Darke,
J. A. Zugay, E. A. Emini, W. A. Schleif, J. C. Quintero, J. H. Lin, I.-W. Chen, M. K. Holloway, P. M. D.
Fitzgerald, M. G. Axel, D. Ostovic, P. S. Anderson and J. R. Huff, J. Med. Chem., 1994, 37, 3443.
12. D. L. Boger, S. Ichikawa and W. Zhong, J. Am. Chem. Soc., 2001, 123, 4161; D. E. Chavez and E. N.
Jacobsen, Angew. Chem., Int. Ed., 2001, 40, 3667.
13. D. Seebach, Angew. Chem. Int. Ed., 1990, 29, 1320; K. C. Nicolaou, E. J. Sorensen and N. Winssinger,
J. Chem. Ed., 1998, 75, 1225.
2 Chemoselectivity
Definitions
Introduction: three types of control
Chemoselectivity: simple examples and rules
Chemoselectivity by Reactivity and Protection: An anti-Malaria Drug
Protection to allow a less reactive group to react
When Protection is not Needed
Dianions: wasting reagent to achieve selectivity
Chemoselectivity by Reagent: The Pinacol Rearrangement
Selectivity between secondary and tertiary alcohols by reagent
Corey’s longifolene synthesis
Chemoselectivity in Enol and Enolate Formation
General discussion of enols and enolates
Formation of specific enol equivalents
Lithium enolates, enamines and silyl enol ethers
Enamines
Silyl enol ethers
Synthesis of the ant alarm pheromone mannicone
Examples of Chemoselectivity in Synthesis
Synthesis of lipstatin, rubrynolide and hirsutene
Definitions
Introduction: three types of control
Behind all grand strategic designs in organic synthesis must lie the confidence that molecules can
be compelled to combine in the ways that we require. We shall call this control and divide it into three
sections by mechanistic arguments. These sections are so important that we shall devote the next three
chapters to the more detailed explanation of just what the divisions mean. If you can recognise what
might go wrong you are in a better position to anticipate the problem and perhaps avoid it altogether.
Our three types of control are over chemoselectivity (selectivity between different functional groups),
regioselectivity (control between different aspects of the same functional group), and stereoselectivity
(control over stereochemistry). Examples of selectivity of all three kinds are given in The Disconnection
Approach: Chemoselectivity in chapter 5, Regioselectivity in chapter 14, and Stereoselectivity in
chapters 12 and 38. These aspects will not be addressed again in the present book.
Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren
Copyright © 2007 John Wiley & Sons, Ltd
10
2 Chemoselectivity
Chemoselectivity: simple examples and rules
Chemoselectivity is the most straightforward of the three types and might seem too elementary
to appear in an advanced textbook. Counting the number of protecting groups in the average
synthesis reveals this as a naive view. Selectivity between functional groups might involve:
(a) Selective reaction of one among several functional groups of different reactivity, as in the
reduction of the keto-acid 2 to give either product 1 or 3 at will.
O
O
OH
?
OH
?
CO2H
1
CO2H
2
3
(b) Selective reaction of one of several identical functional groups, as in the conversion of the
symmetrical diacid 5 to the half ester, half acid chloride 4, or the lactone 6 in which one of the two
acids has been reduced. There is a more subtle example of this at the end of the chapter.
?
COCl
MeO2C
CO2H
HO2C
substitution
4
?
reduction
O
6
5
O
(c) Selective reaction of a functional group to give a product which could itself react with the
same reagent, as in the classical problem of making a ketone 8 from an acid derivative 7 without
getting the alcohol 9 instead.
O
R1
OEt
OH
O
R2–Metal
and not
R2
R1
?
8
7
R1
R2
R2
9
Organic chemists are developing ever more specific reagents to do these jobs. These reagents
must carry out the reaction they are designed for and must not:
(i) react with themselves.
(ii) react with functional groups other than the one they are aimed at.
(iii) react with the product.
Proviso (ii) is obvious, but (i) and (iii) perhaps need some explanation. It seems hardly worth
stating that a reagent should not react with itself, but it is only too easy to suggest using a reagent
such as 11 without realising that the organo-metallic reagent will act as a base for its own hydroxyl
group 12 and destroy itself. The traditional solution to this problem is protection of the OH group
in 10 but ideally we should like to avoid protection altogether though this is not yet possible.
Br
HO
Mg
MgBr
HO
10
HO
11
n-butanol
H , H 2O
O
H
12
MgBr
O
CH3
13
work-up
OH
CH3
n-butanol
2 Definitions
11
Proviso (iii) is more obvious and yet perhaps more often catches people out. It is not always clear
in exactly what form the product is produced in the reaction mixture, though a good mechanistic
understanding and careful thought should reveal this. The reaction between the simple aldehyde
14 and chloral (Cl3C.CHO) looks like a straightforward route to the aldol 17, and might reasonably
be carried out via the enamine 16.
CHO
+
HN
O
14
O
15
OH
Cl3C.CHO
N
CHO
Cl3C
?
17
16
However, mixtures of 16 and chloral, in any proportion, give only the 2:1 adduct 20 which can
be isolated in 83% yield.1 Obviously the immediate product 19 reacts with chloral at least as fast
as does 16. Fortunately the synthesis can be rescued by acid-catalysed cleavage of 20 with HCl
which gives a good yield of the target 17.
CCl3
CCl3
Cl3C
O
O
O
Cl3C.CHO
N
Cl3C
N
O
O
N
O
O
18
19
20
Enamines are excellent at Michael additions and another plausible synthesis which “goes
wrong” is the addition of acrolein to cyclohexanone mediated by the enamine 21 formed this time
with pyrollidine.
O
O
CHO
N
+
HN
?
CHO
21
22
If the product is isolated by distillation, a good yield (75%) of the bicyclic ketone 23 is obtained.2
A more detailed investigation disclosed that 24 is the immediate product, that 23 is formed from
it on distillation, and that the expected Michael adduct 22 can be isolated in good yield simply
by the hydrolysis of 24. In other words, don’t distil! If things “go wrong” in a synthesis, this may
be a blessing, as here. There are lots of ways to control Michael additions, but few ways to make
bicyclic ketones like 23, and this is now a standard method.3 The moral is to make sure you know
what is happening, and to be prepared to welcome the useful and unexpected result.
O
N
23
N
O
24
N
12
2 Chemoselectivity
Chemoselectivity by Reactivity and Protection: An anti-Malaria Drug
We need to see some of these principles in action and a proper synthesis is overdue. The anti-malarial
drug amopyroquine 25 might have been derived from quinine as it has a quinoline nucleus. It also has
five functional groups – three amines (all different - one aromatic, one tertiary, and one secondary), a
phenol and an aryl chloride. There are four rings, three aromatic and one saturated heterocyclic.
OH
OH
a
N
HN
b
N
Cl
N
Cl
25
Amopyroquine
N
HN
25a, b
Amopyroquine disconnections
There are many possible disconnections, but we should prefer to start in the middle of the molecule
to achieve the greatest simplification. Disconnection 25a would require a nucleophilic displacement
(X ⫽ a leaving group) on an unactivated benzene ring 27 and looks unpromising. Disconnection 25b
requires nucleophilic displacement at position 4 in a pyridine ring, an acceptable reaction because of
the electron-withdrawing effect of the nitrogen atom in the ring, so this is the better route, though we
may be apprehensive about controlling the chemoselectivity as there are three potential nucleophiles
in 26 and two potential electrophiles in 28.
X
+
+
25
N
Cl
28
OH
a
b
N
H2N
N
Cl
NH2
OH
N
X
29
26
27
Further disconnections of 26 by the Mannich reaction4 and of 28 by standard heterocyclic
methods give simple starting materials.5
OH
OH
Mannich
+ CH2O +
N
H2N
H2N
HN
30
26a
X
O
O
quinoline
synthesis
FGI
+
N
Cl
28
N
H
Cl
31
HO
NH2
Cl
32
Protection to allow a less reactive group to react
Now the fun begins! Attempted Mannich reaction on the aminophenol 30 would be dominated by the
more nucleophilic NH2 group and is no good. Acylation moderates the NH2 group by delocalisation
2 When Protection is not Needed
13
and 33 is a good choice for starting material as it is paracetamol, the common analgesic. Mannich
reaction now chemoselectively gives 34 and alkaline hydrolysis of the amide gives 26.
OH
CH2O
··
N
H
AcOH
H2N
OH
OH
O
Ac2O
pyrrolidine
30
N
AcHN
33
34
Michael addition of acrylic acid to the chloroamine 32 is straightforward and Friedel-Crafts
cyclisation of 35 gives only 31, presumably because the position next to the chlorine atom is slightly
disfavoured both sterically and electronically. Chlorination and oxidation are conveniently carried
out in the same step and the two halves (26 and 28) of this convergent synthesis are combined to
give amopyroquine 25.
Cl
CO2H
CO2H
PPA
32
POCl3
26
I2
heat in
aqueous ethanol
31
N
H
Cl
Cl
25
N
28; X = Cl
35
In the last step we return to the original question of chemoselectivity: Only the primary amine
in 26 reacts because it is more nucleophilic than OH and because the more nucleophilic tertiary
amine adds reversibly – it cannot lose a hydrogen atom as it does not have one. Only the 4-chlorine
atom in the pyridine 28 reacts, presumably because addition to the other position would require
the disruption of both aromatic rings. Though this compound has been succeeded by better antimalarials, its synthesis illustrates the all-important principle that predictions of chemoselectivity
must be based on sound mechanistic understanding. If doubt remains it is worth trying a model
reaction on simpler compounds or, of course, an alternative strategy.
When Protection is not Needed
Dianions: wasting reagent to achieve selectivity
In that synthesis we moderated an over-reactive amino group by protection. Sometimes, protection
is not necessary if we are prepared to squander some of our reagents. A trivial example is the
addition of methyl Grignard to the ketoacid 36. We have already seen how acidic protons destroy
Grignard reagents, but if we are prepared to waste one molecule of the Grignard, we get automatic
protection of the carboxylic acid by deprotonation. Nucleophilic MeMgI will not add to the anion
of a carboxylic acid but adds cleanly to the ketone to give, after workup, the alcohol 37.
O
O
MeMgI
CO2H
R
MeMgI
CO2 Mg
R
2
36; Mg salt
36
OH
O
R
CO2 Mg
Me
37; Mg salt
2
H
H2O
R
CO2H
Me
37
14
2 Chemoselectivity
At first sight, the synthesis of Z-38 by the Wittig reaction seems too risky. The phosphonium
salt 39 has a more acidic proton (CO2H) than the one we want to remove to make the ylid, and the
aldehyde 40 not only also has an acidic proton (OH), but it prefers to remain as the cyclic hemiacetal 41 so that there is no carbonyl group at all.
Wittig
Z-38 HO
Wittig
CO2H reagent
39
Ph3P
CO2H
+
hemi-acetal
41
HO
OH
O
CHO
hydroxy-aldehyde
40
However, simply using a large excess of base makes the reaction work without any protection.
The phosphonium salt 39 does indeed lose its first proton from the CO2H group 42, but the second
molecule of base forms the ylid 43 as the two anions are far enough apart not to influence each
other.6 Base also catalyses the equilibrium between the anions 44 and 45 so that 43 and 45 can
react to give the target molecule. The transition state for this reaction has three partial negative
charges, but they are well apart from each other and there is obviously not too much electrostatic
repulsion as the reaction goes well. This case is opposite to the previous ones: careful mechanistic
analysis shows that expected chemoselectivity problems do not materialise.
base
base
Ph3P
39
Ph3P
CO2
42
41
CO2
43
base
O
44
O
O
Z-38
CHO
45
Chemoselectivity by Reagent: The Pinacol Rearrangement
So far we have discussed chemoselectivity between different functional groups. The situation gets
more complicated if the functional groups are similar, or even the same. The pinacol rearrangement
is a useful route to carbonyl compounds from diols, the classical example being the rearrangement
of 46 in acid solution to give the t-alkyl ketone 48. There are no chemoselectivity problems here:
the two hydroxyl groups in 46 are the same so it does not matter which gets protonated and, in the
rearrangement step 47, all four potential migrating groups are methyl.
H
HO
OH
46
'pinacol'
Me
migration
Me
HO
OH2
HO
47
O
48
'pinacolone'
Selectivity between secondary and tertiary alcohols by reagent
Unsymmetrical diols provide a serious problem of chemoselectivity with an ingenious solution.7
Treatment of the diol 49 with acid leads to loss of OH from what would be the more stable
t-alkyl cation and hence, by hydrogen shift, to the ketone 51.
2 Chemoselectivity by Reagent: The Pinacol Rearrangement
H
H
migration
H
H
R
R
15
H
R
H
R
HO
OH
HO
49
HO
OH2
O
51
50
The alternative, more interesting rearrangement to give 53 can be initiated by tosylation of
the diol 49 in weak base. It is impossible to tosylate tertiary alcohols under these conditions, as
both the t-alcohol and TsCl are large, so only the secondary alcohol becomes sulfonylated and so
leaves, and the rearranged ketone 53 is the only product.
H
H
TsCl
R
HO
OH
Me
migration
Me
pyridine
TsO
49
Me
H
R
H
R
OH
OH
R
52
O
53
Corey’s longifolene synthesis
The question of which group migrates in a pinacol rearrangement is also a question of chemoselectivity, and usually groups that can participate because they have lone pair or π-electrons migrate
best. In Corey’s longifolene synthesis,8 the 6/7 fused enone 54 was an important intermediate. Synthesis from the readily available Robinson annelation product 57 is very attractive, but this demands
a ring expansion step such as the pinacol rearrangement of 55 of unknown selectivity. 1,2-Diols such
as 55 normally come from the hydroxylation of an alkene, in this case the diene 56 which might be
made by a Wittig reaction on the dione 57. Every step in this sequence raises a question of chemoselectivity. Which of the two ketones in 57 is more reactive? Which of the two double bonds in 56 is
more easily hydroxylated? Which side of the ring migrates in the pinacol rearrangement on 55?
pinacol
O
O
O
O
?
HO
O
FGI
Wittig
?
?
HO
O
54
55
56
57
One of the ketones in 57 is conjugated, and one is not. The unconjugated one is less stable and
we can therefore use thermodynamic control if we protect as an acetal, a reversible process. The
unconjugated ketone would also be more kinetically reactive towards the Wittig reagent. Of the
two double bonds in 59, the one outside the ring is more reactive towards electrophilic reagents,
again for both kinetic and thermodynamic reasons. The tosylation route ensures that the right
OH group leaves in the pinacol rearrangement and because the remaining π-bond migrates better
than the simple alkyl group when 60 rearranges with a weak Lewis acid, all is well. The synthesis
therefore follows the route below, with all questions of chemoselectivity neatly solved. The acetal
protecting group was also useful later in the synthesis.
16
2 Chemoselectivity
O
O
O
O
Ph3P
HO
57
OsO4
OH
TsOH
pyridine
O
E and Z-59
58
O
O
O
O
O
HO
HO
O
LiClO4
TsCl
pyridine
TsO
HO
O
CaCO3
THF
60
61
Chemoselectivity in Enol and Enolate Formation
General discussion of enols and enolates
We have concentrated so far on two functional groups within the same molecule. The chemoselectivity problem is just as important when we want two molecules to react together in a certain
way, but, because both molecules have similar functional groups, the reaction can occur the other
way round, or one of the molecules may react with itself and ignore the other. This problem is
particularly acute in reactions involving enolisation. The alkylation or acylation of enols or enolates and the reaction of one carbonyl compound with another, the aldol reaction, are classical
and important examples summarised in the general scheme below. We shall concentrate in this
chapter on the chemoselectivity of these processes, that is we shall look at the enolisation of esters,
aldehydes, and the like.
OH
O
R1
O
R1
carbonyl compound
enol
O
R1
enolate
O
aldol
O
O
OH
R2
R1
R1
acylation O
aldol
H
R2
R1
RBr
O
aldol
R2
H
O
alkylation
R2
H
OH
O
R2
R2
X
R2
OH
O
R2
R1
O
R2
R1
Reaction of an ester 62 with its own alkoxide ion produces a small amount of enolate 63 that
reacts with unenolised ester to give the ketoester 64. This reaction, though useful in its own right,
precludes the direct alkylation of esters under these conditions.
H
H
H
EtO
R
OEt
R
OEt
R
O
O
62; ester
63; ester enolate
R
CO2Et
R
CO2Et
O
64; β-keto-ester
2 Chemoselectivity in Enol and Enolate Formation
17
Formation of specific enol equivalents
What is needed for the alkylation is rapid conversion of the ester into a reasonably stable enolate, so rapid in fact that there is no unenolised ester left. In other words the rate of proton
removal must be faster than the rate of combination of enolate and ester. These conditions are
met when lithium enolates are made from esters with lithium amide bases at low temperature,
often ⫺78 ⬚C. Hindered bases must be used as otherwise nucleophilic displacement will occur
at the ester carbonyl group to give an amide. Popular bases are LDA (Lithium Di-isopropyl
Amide, 66), lithium hexamethyldisilazide 67, and lithium tetramethylpiperidide 68, the most
hindered of all. These bases are conveniently prepared from the amine, e.g. 65 for LDA, and
BuLi in dry THF solution.
Me3Si
BuLi
N
H
N
THF
65
Di-isopropylamine
Li
SiMe3
N
N
Li
Li
66; LDA
67; LHMDS
Li Di-isopropyl Amide Li HexaMethylDiSilyazide
68; LTMP
Li TetraMethylPiperidide
Treatment of a simple ester 62 with one of these bases at ⫺78 ⬚C leads to a stable lithium
enolate 70 by initial coordination of lithium to the carbonyl group 69 and proton removal via a
six-membered cyclic transition state 69a.
H
H
O
R
H
LDA
–78 °C
THF
OEt
R2N
Li
H
H
O
R
R2N
Li
H
O
R
OEt
62; ester
H
OLi
R
OEt
OEt
69a
69
+ R2NH
70
Lithium enolates, enamines and silyl enol ethers
Direct alkylation of lithium enolates of esters9 62 and lactones 73, via the lithium enolates 71 and
74, with alkyl halides is usually successful.
H
1. LDA
R1
O
R1
CO2Et
Br
2. R 2Br
OEt
R2
62
R2
Li
R1
CO2Et
71
O
O
1. LDA
72
OLi
O
O
O
2. RBr
R
73
74
75
18
2 Chemoselectivity
More impressive and more important is the performance of these lithium enolates in aldol
reactions. Ester enolates are awkward things to use in reactions with enolisable aldehydes and
ketones because of the very efficient self-condensation of the aldehydes and ketones. The traditional
solutions involve such devices as Knoevenagel-style reactions with malonates.11 Lithium enolates
of esters, e.g. 76, react cleanly with enolisable aldehydes and ketones to give high yields of aldols,12
e.g. 79 in a single step also involving a six-membered cyclic transition state 77.
O
MeCO2Et
LDA
Li
O
OLi
+
THF
–78 °C
O
OEt
OEt
76
77
OLi
OH
CO2Et
H
CO2Et
H2O
79, 79-90%
78
They even react cleanly with formaldehyde, thus solving the problem that the Mannich reaction
is not applicable to esters. The synthesis of the exo-methylene lactone 80 can be accomplished
this way. Enone disconnection13 reveals formaldehyde as the electrophilic component in a crossed
aldol reaction, realised with a lithium enolate 82.14 The mono-adduct 83 of formaldehyde and the
lactone 81 can be isolated and the cautious dehydration step is to avoid migration of the double
bond into the ring.
H
H
α,β-unsaturated
carbonyl
O
O
O + CH2=O
O
H
H
80
81
H
81
H
O
LDA
O
CH2O
1. MsCl, Et 3N
80
O
OLi
2. pyridine
H
H
82
OH
83
The same technique can even be applied to carboxylic acids themselves 84 providing two
molecules of base are used. The first removes the acid proton to give the lithium salt 85 and the
second forms the lithium enolate 86.
O
R
OH
84
LDA
O
R
OLi
85
LDA
OLi
R
OLi
2 x LDA
–78 ˚C, THF
84
86
These lithium derivatives are also well behaved in alkylations and aldol reactions. Krapcho’s
synthesis15 of the sesquiterpene α-curcumene 92 starts with the chemoselective condensation of
2 Chemoselectivity in Enol and Enolate Formation
19
the dilithium derivative of the acid 87 with the enolisable aldehyde 89. The aldol product 90 is
converted into the β-lactone 91 and hence by heating and loss of CO2 into α-curcumene 92.
2 x LDA
OLi
CO2H
OLi
88
87
87
CHO
1. 2 x LDA
89
PhSO2Cl
90
73% yield
2. 89
CO2H
pyridine
OH
90, 73% yield
140 °C
O
O
91, 77% yield
92; α-curcumene, 90% yield
You might be forgiven for thinking that lithium enolates solve all problems of enolate chemoselectivity at a stroke and wonder why they are not always used. They are very widely used, but
they require strictly anhydrous conditions at low temperatures (usually ⫺78 ⬚C, the temperature
of a dry ice/acetone bath) and no-one in their right mind would use these conditions if mixing
the reagents in ethanol at room temperature with a catalytic amount of NaOH did nearly as well.
These are the conditions of many simple aldol reactions and are preferred where practical, particularly in industrial practice. The intermediate 93 was needed in a synthesis of geiparvirin. The
best aldol disconnection in the middle of the molecule gives a ketone 94, that must be enolised in
the only possible position, and then react with an unenolisable and more electrophilic aldehyde 95.
No selectivity problems arise and an equilibrating aldol reaction between 94 and 95 catalysed by
NaOEt in EtOH gives 93 in 89% yield.16
O
O
H
aldol
Ph
HO
93
+ O
HO
94
Ph
95
Enamines
Lithium enolates do not even solve all problems of chemoselectivity: most notoriously, they fail
when the specific enolates of aldehydes are needed. The problem is that aldehydes self-condense
so readily that the rate of the aldol reaction can be comparable with the rate of enolate formation
by proton removal. Fortunately there are good alternatives. Earlier in this chapter we showed
examples of what can go wrong with enamines. Now we can set the record straight by extolling the
virtues of the enamines 96 of aldehydes.17 They are easily made without excessive aldol reaction as
they are much less reactive than lithium enolates, they take part well in reactions such as Michael
additions, a standard route to 1,5-dicarbonyl compounds, e.g. 97.18
20
2 Chemoselectivity
CHO
NR2
NR2
R2NH
OMe
96
OMe
R
R
R
O
O
NR2
OMe
R
CHO
H
R
H2O
O
CO2Me
97
An impressive example19 is the Robinson annelation of the unsaturated aldehyde 98 where
neither aldol reaction nor double bond migration in the enamine 99 interferes. The 1,5-dicarbonyl
compound 100 cyclises spontaneously to the enone 101.
O
O
O
CHO
N
+
Ph
OHC
N
H
Ph
Ph
98
Ph
100
99
101
Silyl enol ethers
For all their usefulness, enamines have now largely been superseded by silyl enol ethers. These
(102-104) can be made directly with Me3SiCl from the lithium enolates of esters or acids or
from aldehydes under milder conditions with a tertiary amine. The silicon atom is an excellent
electrophile with a strong preference for more electronegative partners and it combines with the
oxygen atom of an enolate so rapidly that no self condensation occurs even with aldehydes.
Silyl Enol Ethers of Esters
O
LDA
R
R
OEt
OLi
Me3SiCl
OEt
102
OEt
Silyl Enol Ethers of Acids
O
LDA
R
R
OH
OLi
Me3SiCl
OSiMe3
103
OLi
O
H
Et3N
OH
R
H
OSiMe3
R
Silyl Enol Ethers of Aldehydes
R
OSiMe3
R
Me3SiCl
OSiMe3
R
H
104
The silyl enol ethers 102 and 104 are shown as single geometrical isomers for convenience:
in fact they are normally formed as mixtures, though this does not usually affect their reactions.
They are thermodynamically stable compounds but are easily hydrolysed with water or methanol
and are usually prepared when they are needed. They are much less reactive than lithium enolates,
or even enamines, and their reactions with electrophiles are best catalysed by Lewis acids, often
2 Chemoselectivity in Enol and Enolate Formation
21
TiCl4. The aldehydes 105 and 108, the one branched and the other not, are simply converted
into their silyl enol ethers 106 and 109 and combined with two different enolisable aldehydes to
give high yields of aldol products 107 and 110 without any self condensation of any of the four
aldehydes or any cross-condensation the wrong way round.20
Ph
Me3SiCl
CHO
CHO
OSiMe3
Et3N
Ph
TiCl4
CHO
OH
107; 86% yield
106
105
CHO
Me3SiCl
Ph
CHO
Ph
OSiMe3
n-PrCHO
Ph
TiCl4
Et3N
OH
108
110; 78% yield
109
The silyl enol ethers of esters, e.g. 111 and lactones, e.g. 114 similarly take part in efficient aldol
reactions with enolisable aldehydes and ketones with Lewis acid catalysis, again with complete
regioselectivity. Example 113 is particularly impressive as the very enolisable ketone gives a high
yield of an aldol product with two adjacent quaternary centres.21
Ph
OSiMe3
1. LDA
CO2Et
2. Me 3SiCl
OEt
111
O
Ph
TiCl4
113, 94% yield
112
O
Me3SiO
O
1. LDA
O
2. Me 3SiCl
114
CO2Et
OH
+
O
TiCl4
O
O
OH
115
116
Synthesis of the ant alarm pheromone mannicone
The synthesis of the ant alarm pheromone mannicone 117 is a good example. Enone disconnection
reveals that we need a crossed aldol condensation between the symmetrical ketone 118, acting as
the enol component, and the enolisable aldehyde 119.
Mannicone: Analysis
O
O
H
aldol
+ O
enone
117
118
119
The ketone gives a mixture of geometrical isomers of the silyl enol ether 120 which condense
with the aldehyde 119 to give the aldol 121 as a mixture of diastereoisomers which dehydrates to
mannicone 117 in acid.22 It is particularly impressive that the optically active aldehyde 119 has its
22
2 Chemoselectivity
stereogenic centre at the enol position and yet optically active mannicone is formed by this route
without racemisation.
Mannicone: Synthesis
OSiMe3
Me3SiCl
118
O
OH
TsOH
119
Et3N
TiCl4
120
TM 117
121
We shall discuss further aspects of the aldol reaction in the next two chapters where we shall
see how to control the enolisation of unsymmetrical ketones, and how to control the stereochemistry of aldol products such as 121. We shall return to a more comprehensive survey of specific
enol equivalents in chapter 10. In this chapter we are concerned to establish that chemoselective
enolisation of esters, acids, aldehydes, and symmetrical ketones can be accomplished with lithium
enolates, enamines, or silyl enol ethers, and we shall be using all these intermediates extensively
in the rest of the book.
Examples of Chemoselective Reactions in Synthesis
Synthesis of lipstatin
The synthesis of lipstatin 122 is too complex to discuss here in detail but an early stage in one
synthesis uses a clever piece of chemoselectivity.23 Kocienski planned to make the β-lactone by
a cycloaddition with the ketene 124 and to add the amino acid side chain 123 by a Mitsunobu
reaction involving inversion. They therefore needed Z,Z-125 to join these pieces together. This
was to be made in turn by a Wittig reaction from 126. The problem now is that 126 is symmetrical
and cannot carry stereochemistry and that aldehydes are needed at both ends.
•
123
O
O
O
O
NHCHO
NHCHO
OH
O
124
CO2H
Me3Si
CHO
H
H
122; lipstatin
OH
OHC
Z,Z-125
CHO
126
The way they solved the problem was this. (S)-(⫺)-Malic acid is available cheaply. Its dimethyl
ester 127 could be chemoselectively reduced by borane to give 128. Normally borane does not
reduce esters and clearly the borane first reacts with the OH group and then delivers hydride to the
nearer carbonyl group. The primary alcohol was chemoselectively tosylated 129 and the remaining (secondary) OH protected with a silyl group 130 (TBDMS stands for t-butyldimethylsilyl and
is sometimes abbreviated to TBS). Now the remaining ester can be reduced to an aldehyde 131
and protected 132. Displacement of tosylate by cyanide puts in the extra carbon atom 133 and
reduction gives 134, that is the dialdehyde 126 in which one of the two aldehydes is protected.
This compound was used in the successful synthesis of lipstatin.
2 Examples of Chemoselective Reactions in Synthesis
OH
MeO2C
OH
BH3•SMe2
NaBH4
CO2Me
OH
TsCl
HO
CO2Me
127; (S)-(–)-malic acid
dimethyl ester
128; 83% yield
OTBDMS
TsO
imidazole
OTBDMS
TsO
CH2Cl2
130; 92% yield
OTBDMS
DMSO
(i-PrO)3CH
CHO
TsOH
131; 78% yield
NaCN
CH(Oi-Pr)2
TsO
CO2Me
129; 75% yield
DIBAL
CO2Me
OTBDMS
TsO
pyridine
CH2Cl2
THF
t-BuMe2SiCl
23
NC
CH(Oi-Pr)2
132; 92% yield
OTBDMS
DIBAL
OHC
CH2Cl2
133; 83% yield
CH(Oi-Pr)2
134; 81% yield
The synthesis of rubrynolide
The synthesis of the natural product rubrynolide from th
Organic Synthesis:
Strategy and Control
Paul Wyatt
Senior Lecturer and Director of Undergraduate Studies, School of Chemistry,
University of Bristol, UK
and
Stuart Warren
Reader in Organic Chemistry, Department of Chemistry,
University of Cambridge, UK
Copyright © 2007
John Wiley & Sons Ltd,
The Atrium, Southern Gate, Chichester,
West Sussex PO19 8SQ, England
Telephone (+44) 1243 779777
Email (for orders and customer service enquiries): [email protected]
Visit our Home Page on www.wiley.com
All Rights Reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any
means, electronic, mechanical, photocopying, recording, scanning or otherwise, except under the terms of the Copyright, Designs
and Patents Act 1988 or under the terms of a licence issued by the Copyright Licensing Agency Ltd, 90 Tottenham Court Road,
London WIT 4LP, UK, without the permission in writing of the Publisher. Requests to the Publisher should be addressed to the
Permissions Department, John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex PO19 8SQ, England, or
emailed to [email protected], or faxed to (+44) 1243 770620.
Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product names
used in this book are trade names, service marks, trademarks or registered trademarks of their respective owners. The Publisher is
not associated with any product or vendor mentioned in this book.
This publication is designed to provide accurate and authoritative information in regard to the subject matter covered. It is sold on the
understanding that the Publisher is not engaged in rendering professional services. If professional advice or other expert assistance is
required, the services of a competent professional should be sought.
Other Wiley Editorial Offices
John Wiley & Sons Inc., 111 River Street, Hoboken, NJ 07030, USA
Jossey-Bass, 989 Market Street, San Francisco, CA 94103-1741, USA
Wiley-VCH Verlag GmbH, Boschstr. 12, D-69469 Weinheim, Germany
John Wiley & Sons Australia Ltd, 42 McDougall Street, Milton, Queensland 4064, Australia
John Wiley & Sons (Asia) Pte Ltd, 2 Clementi Loop #02-01, Jin Xing Distripark, Singapore 129809
John Wiley & Sons Canada Ltd, 6045 Freemont Blvd, Mississauga, ONT, L5R 4J3, Canada
Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in
electronic books.
Anniversary Logo Design: Richard J. Pacifico
Library of Congress Cataloging-in-Publication Data
Wyatt, Paul.
Organic synthesis: strategy and control / Paul Wyatt and Stuart Warren.
p. cm.
Includes bibliographical references.
ISBN: 978-0-470-48940-5
ISBN: 978-0-471-92963-5
1. Organic compounds – Synthesis. 2. Stereochemistry. I. Warren, Stuart
G. II. Title.
QD262.W89 2007
547´.2–dc22
2006034932
British Library Cataloguing in Publication Data
A catalogue record for this book is available from the British Library
ISBN: 978-0-471-48940-5 (HB)
ISBN: 978-0-471-92963-5 (PB)
Typeset in 10/12pt Times by Thomson Digital
Printed and bound in Great Britain by Antony Rowe Ltd, Chippenham, Wiltshire
This book is printed on acid-free paper responsibly manufactured from sustainable forestry
in which at least two trees are planted for each one used for paper production.
Contents
Preface
vii
A: Introduction: Selectivity
1. Planning Organic Syntheses: Tactics, Strategy and Control
2. Chemoselectivity
3. Regioselectivity: Controlled Aldol Reactions
4. Stereoselectivity: Stereoselective Aldol Reactions
5. Alternative Strategies for Enone Synthesis
6. Choosing a Strategy: The Synthesis of Cyclopentenones
3
9
27
43
55
71
B: Making Carbon–Carbon Bonds
7. The Ortho Strategy for Aromatic Compounds
8. σ-Complexes of Metals
9. Controlling the Michael Reaction
10. Specific Enol Equivalents
11. Extended Enolates
12. Allyl Anions
13. Homoenolates
14. Acyl Anion Equivalents
91
113
127
139
155
173
189
203
C: Carbon–Carbon Double Bonds
15. Synthesis of Double Bonds of Defined Stereochemistry
16. Stereo-Controlled Vinyl Anion Equivalents
17. Electrophilic Attack on Alkenes
18. Vinyl Cations: Palladium-Catalysed C–C Coupling
19. Allyl Alcohols: Allyl Cation Equivalents (and More)
223
255
277
307
339
D: Stereochemistry
20. Control of Stereochemistry – Introduction
21. Controlling Relative Stereochemistry
22. Resolution
23. The Chiral Pool — Asymmetric Synthesis with Natural Products as
Starting Materials —
24. Asymmetric Induction I Reagent-Based Strategy
25. Asymmetric Induction II Asymmetric Catalysis: Formation of C–O
and C–N Bonds
26. Asymmetric Induction III Asymmetric Catalysis: Formation of C–H
and C–C Bonds
27. Asymmetric Induction IV Substrate-Based Strategy
371
399
435
465
505
527
567
599
vi
Contents
28. Kinetic Resolution
29. Enzymes: Biological Methods in Asymmetric Synthesis
30. New Chiral Centres from Old — Enantiomerically Pure
Compounds & Sophisticated Syntheses —
31. Strategy of Asymmetric Synthesis
E: Functional Group Strategy
32. Functionalisation of Pyridine
33. Oxidation of Aromatic Compounds, Enols and Enolates
34. Functionality and Pericyclic Reactions: Nitrogen Heterocycles by
Cycloadditions and Sigmatropic Rearrangements
35. Synthesis and Chemistry of Azoles and other Heterocycles with Two
or more Heteroatoms
36. Tandem Organic Reactions
627
651
681
717
749
777
809
835
863
General References
893
Index
895
Preface
We would like to thank those who have had the greatest influence on this book, namely the undergraduates at the Universities of Bristol and Cambridge. But, particularly we would like to thank
the organic chemists at Organon (Oss), AstraZeneca (Alderley Park, Avlon Works, Mölndal and
Macclesfield), Lilly (Windlesham), Solvay (Weesp) and Novartis (Basel) who contributed to the
way the book was written more than they might realise. These chemists will recognise material
from our courses on The Disconnection Approach, Advanced Heterocyclic Chemistry, New Synthetic Methods and Asymmetric Synthesis. Additionally we would like to thank the participants
at the SCI courses organised by the Young Chemists Panel. All these industrial chemists participated in our courses and allowed us to find the best way to explain concepts that are difficult to
grasp. This book has changed greatly over the ten years it was being written as we became more
informed over what was really needed. The book is intended for that very audience – final year
undergraduates, graduate students and professional chemists in industry.
PJW
SGW
July 2006
Section A:
Introduction: Selectivity
1. Planning Organic Syntheses: Tactics, Strategy and Control ............................................. 3
2. Chemoselectivity ..................................................................................................................... 9
3. Regioselectivity: Controlled Aldol Reactions ..................................................................... 27
4. Stereoselectivity: Stereoselective Aldol Reactions ............................................................. 43
5. Alternative Strategies for Enone Synthesis ........................................................................ 55
6. Choosing a Strategy: The Synthesis of Cyclopentenones .................................................. 71
1
Planning Organic Syntheses:
Tactics, Strategy and Control
The roll of honour inscribed with successful modern organic syntheses is remarkable for the
number, size, and complexity of the molecules made in the last few decades. Woodward and
Eschenmoser’s vitamin B12 synthesis,1 completed in the 1970s, is rightly regarded as a pinnacle of
achievement, but since then Kishi2 has completed the even more complex palytoxin. The smaller
erythromycin and its precursors the erythronolides3 1, and the remarkably economical syntheses of
the possible stereoisomers of the cockroach pheromones 2 by Still4 deal with a greater concentration
of problems.
O
O
O
HO
1
erythronolide A
OH
OH
2
periplanone-B
O
O
OH
O
OH
Less applauded, but equally significant, is the general advance in synthetic methods and their
industrial applications. AstraZeneca confess that it took them nearly a century to bring Victor
Grignard’s methods into use, but are proud that Corey’s sulfur ylid chemistry made it in a decade.
Both are used in the manufacture of the fungicide flutriafol5 3.
Me
F
S
CH2
O
F
O
F
F
Me
F
N
RCO3H
F
N
N N
N
N
base
F
F
F
Br2
F
X
O
MgI
OH
F
F
AlCl3
+
H2O
F
3
Flutriazole
fungicide
OH
Cl
COCl
Cl
Optically active and biodegradable deltamethrin6 4 has taken a large share of the insecticide
market, and asymmetric hydrogenation is used in the commercial synthesis of DOPA 5 used to
treat Parkinson’s disease.7 These achievements depend both on the development of new methods
and on strategic planning:8 the twin themes of this book.
Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren
Copyright © 2007 John Wiley & Sons, Ltd
4
1 Planning Organic Syntheses: Tactics, Strategy and Control
Br
HO
R R
O S
Br
O
NH2
O
H
CO2H
HO
CN
4
deltamethrin
5
(S)-L-(–)-DOPA
To make any progress in this advanced area, we have to assume that you have mastered the
basics of planning organic synthesis by the disconnection approach, roughly the material covered
in our previous books.9 There, inspecting the target molecule, identifying the functional groups,
and counting up the relationships between them usually gave reliable guidelines for a logical
synthesis. All enones were tackled by some version of the aldol reaction; thus 6 would require the
attack of enolate 7 on acetone. We hope you already have the critical judgement to recognise that
this would need chemoselectivity in enolising 7 rather than acetone or 6, and regioselectivity in
enolising 7 on the correct side.
O
O
aldol
O
+
O
6
7
In this book we shall explore two new approaches to such a problem. We shall see how to make
specific enol equivalents for just about any enolate you might need, and we shall see that alternative disconnections such as 6a, the acylation of a vinyl anion 8, can be put into practice. Another
way to express the twin themes of this book is strategy and control: we solve problems either
by finding an alternative strategy or by controlling any given strategy to make it work. This will
require the introduction of many new methods - a whole chapter will be devoted to reagents for
vinyl anions such as 8, and this will mean exploring modern organometallic chemistry.
O
O
+
6a
X
8
9
We shall also extend the scope of established reactions. We hope you would recognise the aldol
disconnection in TM 10, but the necessary stereochemical control might defeat you. An early
section of this book describes how to control every aspect of the aldol reaction: how to select
which partner, i.e. 11 or 12, becomes an enolate (chemoselectivity), how to control which enolate
of the ketone 12 is formed (regioselectivity), and how to control the stereochemistry of the product
10 (stereoselectivity). As we develop strategy, we shall repeatedly examine these three aspects of
control.
OH
O
O
R2
R1
?
10
O
aldol
R1
H
11
+
O
R2
R2
?
12
The target molecules we shall tackle in this book are undoubtedly more difficult in several ways
than this simple example 10. They are more complex quantitatively in that they combine functional
1 Planning Organic Syntheses: Tactics, Strategy and Control
5
groups, rings, double bonds, and chiral centres in the same target, and qualitatively in that they
may have features like large rings, double bonds of fixed configuration, or relationships between
functional groups or chiral centres which no standard chemistry seems to produce. Molecules 1 to
5 are examples: a quite different one is flexibilene 13, a compound from Indonesian soft coral. It
has a fifteen-membered ring, one di- and three tri-substituted double bonds, all E but none
conjugated, and a quaternary centre. Mercifully there are no functional groups or chiral centres.
How on earth would you tackle its synthesis? One published synthesis is by McMurry.10
OMe
1. 2 x BuLi
H
2.
Cp2ZrHCl
1. PCC
HO
O
14
15; 78% yield
2. H ,
CH(OMe)3
17
MeO
16; 83% yield
OMe
MeO
17
O
ZrCp2Cl
18
Pd(OCOCF3)2
TiCl3
OHC
O
Zn/Cu
13; flexibilene, 52% yield
19; 76% yield from 16 + 18
This short synthesis uses seven metals (Li, Cr, Zr, Pd, Ti, Zn, and Cu), only one protecting
group, achieves total control over double bond geometry, remarkable regioselectivity in the Zr-Pd
coupling reaction, and a very satisfactory large ring synthesis. The yield in the final step (52%)
may not look very good, but this is a price worth paying for such a short synthesis. Only the first
two steps use chemistry from the previous books: all the other methods were unknown only ten
years before this synthesis was carried out but we shall meet them all in this book.
An important reason for studying alternative strategies (other than just making the compound!)
is the need to find short cheap large scale routes in the development of research lab methods into
production. All possible routes must be explored, at least on paper, to find the best production
method and for patent coverage. Many molecules suffer this exhaustive process each year, and
some sophisticated molecules, such as Merck’s HIV protease inhibitor 20, a vital drug in the fight
against AIDS, are in current production on a large scale because a good synthesis was found by
this process.11
N
OH
Ph
H
N
N
N
t-BuNH
O
OH
O
20; Crixivan
You might think that, say organometallic chemistry using Zr or Pd would never be used in
manufacture. This is far from true as many of these methods are catalytic and the development
of polymer-supported reagents for flow systems means that organo-metallic reagents or enzymes
may be better than conventional organic reagents in solution with all the problems of by-product
disposal and solvent recovery. We shall explore the chemistry of B, Si, P, S, and Se, and of metals
6
1 Planning Organic Syntheses: Tactics, Strategy and Control
such as Fe, Co, Ni, Pd, Cu, Ti, Sn, Ru and Zr because of the unique contribution each makes to
synthetic methods.
In the twenty years since McMurry’s flexibilene synthesis major developments have changed
the face of organic synthesis. Chiral drugs must now be used as optically pure compounds and
catalytic asymmetric reactions (chapters 25 and 26) have come to dominate this area, an achievement crowned by the award of the 2001 Nobel prize for Chemistry to Sharpless, Noyori and
Knowles. Olefin metathesis (chapter 15) is superseding the Wittig reaction. Palladium-catalysed
coupling of aromatic rings to other aromatic rings, to alkenes, and to heteroatoms (chapter 18)
makes previously impossible disconnections highly favourable. These and many more important
new methods make a profound impact on the strategic planning of a modern synthesis and find
their place in this book.
A Modern Synthesis: Fostriecin (CI-920)
The anti-cancer compound Fostriecin 21 was discovered in 1983 and its stereochemistry elucidated in 1997. Not until 2001 was it synthesised and then by two separate groups.12 Fostriecin is
very different from flexibilene. It still has alkene geometry but it has the more challenging threedimensional chirality as well. It has plenty of functionality including a delicate monophosphate
salt. A successful synthesis must get the structure right, the geometry of the alkenes right, the
relative stereochemistry right, and it must be made as a single enantiomer.
O
HO
O
O
P
Na
OH
OH
O
O
OH
21; Fostriecin (CI-920)
The brief report of Jacobsen’s total synthesis starts with a detailed retrosynthetic analysis.
The compound was broken into four pieces 21a after removal of the phosphate. The unsaturated
lactone 24 (M is a metal) could be made by an asymmetric oxo-Diels-Alder reaction from diene 22
and ynal 23. The epoxide 25 provides a second source of asymmetry. One cis alkene comes from
an alkyne 26 and the rest from a dienyl tin derivative 27.
O O
P
HO
O
21a
Disconnection of
Fostriecin (CI-920)
O
Na
OH
OH
O
OH
OR
DielsAlder
RO
O
O
22
O
23
SiMe3
24
S
SnBu3
H
M
O
S
O
25
26
SiMe3
27
The synthesis is a catalogue of modern asymmetric catalytic methods. The epoxide 25 was
resolved by a hydrolytic kinetic resolution (chapter 28) using a synthetic asymmetric cobalt complex. The asymmetric Diels-Alder reaction (chapter 26) was catalysed by a synthetic chromium
1 References
7
complex. The vinyl metal derivative 24 was made by hydrozirconation of an alkyne (this at least
is similar to the flexibilene synthesis) and the secondary alcohol chiral centre was derived from
the dithian 26 by hydrolysis to a ketone and asymmetric reduction with a synthetic ruthenium
complex (chapter 24). The dienyl tin unit 27 was coupled to the rest of the molecule using catalytic
palladium chemistry (chapter 18). Almost none of these catalytic methods was available in 1983
when flexibilene was made and such methods are a prominent feature of this book. Organic synthesis nowadays can tackle almost any problem.13
Please do not imagine that we are abandoning the systematic approach or the simpler reagents
of the previous books. They are more essential than ever as new strategy can be seen for what
it is only in the context of what it replaces. Anyway, no-one in his or her right mind would use
an expensive, toxic, or unstable reagent unless a friendlier one fails. Who would use pyrophoric
tertiary butyl-lithium in strictly dry conditions when aqueous sodium hydroxide works just as
well? In most cases we shall consider the simple strategy first to see how it must be modified. The
McMurry flexibilene synthesis is unusual in deploying exotic reagents in almost every step. A
more common situation is a synthesis with one exotic reagent and six familiar ones. The logic of
the previous books is always our point of departure.
The organisation of the book
The book has five sections:
A:
B:
C:
D:
E:
Introduction, selectivity, and strategy
Making Carbon-Carbon bonds
Carbon-Carbon double bonds
Stereochemistry
Functional Group Strategy
The introductory section uses aldol chemistry to present the main themes in more detail and
gives an account of the three types of selectivity: chemo-, regio-, and stereo-selectivity. We shall
explore alternative strategies using enones as our targets, and discuss how to choose a good route
using cyclopentenones as a special case among enones. Each chapter develops strategy, new
reagents, and control side-by-side. To keep the book as short as possible (like a good synthesis),
each chapter in the book has a corresponding chapter in the workbook with further examples,
problems, and answers. You may find that you learn more efficiently if you solve some problems
as you go along.
References
General references are given on page 893
1. R. B. Woodward, Pure Appl. Chem., 1973, 33, 145; A. Eschenmoser and C. E. Wintner, Science, 1977,
196, 1410; A. Eschenmoser, Angew. Chem., Int. Ed. Engl., 1988, 27, 5.
2. Y. Kishi, Tetrahedron, 2002, 58, 6239.
3. E. J. Corey, K. C. Nicolaou, and L. S. Melvin, J. Am. Chem. Soc., 1975, 97, 654; G. Stork and S. D.
Rychnovsky, J. Am. Chem. Soc., 1987, 109, 1565; Pure Appl. Chem., 1987, 59, 345; A. F. Sviridov, M. S.
Ermolenko, D. V. Yashunsky, V. S. Borodkin and N. K. Kochetkov, Tetrahedron Lett., 1987, 28, 3835,
and references therein.
4. W. C. Still, J. Am. Chem. Soc., 1979, 101, 2493. See also S. L. Schreiber and C. Santini, J. Am. Chem.
Soc., 1984, 106, 4038; T. Takahashi, Y. Kanda, H. Nemoto, K. Kitamura, J. Tsuji and Y. Fukazawa,
J. Org. Chem., 1984, 51, 3393; H. Hauptmann, G. Mühlbauer and N. P. C. Walker, Tetrahedron Lett.,
1986, 27, 1315; T. Kitahara, M. Mori and K. Mori, Tetrahedron, 1987, 43, 2689.
8
1 Planning Organic Syntheses: Tactics, Strategy and Control
5. P. A. Worthington, ACS Symposium 355, Synthesis and Chemistry of Agrochemicals, eds D. R. Baker,
J. G. Fenyes, W. K. Moberg, and B. Cross, ACS, Washington, 1987, p 302.
6. M. Elliott, A. W. Farnham, N. F. Janes, P. H. Needham, and D. A. Pullman, Nature, 1974, 248, 710;
M. Elliott, Pestic. Sci., 1980, 11, 119.
7. J. Halpern, H. B. Kagan, and K. E. Koenig, Morrison, vol 5, pp 1–101.
8. Corey, Logic; Nicolaou and Sorensen.
9. Designing Syntheses, Disconnection Textbook, and Disconnection Workbook.
10. J. McMurry, Acc. Chem. Res., 1983, 16, 405.
11. D. Askin, K. K. Eng, K. Rossen, R. M. Purick, K. M. Wells, R. P. Volante and P. J. Reider, Tetrahedron
Lett., 1994, 35, 673; B. D. Dorsey, R. B. Levin, S. L. McDaniel, J. P. Vacca, J. P. Guare, P. L. Darke,
J. A. Zugay, E. A. Emini, W. A. Schleif, J. C. Quintero, J. H. Lin, I.-W. Chen, M. K. Holloway, P. M. D.
Fitzgerald, M. G. Axel, D. Ostovic, P. S. Anderson and J. R. Huff, J. Med. Chem., 1994, 37, 3443.
12. D. L. Boger, S. Ichikawa and W. Zhong, J. Am. Chem. Soc., 2001, 123, 4161; D. E. Chavez and E. N.
Jacobsen, Angew. Chem., Int. Ed., 2001, 40, 3667.
13. D. Seebach, Angew. Chem. Int. Ed., 1990, 29, 1320; K. C. Nicolaou, E. J. Sorensen and N. Winssinger,
J. Chem. Ed., 1998, 75, 1225.
2 Chemoselectivity
Definitions
Introduction: three types of control
Chemoselectivity: simple examples and rules
Chemoselectivity by Reactivity and Protection: An anti-Malaria Drug
Protection to allow a less reactive group to react
When Protection is not Needed
Dianions: wasting reagent to achieve selectivity
Chemoselectivity by Reagent: The Pinacol Rearrangement
Selectivity between secondary and tertiary alcohols by reagent
Corey’s longifolene synthesis
Chemoselectivity in Enol and Enolate Formation
General discussion of enols and enolates
Formation of specific enol equivalents
Lithium enolates, enamines and silyl enol ethers
Enamines
Silyl enol ethers
Synthesis of the ant alarm pheromone mannicone
Examples of Chemoselectivity in Synthesis
Synthesis of lipstatin, rubrynolide and hirsutene
Definitions
Introduction: three types of control
Behind all grand strategic designs in organic synthesis must lie the confidence that molecules can
be compelled to combine in the ways that we require. We shall call this control and divide it into three
sections by mechanistic arguments. These sections are so important that we shall devote the next three
chapters to the more detailed explanation of just what the divisions mean. If you can recognise what
might go wrong you are in a better position to anticipate the problem and perhaps avoid it altogether.
Our three types of control are over chemoselectivity (selectivity between different functional groups),
regioselectivity (control between different aspects of the same functional group), and stereoselectivity
(control over stereochemistry). Examples of selectivity of all three kinds are given in The Disconnection
Approach: Chemoselectivity in chapter 5, Regioselectivity in chapter 14, and Stereoselectivity in
chapters 12 and 38. These aspects will not be addressed again in the present book.
Organic Synthesis: Strategy and Control, Written by Paul Wyatt and Stuart Warren
Copyright © 2007 John Wiley & Sons, Ltd
10
2 Chemoselectivity
Chemoselectivity: simple examples and rules
Chemoselectivity is the most straightforward of the three types and might seem too elementary
to appear in an advanced textbook. Counting the number of protecting groups in the average
synthesis reveals this as a naive view. Selectivity between functional groups might involve:
(a) Selective reaction of one among several functional groups of different reactivity, as in the
reduction of the keto-acid 2 to give either product 1 or 3 at will.
O
O
OH
?
OH
?
CO2H
1
CO2H
2
3
(b) Selective reaction of one of several identical functional groups, as in the conversion of the
symmetrical diacid 5 to the half ester, half acid chloride 4, or the lactone 6 in which one of the two
acids has been reduced. There is a more subtle example of this at the end of the chapter.
?
COCl
MeO2C
CO2H
HO2C
substitution
4
?
reduction
O
6
5
O
(c) Selective reaction of a functional group to give a product which could itself react with the
same reagent, as in the classical problem of making a ketone 8 from an acid derivative 7 without
getting the alcohol 9 instead.
O
R1
OEt
OH
O
R2–Metal
and not
R2
R1
?
8
7
R1
R2
R2
9
Organic chemists are developing ever more specific reagents to do these jobs. These reagents
must carry out the reaction they are designed for and must not:
(i) react with themselves.
(ii) react with functional groups other than the one they are aimed at.
(iii) react with the product.
Proviso (ii) is obvious, but (i) and (iii) perhaps need some explanation. It seems hardly worth
stating that a reagent should not react with itself, but it is only too easy to suggest using a reagent
such as 11 without realising that the organo-metallic reagent will act as a base for its own hydroxyl
group 12 and destroy itself. The traditional solution to this problem is protection of the OH group
in 10 but ideally we should like to avoid protection altogether though this is not yet possible.
Br
HO
Mg
MgBr
HO
10
HO
11
n-butanol
H , H 2O
O
H
12
MgBr
O
CH3
13
work-up
OH
CH3
n-butanol
2 Definitions
11
Proviso (iii) is more obvious and yet perhaps more often catches people out. It is not always clear
in exactly what form the product is produced in the reaction mixture, though a good mechanistic
understanding and careful thought should reveal this. The reaction between the simple aldehyde
14 and chloral (Cl3C.CHO) looks like a straightforward route to the aldol 17, and might reasonably
be carried out via the enamine 16.
CHO
+
HN
O
14
O
15
OH
Cl3C.CHO
N
CHO
Cl3C
?
17
16
However, mixtures of 16 and chloral, in any proportion, give only the 2:1 adduct 20 which can
be isolated in 83% yield.1 Obviously the immediate product 19 reacts with chloral at least as fast
as does 16. Fortunately the synthesis can be rescued by acid-catalysed cleavage of 20 with HCl
which gives a good yield of the target 17.
CCl3
CCl3
Cl3C
O
O
O
Cl3C.CHO
N
Cl3C
N
O
O
N
O
O
18
19
20
Enamines are excellent at Michael additions and another plausible synthesis which “goes
wrong” is the addition of acrolein to cyclohexanone mediated by the enamine 21 formed this time
with pyrollidine.
O
O
CHO
N
+
HN
?
CHO
21
22
If the product is isolated by distillation, a good yield (75%) of the bicyclic ketone 23 is obtained.2
A more detailed investigation disclosed that 24 is the immediate product, that 23 is formed from
it on distillation, and that the expected Michael adduct 22 can be isolated in good yield simply
by the hydrolysis of 24. In other words, don’t distil! If things “go wrong” in a synthesis, this may
be a blessing, as here. There are lots of ways to control Michael additions, but few ways to make
bicyclic ketones like 23, and this is now a standard method.3 The moral is to make sure you know
what is happening, and to be prepared to welcome the useful and unexpected result.
O
N
23
N
O
24
N
12
2 Chemoselectivity
Chemoselectivity by Reactivity and Protection: An anti-Malaria Drug
We need to see some of these principles in action and a proper synthesis is overdue. The anti-malarial
drug amopyroquine 25 might have been derived from quinine as it has a quinoline nucleus. It also has
five functional groups – three amines (all different - one aromatic, one tertiary, and one secondary), a
phenol and an aryl chloride. There are four rings, three aromatic and one saturated heterocyclic.
OH
OH
a
N
HN
b
N
Cl
N
Cl
25
Amopyroquine
N
HN
25a, b
Amopyroquine disconnections
There are many possible disconnections, but we should prefer to start in the middle of the molecule
to achieve the greatest simplification. Disconnection 25a would require a nucleophilic displacement
(X ⫽ a leaving group) on an unactivated benzene ring 27 and looks unpromising. Disconnection 25b
requires nucleophilic displacement at position 4 in a pyridine ring, an acceptable reaction because of
the electron-withdrawing effect of the nitrogen atom in the ring, so this is the better route, though we
may be apprehensive about controlling the chemoselectivity as there are three potential nucleophiles
in 26 and two potential electrophiles in 28.
X
+
+
25
N
Cl
28
OH
a
b
N
H2N
N
Cl
NH2
OH
N
X
29
26
27
Further disconnections of 26 by the Mannich reaction4 and of 28 by standard heterocyclic
methods give simple starting materials.5
OH
OH
Mannich
+ CH2O +
N
H2N
H2N
HN
30
26a
X
O
O
quinoline
synthesis
FGI
+
N
Cl
28
N
H
Cl
31
HO
NH2
Cl
32
Protection to allow a less reactive group to react
Now the fun begins! Attempted Mannich reaction on the aminophenol 30 would be dominated by the
more nucleophilic NH2 group and is no good. Acylation moderates the NH2 group by delocalisation
2 When Protection is not Needed
13
and 33 is a good choice for starting material as it is paracetamol, the common analgesic. Mannich
reaction now chemoselectively gives 34 and alkaline hydrolysis of the amide gives 26.
OH
CH2O
··
N
H
AcOH
H2N
OH
OH
O
Ac2O
pyrrolidine
30
N
AcHN
33
34
Michael addition of acrylic acid to the chloroamine 32 is straightforward and Friedel-Crafts
cyclisation of 35 gives only 31, presumably because the position next to the chlorine atom is slightly
disfavoured both sterically and electronically. Chlorination and oxidation are conveniently carried
out in the same step and the two halves (26 and 28) of this convergent synthesis are combined to
give amopyroquine 25.
Cl
CO2H
CO2H
PPA
32
POCl3
26
I2
heat in
aqueous ethanol
31
N
H
Cl
Cl
25
N
28; X = Cl
35
In the last step we return to the original question of chemoselectivity: Only the primary amine
in 26 reacts because it is more nucleophilic than OH and because the more nucleophilic tertiary
amine adds reversibly – it cannot lose a hydrogen atom as it does not have one. Only the 4-chlorine
atom in the pyridine 28 reacts, presumably because addition to the other position would require
the disruption of both aromatic rings. Though this compound has been succeeded by better antimalarials, its synthesis illustrates the all-important principle that predictions of chemoselectivity
must be based on sound mechanistic understanding. If doubt remains it is worth trying a model
reaction on simpler compounds or, of course, an alternative strategy.
When Protection is not Needed
Dianions: wasting reagent to achieve selectivity
In that synthesis we moderated an over-reactive amino group by protection. Sometimes, protection
is not necessary if we are prepared to squander some of our reagents. A trivial example is the
addition of methyl Grignard to the ketoacid 36. We have already seen how acidic protons destroy
Grignard reagents, but if we are prepared to waste one molecule of the Grignard, we get automatic
protection of the carboxylic acid by deprotonation. Nucleophilic MeMgI will not add to the anion
of a carboxylic acid but adds cleanly to the ketone to give, after workup, the alcohol 37.
O
O
MeMgI
CO2H
R
MeMgI
CO2 Mg
R
2
36; Mg salt
36
OH
O
R
CO2 Mg
Me
37; Mg salt
2
H
H2O
R
CO2H
Me
37
14
2 Chemoselectivity
At first sight, the synthesis of Z-38 by the Wittig reaction seems too risky. The phosphonium
salt 39 has a more acidic proton (CO2H) than the one we want to remove to make the ylid, and the
aldehyde 40 not only also has an acidic proton (OH), but it prefers to remain as the cyclic hemiacetal 41 so that there is no carbonyl group at all.
Wittig
Z-38 HO
Wittig
CO2H reagent
39
Ph3P
CO2H
+
hemi-acetal
41
HO
OH
O
CHO
hydroxy-aldehyde
40
However, simply using a large excess of base makes the reaction work without any protection.
The phosphonium salt 39 does indeed lose its first proton from the CO2H group 42, but the second
molecule of base forms the ylid 43 as the two anions are far enough apart not to influence each
other.6 Base also catalyses the equilibrium between the anions 44 and 45 so that 43 and 45 can
react to give the target molecule. The transition state for this reaction has three partial negative
charges, but they are well apart from each other and there is obviously not too much electrostatic
repulsion as the reaction goes well. This case is opposite to the previous ones: careful mechanistic
analysis shows that expected chemoselectivity problems do not materialise.
base
base
Ph3P
39
Ph3P
CO2
42
41
CO2
43
base
O
44
O
O
Z-38
CHO
45
Chemoselectivity by Reagent: The Pinacol Rearrangement
So far we have discussed chemoselectivity between different functional groups. The situation gets
more complicated if the functional groups are similar, or even the same. The pinacol rearrangement
is a useful route to carbonyl compounds from diols, the classical example being the rearrangement
of 46 in acid solution to give the t-alkyl ketone 48. There are no chemoselectivity problems here:
the two hydroxyl groups in 46 are the same so it does not matter which gets protonated and, in the
rearrangement step 47, all four potential migrating groups are methyl.
H
HO
OH
46
'pinacol'
Me
migration
Me
HO
OH2
HO
47
O
48
'pinacolone'
Selectivity between secondary and tertiary alcohols by reagent
Unsymmetrical diols provide a serious problem of chemoselectivity with an ingenious solution.7
Treatment of the diol 49 with acid leads to loss of OH from what would be the more stable
t-alkyl cation and hence, by hydrogen shift, to the ketone 51.
2 Chemoselectivity by Reagent: The Pinacol Rearrangement
H
H
migration
H
H
R
R
15
H
R
H
R
HO
OH
HO
49
HO
OH2
O
51
50
The alternative, more interesting rearrangement to give 53 can be initiated by tosylation of
the diol 49 in weak base. It is impossible to tosylate tertiary alcohols under these conditions, as
both the t-alcohol and TsCl are large, so only the secondary alcohol becomes sulfonylated and so
leaves, and the rearranged ketone 53 is the only product.
H
H
TsCl
R
HO
OH
Me
migration
Me
pyridine
TsO
49
Me
H
R
H
R
OH
OH
R
52
O
53
Corey’s longifolene synthesis
The question of which group migrates in a pinacol rearrangement is also a question of chemoselectivity, and usually groups that can participate because they have lone pair or π-electrons migrate
best. In Corey’s longifolene synthesis,8 the 6/7 fused enone 54 was an important intermediate. Synthesis from the readily available Robinson annelation product 57 is very attractive, but this demands
a ring expansion step such as the pinacol rearrangement of 55 of unknown selectivity. 1,2-Diols such
as 55 normally come from the hydroxylation of an alkene, in this case the diene 56 which might be
made by a Wittig reaction on the dione 57. Every step in this sequence raises a question of chemoselectivity. Which of the two ketones in 57 is more reactive? Which of the two double bonds in 56 is
more easily hydroxylated? Which side of the ring migrates in the pinacol rearrangement on 55?
pinacol
O
O
O
O
?
HO
O
FGI
Wittig
?
?
HO
O
54
55
56
57
One of the ketones in 57 is conjugated, and one is not. The unconjugated one is less stable and
we can therefore use thermodynamic control if we protect as an acetal, a reversible process. The
unconjugated ketone would also be more kinetically reactive towards the Wittig reagent. Of the
two double bonds in 59, the one outside the ring is more reactive towards electrophilic reagents,
again for both kinetic and thermodynamic reasons. The tosylation route ensures that the right
OH group leaves in the pinacol rearrangement and because the remaining π-bond migrates better
than the simple alkyl group when 60 rearranges with a weak Lewis acid, all is well. The synthesis
therefore follows the route below, with all questions of chemoselectivity neatly solved. The acetal
protecting group was also useful later in the synthesis.
16
2 Chemoselectivity
O
O
O
O
Ph3P
HO
57
OsO4
OH
TsOH
pyridine
O
E and Z-59
58
O
O
O
O
O
HO
HO
O
LiClO4
TsCl
pyridine
TsO
HO
O
CaCO3
THF
60
61
Chemoselectivity in Enol and Enolate Formation
General discussion of enols and enolates
We have concentrated so far on two functional groups within the same molecule. The chemoselectivity problem is just as important when we want two molecules to react together in a certain
way, but, because both molecules have similar functional groups, the reaction can occur the other
way round, or one of the molecules may react with itself and ignore the other. This problem is
particularly acute in reactions involving enolisation. The alkylation or acylation of enols or enolates and the reaction of one carbonyl compound with another, the aldol reaction, are classical
and important examples summarised in the general scheme below. We shall concentrate in this
chapter on the chemoselectivity of these processes, that is we shall look at the enolisation of esters,
aldehydes, and the like.
OH
O
R1
O
R1
carbonyl compound
enol
O
R1
enolate
O
aldol
O
O
OH
R2
R1
R1
acylation O
aldol
H
R2
R1
RBr
O
aldol
R2
H
O
alkylation
R2
H
OH
O
R2
R2
X
R2
OH
O
R2
R1
O
R2
R1
Reaction of an ester 62 with its own alkoxide ion produces a small amount of enolate 63 that
reacts with unenolised ester to give the ketoester 64. This reaction, though useful in its own right,
precludes the direct alkylation of esters under these conditions.
H
H
H
EtO
R
OEt
R
OEt
R
O
O
62; ester
63; ester enolate
R
CO2Et
R
CO2Et
O
64; β-keto-ester
2 Chemoselectivity in Enol and Enolate Formation
17
Formation of specific enol equivalents
What is needed for the alkylation is rapid conversion of the ester into a reasonably stable enolate, so rapid in fact that there is no unenolised ester left. In other words the rate of proton
removal must be faster than the rate of combination of enolate and ester. These conditions are
met when lithium enolates are made from esters with lithium amide bases at low temperature,
often ⫺78 ⬚C. Hindered bases must be used as otherwise nucleophilic displacement will occur
at the ester carbonyl group to give an amide. Popular bases are LDA (Lithium Di-isopropyl
Amide, 66), lithium hexamethyldisilazide 67, and lithium tetramethylpiperidide 68, the most
hindered of all. These bases are conveniently prepared from the amine, e.g. 65 for LDA, and
BuLi in dry THF solution.
Me3Si
BuLi
N
H
N
THF
65
Di-isopropylamine
Li
SiMe3
N
N
Li
Li
66; LDA
67; LHMDS
Li Di-isopropyl Amide Li HexaMethylDiSilyazide
68; LTMP
Li TetraMethylPiperidide
Treatment of a simple ester 62 with one of these bases at ⫺78 ⬚C leads to a stable lithium
enolate 70 by initial coordination of lithium to the carbonyl group 69 and proton removal via a
six-membered cyclic transition state 69a.
H
H
O
R
H
LDA
–78 °C
THF
OEt
R2N
Li
H
H
O
R
R2N
Li
H
O
R
OEt
62; ester
H
OLi
R
OEt
OEt
69a
69
+ R2NH
70
Lithium enolates, enamines and silyl enol ethers
Direct alkylation of lithium enolates of esters9 62 and lactones 73, via the lithium enolates 71 and
74, with alkyl halides is usually successful.
H
1. LDA
R1
O
R1
CO2Et
Br
2. R 2Br
OEt
R2
62
R2
Li
R1
CO2Et
71
O
O
1. LDA
72
OLi
O
O
O
2. RBr
R
73
74
75
18
2 Chemoselectivity
More impressive and more important is the performance of these lithium enolates in aldol
reactions. Ester enolates are awkward things to use in reactions with enolisable aldehydes and
ketones because of the very efficient self-condensation of the aldehydes and ketones. The traditional
solutions involve such devices as Knoevenagel-style reactions with malonates.11 Lithium enolates
of esters, e.g. 76, react cleanly with enolisable aldehydes and ketones to give high yields of aldols,12
e.g. 79 in a single step also involving a six-membered cyclic transition state 77.
O
MeCO2Et
LDA
Li
O
OLi
+
THF
–78 °C
O
OEt
OEt
76
77
OLi
OH
CO2Et
H
CO2Et
H2O
79, 79-90%
78
They even react cleanly with formaldehyde, thus solving the problem that the Mannich reaction
is not applicable to esters. The synthesis of the exo-methylene lactone 80 can be accomplished
this way. Enone disconnection13 reveals formaldehyde as the electrophilic component in a crossed
aldol reaction, realised with a lithium enolate 82.14 The mono-adduct 83 of formaldehyde and the
lactone 81 can be isolated and the cautious dehydration step is to avoid migration of the double
bond into the ring.
H
H
α,β-unsaturated
carbonyl
O
O
O + CH2=O
O
H
H
80
81
H
81
H
O
LDA
O
CH2O
1. MsCl, Et 3N
80
O
OLi
2. pyridine
H
H
82
OH
83
The same technique can even be applied to carboxylic acids themselves 84 providing two
molecules of base are used. The first removes the acid proton to give the lithium salt 85 and the
second forms the lithium enolate 86.
O
R
OH
84
LDA
O
R
OLi
85
LDA
OLi
R
OLi
2 x LDA
–78 ˚C, THF
84
86
These lithium derivatives are also well behaved in alkylations and aldol reactions. Krapcho’s
synthesis15 of the sesquiterpene α-curcumene 92 starts with the chemoselective condensation of
2 Chemoselectivity in Enol and Enolate Formation
19
the dilithium derivative of the acid 87 with the enolisable aldehyde 89. The aldol product 90 is
converted into the β-lactone 91 and hence by heating and loss of CO2 into α-curcumene 92.
2 x LDA
OLi
CO2H
OLi
88
87
87
CHO
1. 2 x LDA
89
PhSO2Cl
90
73% yield
2. 89
CO2H
pyridine
OH
90, 73% yield
140 °C
O
O
91, 77% yield
92; α-curcumene, 90% yield
You might be forgiven for thinking that lithium enolates solve all problems of enolate chemoselectivity at a stroke and wonder why they are not always used. They are very widely used, but
they require strictly anhydrous conditions at low temperatures (usually ⫺78 ⬚C, the temperature
of a dry ice/acetone bath) and no-one in their right mind would use these conditions if mixing
the reagents in ethanol at room temperature with a catalytic amount of NaOH did nearly as well.
These are the conditions of many simple aldol reactions and are preferred where practical, particularly in industrial practice. The intermediate 93 was needed in a synthesis of geiparvirin. The
best aldol disconnection in the middle of the molecule gives a ketone 94, that must be enolised in
the only possible position, and then react with an unenolisable and more electrophilic aldehyde 95.
No selectivity problems arise and an equilibrating aldol reaction between 94 and 95 catalysed by
NaOEt in EtOH gives 93 in 89% yield.16
O
O
H
aldol
Ph
HO
93
+ O
HO
94
Ph
95
Enamines
Lithium enolates do not even solve all problems of chemoselectivity: most notoriously, they fail
when the specific enolates of aldehydes are needed. The problem is that aldehydes self-condense
so readily that the rate of the aldol reaction can be comparable with the rate of enolate formation
by proton removal. Fortunately there are good alternatives. Earlier in this chapter we showed
examples of what can go wrong with enamines. Now we can set the record straight by extolling the
virtues of the enamines 96 of aldehydes.17 They are easily made without excessive aldol reaction as
they are much less reactive than lithium enolates, they take part well in reactions such as Michael
additions, a standard route to 1,5-dicarbonyl compounds, e.g. 97.18
20
2 Chemoselectivity
CHO
NR2
NR2
R2NH
OMe
96
OMe
R
R
R
O
O
NR2
OMe
R
CHO
H
R
H2O
O
CO2Me
97
An impressive example19 is the Robinson annelation of the unsaturated aldehyde 98 where
neither aldol reaction nor double bond migration in the enamine 99 interferes. The 1,5-dicarbonyl
compound 100 cyclises spontaneously to the enone 101.
O
O
O
CHO
N
+
Ph
OHC
N
H
Ph
Ph
98
Ph
100
99
101
Silyl enol ethers
For all their usefulness, enamines have now largely been superseded by silyl enol ethers. These
(102-104) can be made directly with Me3SiCl from the lithium enolates of esters or acids or
from aldehydes under milder conditions with a tertiary amine. The silicon atom is an excellent
electrophile with a strong preference for more electronegative partners and it combines with the
oxygen atom of an enolate so rapidly that no self condensation occurs even with aldehydes.
Silyl Enol Ethers of Esters
O
LDA
R
R
OEt
OLi
Me3SiCl
OEt
102
OEt
Silyl Enol Ethers of Acids
O
LDA
R
R
OH
OLi
Me3SiCl
OSiMe3
103
OLi
O
H
Et3N
OH
R
H
OSiMe3
R
Silyl Enol Ethers of Aldehydes
R
OSiMe3
R
Me3SiCl
OSiMe3
R
H
104
The silyl enol ethers 102 and 104 are shown as single geometrical isomers for convenience:
in fact they are normally formed as mixtures, though this does not usually affect their reactions.
They are thermodynamically stable compounds but are easily hydrolysed with water or methanol
and are usually prepared when they are needed. They are much less reactive than lithium enolates,
or even enamines, and their reactions with electrophiles are best catalysed by Lewis acids, often
2 Chemoselectivity in Enol and Enolate Formation
21
TiCl4. The aldehydes 105 and 108, the one branched and the other not, are simply converted
into their silyl enol ethers 106 and 109 and combined with two different enolisable aldehydes to
give high yields of aldol products 107 and 110 without any self condensation of any of the four
aldehydes or any cross-condensation the wrong way round.20
Ph
Me3SiCl
CHO
CHO
OSiMe3
Et3N
Ph
TiCl4
CHO
OH
107; 86% yield
106
105
CHO
Me3SiCl
Ph
CHO
Ph
OSiMe3
n-PrCHO
Ph
TiCl4
Et3N
OH
108
110; 78% yield
109
The silyl enol ethers of esters, e.g. 111 and lactones, e.g. 114 similarly take part in efficient aldol
reactions with enolisable aldehydes and ketones with Lewis acid catalysis, again with complete
regioselectivity. Example 113 is particularly impressive as the very enolisable ketone gives a high
yield of an aldol product with two adjacent quaternary centres.21
Ph
OSiMe3
1. LDA
CO2Et
2. Me 3SiCl
OEt
111
O
Ph
TiCl4
113, 94% yield
112
O
Me3SiO
O
1. LDA
O
2. Me 3SiCl
114
CO2Et
OH
+
O
TiCl4
O
O
OH
115
116
Synthesis of the ant alarm pheromone mannicone
The synthesis of the ant alarm pheromone mannicone 117 is a good example. Enone disconnection
reveals that we need a crossed aldol condensation between the symmetrical ketone 118, acting as
the enol component, and the enolisable aldehyde 119.
Mannicone: Analysis
O
O
H
aldol
+ O
enone
117
118
119
The ketone gives a mixture of geometrical isomers of the silyl enol ether 120 which condense
with the aldehyde 119 to give the aldol 121 as a mixture of diastereoisomers which dehydrates to
mannicone 117 in acid.22 It is particularly impressive that the optically active aldehyde 119 has its
22
2 Chemoselectivity
stereogenic centre at the enol position and yet optically active mannicone is formed by this route
without racemisation.
Mannicone: Synthesis
OSiMe3
Me3SiCl
118
O
OH
TsOH
119
Et3N
TiCl4
120
TM 117
121
We shall discuss further aspects of the aldol reaction in the next two chapters where we shall
see how to control the enolisation of unsymmetrical ketones, and how to control the stereochemistry of aldol products such as 121. We shall return to a more comprehensive survey of specific
enol equivalents in chapter 10. In this chapter we are concerned to establish that chemoselective
enolisation of esters, acids, aldehydes, and symmetrical ketones can be accomplished with lithium
enolates, enamines, or silyl enol ethers, and we shall be using all these intermediates extensively
in the rest of the book.
Examples of Chemoselective Reactions in Synthesis
Synthesis of lipstatin
The synthesis of lipstatin 122 is too complex to discuss here in detail but an early stage in one
synthesis uses a clever piece of chemoselectivity.23 Kocienski planned to make the β-lactone by
a cycloaddition with the ketene 124 and to add the amino acid side chain 123 by a Mitsunobu
reaction involving inversion. They therefore needed Z,Z-125 to join these pieces together. This
was to be made in turn by a Wittig reaction from 126. The problem now is that 126 is symmetrical
and cannot carry stereochemistry and that aldehydes are needed at both ends.
•
123
O
O
O
O
NHCHO
NHCHO
OH
O
124
CO2H
Me3Si
CHO
H
H
122; lipstatin
OH
OHC
Z,Z-125
CHO
126
The way they solved the problem was this. (S)-(⫺)-Malic acid is available cheaply. Its dimethyl
ester 127 could be chemoselectively reduced by borane to give 128. Normally borane does not
reduce esters and clearly the borane first reacts with the OH group and then delivers hydride to the
nearer carbonyl group. The primary alcohol was chemoselectively tosylated 129 and the remaining (secondary) OH protected with a silyl group 130 (TBDMS stands for t-butyldimethylsilyl and
is sometimes abbreviated to TBS). Now the remaining ester can be reduced to an aldehyde 131
and protected 132. Displacement of tosylate by cyanide puts in the extra carbon atom 133 and
reduction gives 134, that is the dialdehyde 126 in which one of the two aldehydes is protected.
This compound was used in the successful synthesis of lipstatin.
2 Examples of Chemoselective Reactions in Synthesis
OH
MeO2C
OH
BH3•SMe2
NaBH4
CO2Me
OH
TsCl
HO
CO2Me
127; (S)-(–)-malic acid
dimethyl ester
128; 83% yield
OTBDMS
TsO
imidazole
OTBDMS
TsO
CH2Cl2
130; 92% yield
OTBDMS
DMSO
(i-PrO)3CH
CHO
TsOH
131; 78% yield
NaCN
CH(Oi-Pr)2
TsO
CO2Me
129; 75% yield
DIBAL
CO2Me
OTBDMS
TsO
pyridine
CH2Cl2
THF
t-BuMe2SiCl
23
NC
CH(Oi-Pr)2
132; 92% yield
OTBDMS
DIBAL
OHC
CH2Cl2
133; 83% yield
CH(Oi-Pr)2
134; 81% yield
The synthesis of rubrynolide
The synthesis of the natural product rubrynolide from th