INFORMATION TO USERS
This m anuscript has been reproduced from the microfilm m aster. UM I films the text directly from the original or copy submitted. Thus, some thesis and dissertatio n copies are in typewriter face, while others may be from any type o f com puter printer.
T h e q u a lity o f th is re p ro d u c tio n is d e p en d e n t u p o n th e q u a lity o f th e copy s u b m itte d . B roken or indistinct print, colored or p o o r quality illustrations and photographs, print bleedthrough, substandard margins, and im proper alignm ent can adversely affect reproduction.
In the unlikely event that the author did not send U M I a com plete manuscript and th e re are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.
Oversize m aterials (e.g., maps, drawings, charts) are reproduced by sectioning th e original, beginning at the upper left-hand c o m er and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the b ack o f th e book.
Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6” x 9” black and white photographic p rin ts are available for any photographs o r illustrations appearing in th is co p y fo r an additional charge. C ontact UM I directly to order.
UMI
A Bell & Howell Information Company
300 North Zeeb Road, Ann Arbor MI 48106-1346 USA 313/76M 700 800/521-0600
CO M PON EN TS O F IO N TRANSPORTERS by
Lynn M ichele Cam eron B.Sc., Saint M ary’s University, 1986
M .Sc., M cM aster University, 1990
A Dissertation Subm itted in Partial Fulfillment o f the Requirem ents for the D egree o f
D O CTO R OF PH ILO SO PH Y in the D epartm ent o f Chemistry
W e accept this thesis as conforming to the required standard
esTSupepvisor
Dr. T.M. Fyles, S u p e ^ s o r (D epartm ent o f Chemistry)
,'O epartm ental IMe
Dr. A. ^ ley,'D epartm ental M em ber (Departm ent o f Chem istry)
Dr. P C. Wan, Departmental M ember (D epartm ent o f Chemistry)
(Departm ent o f Biology)
Dr. L. W eileryExtemal Exam iner (University o f British Columbia) LYNN M IC H E LE CAM ERON
1494-U niversity o f Victoria
All rights reserved. Dissertation m ay not be reproduced in whole o r in part, by photocopying o r other means, without the permission o f the author.
Supervisor: Dr. T.M . Fyles
ABSTRACT
The thesis describes the development o f a new set o f macrocycles that could be used as part o f an existing modular set o f com ponents, that w hen assembled, probe the structure-activity relationship o f ion transporters. A property directed synthesis
approach w as adopted in which the properties, rath er than th e specific chemical makeup o f the final target, guide the synthetic path. This allows fo r incorporation o f the favorable properties and avoidance o f the troublesom e aspects o f the current set o f macrocycles into the synthetic design o f the new m odular components. As
com ponents in a m odular set, the macrocycles are required in sufficient yield, so the property directed synthesis must provide an econom ical and efficient route to the targets. M olecular mechanics was used as a to o l to estim ate the length and rigidity o f the macrocyclic systems. The design w as also dictated by the need to incorporate high yielding m ethods for macrocyclization into the synthesis. All targets are formally meta cyclophanes involving overall axial symmetry to limit regioisomerism.
The syntheses o f 23, 25, 27, and 28 w ere based on the macrocyclization o f bis- a-h alo amides with bis-nucleophiles. T arget 2 2 incorporated the bis-chloroamide o f /neto-phenylene diamine and a diol 23 derived fi'om 1,3 bisbromomethyl benzene and butane diol. Conditions sufficient to deprotonate 23 apparently resulted in
concom itant deprotonation o f the amides leading to com plex mixtures o f products. Based on these observations the secondary am ides o f 23 w ere replaced with the hom ologous N -M e tertiary amides to give targ et 25. As anticipated, macrocyclization
o f 23 and th e bis chloram ide o f N ,N ’ dimethyl m etaphenyiene diam ine gave 25a in a yield o f 7% .
T he design requires additional functional groups at the 5 and 5’ positions o f the cyclophane. The chem istry leading to com binations o f nitro and /-B oc protected amino g roups in these positions w as explored. Additional side reactions o f the alkoxide nucleophile w ith nitro substituted arom atics precluded the desired
m acrocyclization reactions, and in all cases complex product m ixtures w ere obtained. Additionally, the diol com ponent 23 w as replaced with a diol 26 o f similar length in w hich the ether linkages w ere replaced by secondary amides. This gave a further targ et 27. In this series as well the side reaction o f alkoxide w ith the nitro arom atic resulted in destruction o f starting materials w ithout production o f product macrocycle. A corresponding target incorporating a dithiol in place o f the diol 23 was explored in a preliminary fashion, but the physical properties o f the dithiol complicated the purification o f the precursor.
T he synthesis o f a series o f tetraester m acrocycles derived from isophthalic acid and diols w as explored. D irect condensation o f 5-nitro isophthaloyl chloride and
1,8-octanediol gave a series o f symetrical 2+2, 3+3, and 4+4 tetraesters in high
conversion. A step-w ise synthesis from the m ono-tetrahydropyranyl derivative o f 1.8- octane diol gave the 2+2 m acrocycle 67 substituted w ith nitro and /-B o c protected amino in th e 5 and 5’ positions in an overall yield o f 16%.
The efficiency o f the synthesis o f 67 is analysed using plan graphs as initially described by Hendrickson for natural products synthesis. The synthesis plan is com pared with the actual perform ance in term s o f reagent consumption, total weight manipulated, and the time for the synthesis. These results are com pared with the synthesis o f a
comparable target in the existing series o f macrocycles. A lthough tetraester 67 is somewhat less efficiently prepared than earlier examples, it is better functionalized to lead to candidate ion transporters.
Examiners:
Dr. T.M. Fyles, Supervisor (D epartm ent o f Chemistry)
Dr. A. McAuley, D epartm ental M ember (D epartm ent o f Chemistry)
Dr. P C Wan, Departmental M ember (Department o f Chemistry)
Dr. J.N. Owens, O utside M em ber (Departm ent o f Biology)
TITLE PAGE i
ABSTRACT ü
TABLE OF CONTEN TS v
LIST OF TABLES viii
LIST OF FIGURES ix
LIST OF SCHEMES xii
LIST OF ABBREVIATIONS xiv
ACKNOW LEDGEM ENTS xv
CHAPTER 1 IN T R O D U C TIO N 1
1.1 Ion T ransport 2
1.2 Examples o f N atural Channels 3
1.2.1 Gramicidin A 3
1.2.2 A m photericin B 3
1.3 Artificial Ion C hannels 5
1.3.1 Design C riteria 5
1.3.2 Channels F o rm ed From Dimers 7
1.3.3 A ggregate P o res 9
1.3.4 Single M olecule Channels 15
CHAPTER 2 PRO JECT P R O P O S A L 25
2.2 Strengths and W eaknesses o f Previous M acrocodes 25
2.3 Property D irected Synthesis 29
2.3.1 Introduction to Synthetic Efficiency 3 1 2.3. I . l N ature o f th e Synthetic Sequence 32
2.3.1.2 M aterials 34 2.3.1.3 R eagents 35 2.3.1.4 Tim e 35 2.4 M acrocyclization Techniques 37 2.5 M olecular M odelling 41 2.6 Target Evolution 43 2.6.1 General Strategy 43
2.6.2 D esign o f M acrocycles Based on a-IIeteroam id es 44 2.4.4 D esign o f T etraester M acrocycles 56 C H A PTER 3 SY N TH ESIS A ND D ISC U SSIO N 58 3.1 Tow ards th e Synthesis o f the Diamide M acroo'cles 22 and 25 58 3.2 Synthesis o f the T etraester M acrocycles 95 C H A PTER 4 C O N C L U SIO N S AND FU T U R E W ORK 107
4.1 Summary 107
4.2 Synthetic Efficiency 110
4.2.1 C om parison o f the M onoprotected M acrocycle 6 8
with th e M onosubstituted 8% 117
with the Pore F o rm er Derived from 82 120 4.2.3 Conclusions on Synthetic EfiBciency 126
4.3 Future W ork 128
4.4 Conclusions 129
C H A P T E R 5 EXPERIM ENTAL 131
5.1 Apparatus 131
5.2 Procedures 132
5.3 M olecular Mechanics M ethodology 172
LIST OF TABLES
Table 1 ; Tabulation o f Predicted Yields, Inverse Yields and Sum o f
Inverse Yields 34
Table 2; Com parison o f the " C nmr data for Com pounds 54 and 5 4 a 91 Table 3 ; Com parison o f the nmr data for Com pounds 54 and 54a 92 Table 4: '^C nm r Chemical Shifts for Tetraester M acrocycles 29, 67
and 61 106
Table 5: A dvantages and Disadvantages for Isolated T a get M acrocycles 109 Table 6; W eight Summaries for M acrocycle 68 Assuming 80% Yield
at Each Step 115
Table 7: Actual Yield for M acrocycle 6 8 116 Table 8: W eight Summaries for M acrocycle 68 Using Real Yields 116 Table 9; Yields, Inverse Yields and Sum o f Inverse Yields for
M onosubstituted 82, 71 119
Table 10: W eight Summaries for M onosubstituted 82, 71 Using Real Yields 119
Table 11 : Yields fo r Pore Form er 73 121
Table 12: W eight Summaries for Pore Form er 73 122 Table 13: Yield Values for Pore Form er Derived from 82 124 Table 14: W eight Summaries for Pore Form er Derived from 82 125
L IS T O F F IG U R E S
Figure I: Mechanisms for Ion T ransport 2
Figure 2: a)Gramicidin, b)A m photericin 4
Figure 3: Tabushi’s Cyclodextrin Channel 7
Figure 4: K obuke’s Calixarene B ased Channel 8
Figure 5; Fyles Suite o f Pore Form ers 10
Figure 6; K obuke’s Ion Pair A ggregates 12
Figure 7: Ghadiri’s Cyclic Peptide 13
Figure 8: R egen’s Steroid B ased Transporter 14
Figure 9: N olte’s Isocyanide Polym er 15
Figure 10; Fyles Group o f Single Molecule Channels 17 Figure II : Lehn’s Channels, a)18-C row n-6 Based Structure, b)Cyclodextrin
Based Structure 19
Figure 12: G okel’s Family o f Single Molecule Transporters 22
Figure 13: V oyer’s Stacked C row ns 23
Figure 14: Previous Set o f Wall Units 26 Figure 15: Stereo- and Regio- Isom ers From Previous Sel o f Wall Units 28
Figure 16: Example o f a Plan G raph 32
Figure 17: Target 1, M acrocycle 22 44
Figure 18: Target 2, M acrocycle 25 47
M acrocycle 2 2 a 49 Figure 20: A R epresentative Snapshot o f a Low Energy C onform ation o f
M acrocycle 2 5 a 50
Figure 21: 5-N itroisophthalic acid 51
Figure 22: M odified D iol 26 52
Figure 23: T arget 3, M acrocycle 27 52
Figure 24: A R epresentative Snapshot o f a Low Energy C onform ation o f
M acrocycle 27 53
Figure 25: T arget 4, M acro cy cle 27 54
Figure 26: A R epresentative Snapshot o f a Low Energy C onform ation o f
M acrocycle 28 55
Figure 27: T arget 5, M acro cy cle 29 56
Figure 28: A R epresentative Snapshot o f a Low Energy C onform ation o f
M acrocycle 29 57
Figure 29: N M R S p ectru m o f Bis Chloroamide 24a 60 Figure 30: " C N M R S p ectru m o f Compound 39 63 Figure 31: " C N M R S p ectru m o f Diol 23 64 Figure 32: C onform ations o f T ertiary Amides 66
Figure 33: ‘^C N M R S p ectru m o f Bis Chloroamide 24b 69 Figure 34: ‘H N M R S p ectru m o f Bis Chloroamide 24b 70 Figure 35: *^C N M R S p ectru m a t R oom Temperature o f M acrocycle 2 5a
Figure 36; N M R Spectrum a t R o o m Tem perature o f M acrocycle 2 5 a
in D M SO 72
Figure 37: "C N M R Spectrum a t 80“C o f Macrocycle 25a in D M SO 73 Figure 38: N M R Spectrum o f B is lodoam ide 43c 78
Figure 39: N M R Spectrum o f 4 4 a 81
Figure 40: N M R Spectrum o f C bz Protected Compound 43b 86
Figure 41 : N M R Spectra o f 5 4 and 5 4a 91 Figure 42: N M R Spectra o f 5 4 and 5 4 a 92
Figure 43 : M acrocycle 27 93
Figure 44: N M R Spectrum o f D iol 60 98
Figure 45: M ixture o f M acrocyclic P ro d u cts 100 Figure 46: Mass Spectrum o f M acrocyclic M ixture 61 101 Figure 47: '^C N M R Spectra o f M acrocycles 29, 67 and mixture 61 105 Figure 48: Plan G raph o f M acro cy cle 6 8 Assuming 80% Yield at E ach Step 115 Figure 49: Plan G raph o f M acro cy cle 6 8 Using Real Yields 116 Figure 50: Plan G raph o f M o n o su b stitu ted 8% 118 Figure 51 : Plan G raph o f P o re F o rm e r 73 Using Real Yields 121 Figure 52: Plan G raph o f P o re F o rm e r 75 124
LIST OF SCHEM ES
Scheme 1: R egen’s Steroid Incorporated into Lipid 14
Scheme 2: Property Directed Synthesis 29
Scheme 3; Retrosynthetic Analysis o f a M acrocycle 37 Scheme 4: Com peting Pathways (Cyclization or Polymerization) 38 Scheme 5; General Design o f N ew Wall Units 43 Scheme 6: R etrosynthetic Analysis o f Targets 22 and 25 45 Scheme 7; Preparation o f Bis Chloroamides 46 Scheme 8; R etrosynthetic Analysis o f Target 1 58 Scheme 9; Synthesis o f Bis Chloroamide 24a 59 Scheme 10: Initial A ttem pts at the Synthesis o f Diol 23 62
Scheme 11: Synthesis o f Diol 23 62
Scheme 12: Synthesis o f Bis Chloroamide 24b 66
Scheme 13: Synthesis o f M acrocycle 25a 74 Scheme 14: Synthetic Pathway for M acrocycle 25 75 Scheme 15: Synthesis o f Bis Chloroamide 43a and Bis lodoam ide 43b 76
Scheme 16: Alternate R oute to 47 79
Scheme 17: Synthesis o f Diol 44a 80
Scheme 18: Synthesis o f /-Boc Protected Amine 4 4 b 82
Scheme 19: Synthesis o f 43b 84
Scheme 21 ; Preparation o f 54 88
Scheme 22; Pathway for the Form ation o f 5 4 a 89
Scheme 23 : Synthesis o f 58 95
Scheme 24: Retrosynthetic Analysis o f T etraester Macrocycles 96
Scheme 25 : Synthesis o f 60 97
Scheme 26: Synthesis o f Dinitro M acrocycle 61 99 Scheme 27: Stepwise Synthesis o f th e Asym m etric Macrocycle 67 104 Scheme 28: Synthesis o f M acrocycle 6 8 114 Scheme 29: Synthesis o f M onosubstituted 8% 118 Scheme 30: Synthesis o f Pore Form er 73 120 Scheme 31 : Synthesis o f Pore Form er 75 123
L IS T O F A B B R E V IA T IO N S A r arom atic
D M A N ,N '-dim ethylacetam ide D M F N ,N '-dim ethyIfoniiam ide D M SO dimethylsulfoxide
IR infrared M e methyl mp melting point ms mass spectrum
nm r nuclear magnetic resonance /-B oc /er/-butyloxy carbonyl TH F tetrahydrofuran
Thp tetrahydropyranyl
ACKNOWLEDGEMENTS
I w ould like to thank my supervisor Dr. T. Fyles for his patience, guidance and encouragem ent throughout this work. Thanks are extended to Scott B uckler for preparing many o f the starting materials and to Danny Lau for help w ith the sulfiir chemistry. I w ish to thank th e chemistry department staff and faculty, in particular M rs. Christine G reenw ood and Dr. David McGillivray. I thank K .E.Laidig and fellow graduate students especially B.Cam eron, L.Clouston, K.Coulter, R .H ooper, D .Loock, P.M ontoya-Pelaez, X .Zhou and B.Zeng fo r insightful discussions thro u g h o u t the course o f this w ork. Financial assistance from the University o f V ictoria w as much appreciated.
INTRODUCTION TO ION TRANSPORT
1.1 Ion Transport
Biological membranes are necessary to separate cells and cell compartments, but in o rd er to maintain the com plex biological processes, the cells must communicate with each other. C om m unication between the cell com ponents occurs through the exchange o f ions o r m olecules and these substances must pass through the membrane barrier. In general, m embranes are impermeable to ions because o f the high energy barrier that m ust be overcom e to traverse the bilayer. Thus transporters act as catalysts by low ering this energy barrier and discriminantly allow ions to pass. With few exceptions, m ost natural channels are large proteins o r protein aggregates, the structures o f which are still being investigated Little is know n on the molecular level o f the structure and control o f ion channels'* so through the study o f artificial model systems that mimic the function o f the more complex system, insight into the biological process may be obtained. O f those few exceptions m entioned above, smaller natural products such as gramicidin A or amphotericin B have been widely studied and provide the basis fo r much o f the structural design o f the mimics’ *'^
F o r as many ion transporters exist, there are probably as many mechanisms for transport to occur. The tw o extrem es are shown in Figure 1. These tw o basic
mechanisms for tran sp o rt are the channel mechanism or carrier mechanism. The structural requirem ents for a transporter vary depending on the mechanism For instance, a channel m ust span the thickness o f the bilayer. A carrier may be envisioned
side. It is often difficult to distinguish betw een the tw o modes o f action but the transport kinetics may provide som e insight. Ions may flow through a channel
unhindered at a rate o f lO’ s '\ as in th e case for gramicidin, whereas carriers typically move less than 10^ ions s’* because th e ion-carrier complex must diffuse th ro u g h th e lipid. Carriers have been widely studied and there are several reviews on the subject*.
This thesis deals with the design and synthesis o f artificial transporters o f alkali metal cations via channels across lipid membranes A detailed discussion o f ion
channel transporters was published by Fyles and van Straaten-Nijenhuis’ . A short, but concise review has recently been published by G okel'”. This chapter introduces the structure and mechanism o f som e natural transporters that provide the basis fo r th e design rationale and synthesis o f artificial channels. This is followed by a sh o rt review o f synthetic transporters.
Carrier
Channel
1.2.1 G ra m ic id in A
Gramicidin A is a cation selective transporter. The active form o f the
transporter is a peptide dimer consisting o f 15 alternating d- and l- amino acid residues each, linked through hydrogen b o n d s a t the N terminus. T he polypeptide form s a (3- helix 26Â long with a pore diam eter o f 4 Â (Figure 2 )" . It is believed th at the carbonyl groups line the inside o f the pore creatin g a hydrophilic environm ent suitable for w ater and cations, while the hydrophobic side chains interact favourably w ith the lipophilic membrane.
1.2.2 A m p h o te ric in B
A m photericin B is an example o f a small non-peptidic transporter. It is a macrocycle com posed on one face o f hydrophobic polyene and on the o th er face o f a hydrophilic poly alcohol. It is believed that 12-20 m olecules o f am photericin aggregate to form a pore*^ in such a fashion th at the hydrocarbon face interacts w ith the lipid and the alcohols orient inwards to line the pore. The alcohols thus create a polar opening through w hich the cations may flow . T here is a high degree o f organization involved with am photericin as tw o aggre g a tes m ust dimerize to form a channel (Figure 2).
a)
b) HO -O -OH HO. —OH HO OH1.3.1 Design Criteria
Artificial channels are functional mimics and most often do not resem ble the natural systems in com position, but rather possess the sam e physical qualities that allow them to perform their function. The design criteria for artificial ion channels are:
1) the assembled channel must span the thickness o f the bilayer, 2) there m ust be a polar group at the bilayer interface, 3) there must be a com plem entary balance o f polar and apolar groups, 4) they must incorporate into the membrane and then adopt the correct orientation once inserted and 5) from a practical standpoint, they must be synthetically feasible. The extent to which points 1 -4 are im portant is unknow n, hence the need for further investigation.
The remaining sections o f this chapter will be divided into three sections o f ion channel design. The first section describes a family o f m olecules that are believed to dimerize to form a channel. In general these molecules are designed to span half the thickness o f the bilayer. The second section is dedicated to the group o f molecules which aggregate to form a pore through the bilayer These m olecules are often characterized by the intermolecular interactions which force them to self-assemble. The third section is dedicated to those which are believed to operate as single molecule channels, that is, the molecule forms a tunnel and is designed to span the w idth o f the bilayer. These are typically large synthetic molecules, on the scale o f 4000 daltons. It is difficult to separate the molecules into these groups and the authors referenced
herein often only speculate on th e mechanism. To date only two groups (Fyles’’ Gokel*'*) have undertaken a structure-activity study which is necessary to investigate the sensitivity o f the channels to changes in structure and to gain a better
understanding o f the channel property design criteria.
The synthetic transporters are studied in biomimetic membranes which consist o f suspensions o f lipids in w ater to form vesicles o r planar bilayers that are form ed by painting the lipid across a teflon plate with a small hole*^. Kinetics from vesicles provide information on overall activity, while planar bilayers focus on rates o f ion transport through single active channels'^.
HN.
Figure 3: Tabushi's C yclodextrin channel‘s
T he earliest artificial transporters w ere dim er channels based on a
tetrasubstituted (3-cyclodextrin published by Tabushi'^ and coworkers in 1982, shown in Figure 3. The cyclodextrins are cyclic cone shaped molecules made up o f 6, 7 and 8
a-D -glucose units, referred to as a - , p - and y-cyclodextrin respectively, and the diam eter o f the hole in each is approxim ately 4.5, 7.0 and 8.5 Â respectively. The P- cyclodextrin having a pore diam eter o f 7
A
w ould then allow the passage o f alkali m etal cations, hydrated cations and som e small molecules. The 14 hydroxyls o f the secondary face o f the cyclodextrin serve as th e polar head group and w ould bem ove the Ca^* / ions along the channel.
K obuke’s channels^* are similar in design (Figure 4) but the pore size is governed by a calix-4-arene. T here are 4 alkyl (Ci?) chains which w ould b e long enough to span half the thickness o f th e bilayer. They propose dimer form ation where the phenolic hydroxyls are the polar head groups. The compound had to be applied to
OH
0 H )8
18
R=CH3(CH 2)16 o r C H 3(C H 2)I0
Figure 4: Kobuke's C alixarene Based Channel
both sides o f the bilayer in order to exhibit any activity and rationalization for this was the com pound could not easily difilise through the membrane. The channel w as investigated as a transporter for b oth Na* and K* ions and was shown to b e selective for K*. It is important to note that a sh o rter alkyl chain (C n ) did not support stable ion transport. An explanation for this is th at it is too short to span the bilayer.
W oolley et al.*^ dem onstrated structural control over ion tran sp o rt by
modifying the widely studied gramicidin channel to investigate the regulatory effects o f positively charged substituents and different conformational isomers at th e m outh o f the channel. Ethylene diamine, propylamine o r piperazine were bound at th e C term inus via a carbam ate linkage and it was show n that the positively charged groups as well as the cis-trans isomerization o f the carbam ate controlled the flow o f cations through the channel.
1.3.3 A g g re g ate Pores
Fyles et al. have adopted a m odular synthetic approach to both th e po re form ers’'^ and single molecule channels'^’^*’^ . Both are composed o f po lar head groups , wall units and core linkers as depicted in Figure 5. The sam e wall units and polar head groups can in principle be used as building blocks in the p o re form ers and in the unimolecular channels. A structure-activity relationship was published in 1994^° in which 14 o f the possible 24 bolaamphiphiles w ere assembled using th e m odular approach and investigated for transport activity. Conclusions w ere draw n concerning the structural requirements for efBcient ion transport. Macrocyclic wall units w ere chosen to impart internal rigidity (structural control) to the system. It w as
hypothesized that differences in the lipophilicity o f the wall units, in th e hydrogen bonding capability o f the spacers and in the head group size and charge w ere important. O f the wall units investigated, 1, 2 and 3, vary in their degree o f
HO_
Polar Head Group
Wall Unit o o Spacer o o s OH S. COi S
o
=TN=
o
o oii
w
° = p = °n
o r
OH O-°v V c
O OH 4 Figure 5; Fyles Suite o f P o re Form ers 7^°that wall units 1 and 2 w ere similar but m ore active than 3 and that in general the spacer with the hydrogen bonding capability 4, w as a better spacer than 5.
Perhaps the m ost simple channel to date, from a synthetic perspective, has been achieved by M enger’. The channels have th e general structure R0(CH2CH2 0)nR' w here R, n, and R ' are varied. T he b est results are obtained when
R=CH3(CH2)ioC=0 , n=5, and R '= -C H2Ph. B ecause the “ion-conducting segm ent”, the poly(oxyethylene), is only long enough to span h alf the bilayer a minimum o f 2
molecules would be required for transport. It does not require a huge leap o f the imagination to see th at this tran sp o rter vaguely resembles the amphotericin structure although the authors do no t allude to this possibility. Transport was lost if
R =R '=dodecanoyl o r if R = R '= B z o r if R =C H3(C H2)4C= 0 and the activity was impaired if n=3. T he authors conclude that th e benzyl group is necessary to interact with the m em brane polar head groups (quaternary N) by ion -dipole interaction. It w ould be interesting to know w hat the effect would be if it w ere substituted by a hydrogen.
K o b u k e^ describes an ion channel that is com posed o f aggregates o f the ion pairs o f a poly(oxybutylene) with a carboxylate group at one end . and a quaternary ammonium group disubstituted with tw o long alkyl chains (Cis) (Figure 6). The tw o lengths investigated, n=2 and n=3, resulted in overall lengths o f 24 and 30Â
respectively and both w ere equally active. T he long alkyl chains on the ammonium nitrogen provided the hydrophobic regions fo r incorporation into the membrane. They propose the ion pair w as incorporated into the bilayer followed by self-assembly where the polar g ro u ps form ed a pore through w hich the ions could pass (Figure 6). The channels w ere selective for cations. Substituting the carboxylate head group for a phosphate^"* opened the possibilities for a voltage gated channel.
o
-Figure 6; K obuke's Ion Pair Aggregates
In 1993 Ghadiri reported^^'^*’^’ a self-assembling nanotubular array o f cyclic peptides capable o f ion transport (Figure 7). The macrocyclic peptides were
composed o f 8 alternating o- and l- amino acids which formed a flat ring th at would stack one on to p o f the other and w ould be held in place by 8 alternating d o n o r / acceptor hydrogen bonds to each face. The amide backbone is perpendicular to the plane o f the ring with the R groups o f the amino acids anchoring the assembled channel in the lipid. These channels w ere reported to transport Na* and K^ alm ost 3 times faster than gramicidin A_ W hen 8 amino acid residues w ere used, a pore
diameter o f approximately 7.5 A w as formed. The properties o f the resulting channel may be altered according to the num ber and type o f amino acids used^*. A m olecular dynamics stu d y ^ has shown the aggregate to be fairly rigid and upon minimization the
F ig u re ? ; Ghadiri’s Cyclic Peptide26
tube-like structure does not rearrange. Simulations also found that w ater molecules placed inside the cavity did not com pete for the hydrogen bonding sites.
Regen has designed a steroid based amphotericin mimic**. The steroid has tw o side chains com posed o f poly(oxyethylene) subunits (Figure 8). The steroid moiety is expected to en ter the bilayer, incorporating at least one o f the hydrophilic chains (Scheme I). This structure resembles amphotericin in principle by possessing a lipophilic face and a hydrophilic face. In fact, the activity w as less than th a t for amphotericin.
OH
Figure 8: Regen's S teroid Based T ransporter
o
S
o 7 oS
o?
oS
o HO HO o=< HO HO1.3.4 Single Molecule Channels
AU single molecule channels to date have had tw o characteristics in common. First, as stated above, they m ust span the thickness o f the bilayer. Second, each end o f these channels are sufiBciently hydrophilic to interact w ith the aqueous medium on either side o f the bilayer thus anchoring the channel in place. The im portance o f the o th er design criteria are less straightforw ard to explore and have as yet to be resolved. H ow ever, as w ith the dimers and aggregate pores, it is believed that the interior o f the channel m ust provide both hydrophilic (to complex with ions and allow their passage) and hydrophobic (to interact w ith lipid) regions.
40
R = CH(Me)
Figure 9: N olte's Isocyanide Polymer^°
The first non-peptidic single molecule channel was prepared by N olte et al'^”"” . The isocyanide polymer, substituted at each residue with a crown ether, could
potentiaUy have 4 functioning channels per molecule. The polymer adopts a helical conform ation com pleting one full turn every 4 residues. They built structures com posed o f 40 m onom ers resulting in a length o f approximately 40 Â (Figure 9).
The m ost complete structure-activity study o f single molecule channels to d ate has been accomplished by the Fyles g r o u p T h e ion channels were assembled having the basic structure as shown in Figure 10 which w as a hexa-, tetra- or di substituted 18-crown-6 ether (18-C-6) prepared from 10. The channel consisted o f a core acid. This 18-C-6 could act as the filter. In other words, ion transport selectivity could be governed by the crown ether and display a similar trend to cation selectivity in ion binding. Structural control w as also investigated by changing the
stereochem istry at the core, ie. 6 vs 7 and 8 vs 9 Linked to the core were a variety o f possible macrocyclic wall unit structures that w ere varied in their
hydrophobic/hydrophilic content and in their length. The polar head groups investigated w ere the same as those used in the pore formers’’^ . Using a m odular approach cores, walls, and head groups could be mixed and matched to properly assess the demands o f a functioning channel and optimize those requirements T w enty-one o f the possible one hundred com pounds were prepared. Careful
interpretation o f the data revealed that so m e o f the subunits routinely displayed p o o r activity; these are the wall units 14 w hich possibly are too long and 12 which may be to o hydrophilic. They also dem onstrated that all components o f the channel are necessary by testing partially built m odules for transport. The most active wall units appeared to be 10 and 11 It is these wall units from which the designs in this thesis take root.
HO^
Polar Head Group --- ► “ ®n ?
s ^ o o f
X -*--- Polar Head Group — ~ ^ \ j
o = / ^ o Wall Unit o=T*~Vo 10. X=Y=CH2 11.X=CH2. Y=0 12. X=Y=0
■
S=
^ ( y ^ . . . 0 0 2 QzC O O OzC^^O OyCozcAo
cr
Ozc^o
( T \ ^ ° V ^ o o 2 ' 1 ^ 0 ^ ^ * 02C...O &Y
0 2 ( Y < C ° ^ cfecr^o cr •C0 2 ozCT^o cr 6 k ^ O v ^ k . o ^ 9 7Figure 10; Fyles G roup o f Single M olecule Channels
Jullien and Lehn have designed a channeP^ (Figure 1 la ) in analogy to the Fyles’ channels. The basic design may be thought o f in th e same way with an 18-C-6 core and four wall units oriented axially from the core. It acts as both a hydrophilic com plexation site for the cations as well as serving to select the cations that may pass. The poly(oxyethylene) chains also provide binding sites fo r cations. The length o f the chains is again governed by the thickness o f the bilayer. T h e wall units contain phenyl
g roups that provide som e rigidity fo r the system and also serve as a chrom ophore fo r easy U V detection in the m em branes. The amides provide increased rigidity and the n- butyi group o f the tertiary am ides increase the lipophilicity. C arboxylate anions
provide good anchors by interacting with the aqueous medium at each end o f the channel and should ensure th e m olecule spans the bilayer. M olecular m odelling suggests the fully extended length o f the molecule is 45-50 Â.
Lehn further investigated this type o f structure but changed the 18-C -6 core by replacing it with a p-cyclodextrin (Figure 1 lb). In Lehn’s design^^*^"* th e wall units are attached via an ester linkage to seven prim ary hydroxyls on th e narrow rim and to seven o f the available fourteen secondary hydroxyl’s on the w ide m outh o f the cavity. This provides a total o f fourteen wall units (seven up and seven dow n). C arboxylate serves as the polar head g ro u p and either poly(oxyethylene) units o r polym ethylene units make up the rest o f the arm . T he phenyl groups are present for rigidity and as a UV chrom ophore. The total length o f the channel was estimated to be approxim ately 50 Â in length. W hen tested fo r transport activity^^ the poly(oxyethylene) w as very similar to the polymethylene analogue. These results are similar to the results obtained w ith the Fyles suite o f wall u n i t s A s far as Na* and Li* transport is concerned there w as no significant difference in any o f the molecules. All transport rates w ere very slow, on the o rd er o f 10'^ min'*. T he sm aller molecules that w ere not long enough to span the bilayer w ere confirm ed carriers. This is in accordance with the design.
a) b) MeO R O ' o
y
". ----i; 33J4Figure 11; Lehn's Channels, a)18-C ro\\n-6 Based Structure*. b)Cyc!odcxtnn Based Structure
Gokel et designed a channel (Figure 12) based on three diazacrown
subunits. They p ro p o se that a hydrophilic crown at both ends w ould fix the channel in the bilayer. A nother crow n at the midpoint would provide a hydrophilic site for complexation o f the cations. T he length o f the spacers and sidearm s betw een the crowns correspond to the thickness o f the bilayer The hydrophobic sidearms and spacers also ensure th e incorporation o f the transporter into the hydrophobic lipid bilayer. Transport o f Na* w as approximately 4 fold less active than gramicidin A.
Starting from the active com pound 15, Gokel investigated the structure- activity relationship*'* by testing com pounds that vary in one o f the structural units. If com pound 15 is viewed as having th e basic structure as shown in Figure 12 then to test the design criteria they varied th e sidearms, R (Bn, H, cholesterol-ether linked, and cholesterol-ester linked). O ne o f th e variations was to change the spacer from the hydrophobic polymethylene subunit 15 to a poly(oxyethylene) subunit 16. It w as thought that the more hydrophilic oxygens may help move the cations along. In fact, the m ore active transporter was 15, the m ore hydrophobic compound.
Functional groups may be incorporated into the subunits to im part some control over channel openings. F o r exam ple, two other structural variations o f the channel 15 were to incorporate g ro u p s into the spacer so as to act as a gated channel. Instead o f the channel remaining op en , the cis-trans geometry o f an amide 17 may open and close the channel. Also by incorporating an anthraquinone spacer unit 18 which is capable o f accepting 1 o r 2 electrons to form a stable anion, cation flow may be stopped due to the “blockage” o f the channel. The central unit w as changed from a diazacrown to a poly(oxyethylene) chain. T he structural significance o f the crow n was also investigated using a sm aller crow n ( 15 membered). The sidearm s w ere changed to benzyl groups to investigate the effect o f aryl groups on transport.
C ompound 15 was found to be the best transporter when analyzed using fluorescence m ethods while compounds 16. 17, and 18 w ere comparable to gramicidin. The conclusion from this study was that a to o hydrophilic interior is detrimental to
th e m ore hydrophilic molecule in the lipid. T h e au th o rs conclude that the channels require th e com ponent parts o f 15 due to th e lack o f activity when parts are missing.
In order to address the question o f w h e th er th e transporter is incorporated into the lipid the octanol-w ater partition coefficients w ere determined. Fortunately, ail o f the com pounds w ere found to be completely soluble in the octanol. In an attem pt to categorize the transporters as carriers o r channels G okel et al.^’ have com pared the rate o f transport across bulk C H C I3 relative to th e carrier valinomycin and with th e
rate o f transport across a bilayer membrane relative to the channel gramicidin. Since the distance across the CH CI3 layer was to o g re a t fo r a channel to span, then the fast
rate in C H C I3 and the slow rate in the bilayer w as interpreted to be indicative o f a
carrier. It was previously postulated that th e tran sp o rter 19 acted as a channel and the benzyl ring stabilized the head groups by interaction with the polar head g roups o f the bilayer. They postulate that if the crow ns are head groups through which the ions pass then changes in the electron density o f th e benzyl group should influence
transport activity. Derivatives 20 and 21 w ere prepared and the electron w ithdraw ing N O2 group in 20 reduces the transport while th e electron donating methoxy 21 in enhances the rate” .
Polar H ead G roup Spacer C ore Side arm 15, R=(CH2)nCH3, x=C,2H24 16, R=(CH2)i iCH3, X=-CH2(CH20CH2)3CH2-(CHzk Q (CHik 17, R=(CH2)i 1C H3, X= (CH2K (CHzk 18, R = (C H 2 )n C H 3 , x = 19, R=Bn, x= Ci2H24 2 0 , R = / ^ < ^ - N 0 2 . X = Ci2H24 21, R= , X = Ci2H24
Figure 12: Gokel's Family o f Single M olecule T ransporters ••’■36 j?
Similar to N olte’s helical crow ns, V oyer recently exploited the helix form ing ability o f L-leucine and combined it w ith L-phenylalanine substituted with a 21-crow n-7 m oiety (Crown)^*. The peptide sequence is /B O C -N (Leu-C row n-Leu-Leu-Leu- Crow n-Leu)3 0Me. The crown e th er phenylalanine residues are strategically placed so the crow ns appear stacked, form ing a crow n ether pore adjacent to the helix as shown
in Figure 13. T he N term inus is protected with a f-butyloxycarbonyl group and the C terminus is protected as the methyl ester. These protecting groups decrease the hydrophilic character o f the helix and probably serve to increase the solubility o f the com pound into the lipid. Presumably the crown ethers interact w ith the aqueous medium outside the bilayer.
R=
Figure 13: V oyer’s Stacked C row ns38
1.4 Rationale for Synthetic Study
T o summarize, a w ide range o f compounds facilitate transport. Channels may be categorized as tw o o r m ore molecules that aggregate, as dimers, and as large single molecules Yet, these com pounds have some properties in comm on; they all have a similar length, they all have polar head groups and they all have amphiphilic character. But the structure-activity studies are far from over. For example, m ost o f the
structural issues pertaining to mechanism remain unresolved. The needed structural and functional investigations will require the efBcient synthesis o f a wide variety o f ion
channel mimics. Consequently, w e need to explore facile, efBcient synthetic pathways to ion channels.
CHAPTER 2
INTRODUCTION TO DESIGN AND SYNTHESIS
2.1 Design of Ion Transporters Via a Modular Assembly Approach
The modular approach to assem bling ion transporters provides an eflScient method for assessing the structural requirem ents for an active ion chaimel. As
discussed in the introduction, th e Fyles group has synthesized a number o f macrocyclic wall units for this purpose. T hese w all units have been incorporated in both single molecule charmels and in pore form er molecules. Synthetically, the wall units are the m ost complex o f the building blocks, especially in the aggregate p ore family o f molecules. For this reason, an efBcient and easily modified m ethod for obtaining macrocyclic wall units is desirable. T he remainder o f this dissertation is dedicated to the design and synthesis o f m acrocycles that can be used as wall units. The
macrocycles discussed are not identical to those previously studied but are designed to preserve desireable features while also improving the efSciency o f the synthesis. This will eventually lead to a new m o d u lar set for synthesis o f new candidates for
evaluation by structure-activity studies.
2.2 Strengths and Weaknesses o f the Previous Wall Units
The previous set o f m acrocycles w as designed to investigate the optim um length and lipophilic compatibility. A short-hand nom enclature consistent w ith the modular assembly was developed^^ to assist in identifying each structure. T h e wall
units all have the same basic stru ctu re show n in Figure 14. The label T rg represents triethylene glycol, the numbers 5, 8 and 12 represent the length o f the m ethylene chain and the subscript 2 indicates both sides o f the wall are the same. O f the wall units investigated, 5;, 8% and 12% serve to investigate the length criteria. The wall units 8% and 8T rg were the most active o f this set and w ere estimated to be 13.5 and 14.0 Â long respectively. This length will be incorporated into the design o f the new systems. Conversely, the macrocycles 82, 8T rg and T rgi are all roughly the same length but vary in their degree o f hydrophobicity. The most active transporters contained the 28 membered macrocycles 82 and 8T rg. T he building block Trg2 did not su p po rt transport and a rationalization is th a t it is too hydrophilic.
C c x c s )
122
Figure 14; Previous Set o f Wall Units
The yields o f the m ore active m acrocycles 82 and 8Trg from commercial precursors are 12% and 6.4% respectively^®. In a second generation m odular set it is
desirable to have much higher yielding macrocyciization reactions. Therefore, the new set o f m acrocycles should be designed to include m ethods capable o f high yielding macrocyciization.
The m ost active linkers o f the previous set o f pore form ers w ere derived from tartaric acid^® and it was believed the hydrogen bonding capability o f the linker may assist in the aggregation. Current w ork within the groupé” questions this conclusion but does not alter the basic goal o f this dissertation. New wall units should be designed that may incorporate a variety o f functionality so that investigation o f
hydrogen bonding capability in the wall units can be done. Functionality also could be incorporated to investigate rigidity or further investigate polarity and functionality issues.
An inherent flaw in the design o f the previous set o f wall units arises in the addition o f cores o r linkers and polar head groups. Although James^’ describes the Michael addition as being rather simple chemistry, there is no regio- o r stereochemical control o f the Michael addition o f the sulfur nucleophiles to the maJeate esters (Figure
15). In the case o f 8Trg, w here x and y are different, the addition o f the head group and linker gives rise to 16 possible isomers. So when the wall units are incorporated into a transport molecule, for example, a pore form er where there are tw o moles o f m acrocycle for every mole o f transporter, then there are 16^=256 possible structures. In the simpler case, 82, where there is a plane o f symmetry (x=y), the number o f possible isom ers is 8. This is less than 8T rg but once tw o molecules are linked together then the number o f possible structures for the transporter is still 8^=64.
O' o L o + enantiomer + 2 diastereotners + enantiomer + 2 diastereomers + + enantiomer
+ 2 diastereomers + enantiomer + 2 diastereomers
Figure 15 ; Stereo- and Regio- Isomers From Previous Set o f W ail U nits This is a potential problem when drawing conclusions about the effect o f th e structure on the transport activity since it is possible to argue that o f the m ultiple com pounds present, perhaps only a fraction are active. The only way to tell is to separate and purify each com pound. Obviously this is not a reasonable o p tio n as th ere are too many com pounds with similar physical properties to perform even a crude separation. The numerous isom ers appear to be very similar by nuclear m agnetic resonance and gel permeation chrom atography but the potential problem rem ains. The new wall units m ust avoid this problem by producing single isomers o f the final
2.3 Property Directed Synthesis
The approach taken in this thesis is the property directed synthesis approach as outlined by Gokei'*^. A property directed synthesis differs from a natural products o r target directed synthesis in th a t the exact target is not dictated by structure but by function. This permits the "target" to be modified by th e dem ands o f th e synthesis and so as long as it perform s the desired function, the synthesis will have been achieved. Scheme 2 is a flow chart that dem onstrates the typical sequence o f events in a property directed synthesis. First a ta rg e t is identified. In the context o f this
dissertation the target is a wall macrocycle o f a certain length (betw een 15 and 17 Â) and sufficiently rigid so that it does not fold. This targ et should have a com plem entary balance o f hydrophobic and hydrophilic subunits, and should have the potential to incorporate a variety o f functionality to investigate the different properties (hydrogen bonding, rigidity requirem ents). In o rd e r that the tran sp o rters can be studied in detail, they must have an efficient, facile synthesis.
D esign Target
Evaluate Functionality Devise Synthetic Plan
Evaluate Synthesis
Synthesis
Once a target has been designed one must then devise a synthetic plan. In this case macrocycles are the target. W hen preparing the synthetic plan it should include m ethods that produce high yielding reactions so bulk material is available and readily incorporated as a building block into a transporter molecule. As w ith the synthesis o f a previously unknown com pound, th e synthesis is constantly being evaluated and revised throughout the course o f carrying out the plan.
Property directed synthesis differs from target directed synthesis such that not only can the synthetic plan be m odified but also the structure o f the targ et molecule may be changed in response to synthetic setbacks. This is possible because it does not m atter what the actual structure is, only that it meets the predeterm ined requirements and performs the desired function. In this manner the synthetic path to a target will be optimized and a macrocycle will b e obtained. The m acrocycle will be assessed on the basis o f structural properties and synthetic efficiency. O nce a m acrocycle is realized it may then be incorporated as a building block into potential transporter molecules, using a variety o f linkers o r cores and head groups, and the properties o f the transporter may be evaluated. From this knowledge new m acrocycles possessing different properties may be designed, synthesized and tested
In practice th e n , we are using a property directed modular assembly approach to study ton transport. The flow ch art in Scheme 2 is applied first in designing
macrocyclic wall units, as in this dissertation, and then it may be subsequently applied to the actual transporters.
T o summarize, the desired properties o f the targ et wall units are that they be macrocycles available in high yield, preferably crystalline solids so they may be investigated structurally by X-ray crystallography, o f a length betw een 15 and 17 A, and that they in principle could Incorporate a variety o f functionality.
2.3.1 Introduction to Synthetic Efficiency
Typically in natural product synthesis a synthetic pathw ay is judged based on an overall yield from a defined starting material. I f the overall yield is high, or higher than a competing route to the same molecule, then the synthesis is said to be efBcient. The overall yield can be misleading because it takes into account the yields o f the reaction starting from only one o f the reagents. A paper published by Hendrikson in
1977'*3 (ligcusses the concept o f efficiency more generally by examining the nature o f the synthetic sequence, the number o f steps, which is effectively a discussion o f yields, the time required for each step including purification, and the cost and amounts o f starting materials and reagents
2.3.1.1 Nature of the Synthetic Sequence
The nature o f the synthetic sequence pertains to the o rd e r in w hich the reactions are done and to th e ty p e o f connections made. In o rd e r to exam ine this Hendrikson^^ uses a plan graph. Every d o t on the graph represents an isolable intermediate and the dots are coim ected by lines that are indicative o f the type o f reaction. The different types o f reactions are those that build th e targ et skeleton and then there are extra steps w hich refiinctionaiize, protect and deprotect.
i= 2
s=5
4 3 2 1 0
Figure 16: Example o f a Plan Graph
Figure 16 is an exam ple o f the plan graph. There are / starting m aterials and on the graph they are labeled from /=1, 2, 3,...and up. They are assigned a rank, /, which corresponds to the num ber o f steps they are away from th e target. Horizontal lines connecting dots are co u n ted as steps in the synthesis while diagonal lines are used to converge the synthesis. F ig u re 16 show s a plan for a synthesis from 3 starting m aterials that is com pleted in 5 steps. The target is at rank 1=0 and th e first starting material has the largest rank because it is furthest away from th e target. The
horizontal line containing r= l is know n as the main line. The 5 steps are a result o f the four steps along the main line plus the step for starting material /=3 from rank 1=2 to
1=2. Instead o f calculating an overall yield which w ould presumably arise from the starting material w ith the largest rank, w e calculate the total am ount o f starting material used (W). The overall yield from one starting material is the inverse o f the am ount o f the starting m aterial required to yield a mole o f final product. For example, if an overall yield is 3% then 1/.03=33.33 moles o f starting material is needed to produce 1 mole o f target. In th is analysis w e take into account the amounts required o f all the starting materials.
For the purposes o f th e analysis, one assumes an average yield o f 80% per step, with the inverse o f the fa c to r 0 .8 0 being 1.25. This means that in a one step reaction, 1.25 moles o f starting m aterial is needed to obtain 1 mole o f product. If the reaction is a tw o step process th e n 1/(0.80 x 0.80)= 1/0.64= 1.56 moles o f starting material is required. Since w e a re using an average, then steps known to give low yields may be treated as extra steps. F or example, a m onoprotection o f a symmetrical precursor may be treated as 3 step s (0.80 x 0.80 x 0.80 = 0.51) since the statistical maximum yield is 50%. In som e cases there are methods to get around this problem (for example, as the product form s it m ay be removed from the reaction mixture by distillation). If such m ethods a re possible then higher yields are expected. Table I provides the calculated yields, y , inverse yields, x, and sum o f inverse yields, 5/, for an average 80% yield at each step, /. O f course, if the real yields are known, then the actual performance can be calculated in the same way.
Table I; Tabulation o f predicted yields, inverse yields, and sum o f inverse yields. 1 y X s, I 0.800 1.25 1.25 2 0.640 1.56 2.81 3 0.512 1.95 4.77 4 0.410 2.44 7.21 5 0.328 3.05 10.26 6 0.262 3.81 14.07 7 0.210 4.77 18.84 8 0.168 5.96 24.80 9 0.134 7.45 32.25 10 0.107 9.31 41.57 11 0.086 11.6 53.2 12 0.069 14.6 67.8 13 0.055 18.2 85.9 14 0.044 22.7 108.7 15 0.035 28.4 137.1 16 0.028 35.5 172.6 17 0.023 44.4 217.0 18 0.018 55.5 272.6 19 0.014 69.4 341.9 20 0.012 86.7 428.7
2.3.1.2 Materials
The weight (W,) o f the starting material (/) required to obtain I mole o f target is W,=M,x''=Fn,x''. F is a constant used to facilitate com parisons betw een different syntheses where all the details are not yet resolved. It is estimated to be 14 which is the weight o f a CHz group or N. O xygen atom s are underrepresented because their weight is 16 but the error is small and should be fairly consistent in the comparisons. The total weight o f all starting m aterials (W») is given by Wo=FZn^x*, w here n, is the number o f structural and heavy atom s in starting material /. Since F i s a constant we can leave it out and discuss the relative w eights obtained from W=Sn,x^'.
2.3.1.3 Reagents
In estimating the am ounts o f reagents used we w ork in molar amounts. Assuming a 1 ; I stoichiometry o f starting materialsireagents, a rough estim ate may be produced. This is probably only useful if little is know n o f th e synthesis. T o do this the main line is taken as the line with the highest ranking starting material. A
convergent plan has sublines that merge with the main line. In order that each intermediate is assigned to only one starting m aterial, /, w hich passes through it, a new term (/') is introduced. E ach subline begins at rank / and ends at rank /' which is the rank o f the last independent interm ediate before it joins th e main line. So the sum, 5/=Zx, corresponds to the molar amounts o f starting materials o r reagents since w e assume a 1:1 stoichiometry. So that w e only count the reagents once, w e need to take the sum o f the difference o f the m olar amounts, which leaves the total number o f moles o f reagent as, R=ZAS.
2.3.1.4 T im e
In general the time required for a reaction depends o n the number o f steps, s, and the am ount o f material manipulated. I f we assume that th e reaction tim es average out to a time, To, then the time, T , to reach a target may be w ritten as, T=sToU where u is an upscaling factor taking into account the weight. H endrikson uses the
exponential scaling factor that w as used by Powers'*^. P ow ers found that TqcW^ where z=0.3. This is equivalent to saying that it takes tw ice as long to manipulate ten times the amount. So to calculate the tim e w e need to know the w eight o f each synthon
each time it is p art o f an interm ediate used in a reaction. The to tal w eight
manipulated, TW , becom es FZn^Sj, but again F is a constant and can thus be ignored as w e are com paring only the relative weights and the equation becom es, TW =Sn,S/,. The total w eight TW divided by the number o f steps, s, gives an average weight. This substituted into P o w ers’ equation gives Tqc(TW /s)°^. So, substituted into
H endrikson’s equation w e g et the approximate time ,T=s(TW /s)° ^. Tim e is a relative measure and is useful in com paring tw o o r more syntheses.
These ideas will be developed fo r the specific syntheses o f this dissertation in C hapter 4. W e will develop the corresponding synthetic graphs and calculate the actual W, T, etc. and then com pare them to a current building block'*^ in the m odular set.
2.4 M acro cy clizatio n T e c h n iq u e s
Chem ists have been challenged by the preparation o f m acrocycles and it is only in the past 30 years have they been attainable in appreciable yield**^. T he following section presents some o f the challenges fo r efiScient macrocyclization.
A retrosynthetic analysis o f any macrocycle C w ould first involve the design o f the bifunctionalized precursor B F C that could then react intramolecularly (Schem e 3). Com pound B F C could then be obtained from the bifunctionalized reactants A and B F or synthetic simplicity, the reactants are usually symmetrical with th e sam e
functionality at each end o f the chain'*^. This design avoids the added complexity o f multiple products. A one p o t synthesis is possible since the same chem istry is used to join the fragm ents as to close the ring, resulting in maximum yields. T he chains may
vary but m ust be non-reactive under the conditions o f A+B.
BFC
A
Scheme 3; Retrosynthetic Analysis o f a Macrocycle
There are several possible reaction pathways (Scheme 4). The concerted double reaction (path a) leads to th e desired macrocycle but it is not a likely
mechanism fi'om a therm odynam ic o r kinetic stand point. M ore likely, one end reacts first (path a .l) , leading to th e bifunctional chain (B FC ). Several pathw ays are
available to this molecule. It may cyclize (path a.2) to afford w hat is know n as the 1+1 product. T w o B F C may react (path b. 1) yielding w hat is referred to as the 2+2
product. This product may also be obtained by the stepw ise reactions o f a B F C w ith a strand A molecule (path b.2) follow ed by a strand B molecule (path b.3) o r vice versa, and subsequently, ring closure. This is sometimes the molecule o f interest but, the main competing reaction is polymerization. T he B FC can react with another A o r B strand, etc. to yield polymers o f varying lengths. It is the problem o f polymerization that demands the most attention.
P O L Y M E R
The factor controlling polym erization is the effective molarity (E .M .)^. Referring to Scheme 4, this is defined as Awa/^mter, where ^«0. is the rate o f the intramolecular reaction a.2 and is the rate o f the intermolecular reaction b.2,
assuming all react ions are irreversible. The first A-B bond is formed as one expects o f the uni functional ized com pounds to form the BFC. Now, whether the molecule
cyclizes or polymerizes depends on the relative rates o f the inter (path a.2) and
intramolecular (path b.2) processes. T he rates o f the reactions depend on the relative free energy o f activation (AG^) o f the polym erization reaction with respect to the intramolecular reaction. One m ust consider th e activation enthalpy (AH^) and the probability for the intram olecular encounter o f A and B (which relates to AS^). T h e primary enthalpy difference will lie in the strain energy on ring formation; a larger strain energy will result in a positive, unfavourable AH^ These strains may be caused by imperfect staggering, bond angle deform ation or transannular interactions
(repulsive interactions across the ring). These effects are usually greater for m edium sized rings how ever in general they tend to decrease as ring size increases. The entropy (AS^) o f the system dictates the probability o f end to end encounters. It is easy to imagine that the frequency o f such encounters decrease as the chain length is increased A disordered open chain th at m ust undergo rearrangement to form an ordered ring loses the rotational fi^eedom and results in a negative and therefore unfavorable AS^
There are several m ethods devised to favor intramolecular reaction. The m ost general and widely used technique follows Z iegler’s high dilution principle T he
E.M . is no t ch an g ed , but the concentration o f the substrates is very low in com parison to E.M . The purpose o f this technique is to keep the concentration o f the reacting species low so that the probability o f a B group encountering an A group is higher intramolecularly than intermolecularly. High dilution typically refers to a final
concentration on the o rder o f 10'^- 10"^ M. The high dilution technique requires each o f the reagents to be dissolved in th e appropriate solvent. T hen they are added separately but simultaneously very slow ly via a syringe pum p device into a round b ottom flask that already contains som e o f the solvent and possibly the other necessary reagents. Addition times vary from a few hours to several days and often additional reaction time is needed. This approach has been successfully applied to many systems but due to the large volume o f solvent needed, the lengthy reaction time, and need for specialized equipment, other m ethods have been sought.
In an effort to develop general m ethods w here high dilution is not necessary, the following techniques have been dem onstrated: tem plated reaction, rigid group syntheses, and the use o f high pressure. The first tw o m ethods change the E.M . so the intramolecular reaction is favoured even at “normal” concentrations while high
pressure is effective for the same reasons as is high dilution (vide infra). In this dissertation both the high dilution and rigid group m ethods are employed.
T he rigid group method refers to control o f the geom etry o f the reactants, namely th at they are already in a suitable conform ation conducive to cyclization. Rigid g roup implies that solvent w ould have minimal effects on the geom etries o f the
exam ple is in the preparation o f SchifFbase macrocycles w h ere th e condensation reaction betw een linked benzaldehydes and amines afford yields as high as 94%.50
2.5 Molecular Modeling
T w o o f the desired properties o f the m acrocycles are that the length be betw een 15 and 17 Â and that the structure be sufficiently rigid th at it will not fold or crum ple on itself. The length assessment is straight forw ard in that the distance is measured from end to end in the low energy conformers. R igidity is difficult to assess so m olecular mechanics and dynamics were used. In the first step, w e are interested in the geom etry o f the molecule at the global minimum so w e perform molecular mechanics^V Since mechanics only seeks out the local m inimum then we perform dynamics as a second step to explore flexibility. The co m pu ter simulations assess the relative geom etric rigidity o f the macrocycles to therm al excitation. Molecular dynamics im parts random velocities to each atom based u p o n the tem perature, (i.e., heats the m olecule) so it can escape a local minimum, and th en follow s each atom over a period o f time steps. Assuming that the total time period is long enough and that we sample th e collection o f geom etries (the trajectory) properly, w e can assume that the molecule can statistically sample the potential energy surface available to it at a given tem perature. The geom etries found during the simulation a re not minima or stationary points on the potential surface, but are just snapshots o f the m olecule during its trajectory. I f a given macrocycle can, at a given tem perature, readily achieve a large number o f distinct, low energy structures, we can safely conclude that the macrocycle
is not rigid w ith respect to therm al excitation. K on the other hand, the m acrocycle leaves the low energy regime o f th e starting point and finds only higher regions o v er a long time period, perhaps finding another low energy region some time later, w e may conclude th at the macrocycle is relatively rigid to thermal excitation. T o assess the length o f the m acrocycle it is necessary to sample some o f the snapshots along the dynamics trajectory and then minimize those using molecular mechanics. The larger the sample size, the greater chance o f finding the global minimum. Both mechanics and dynamics results will be discussed as th e target designs are refined.
2.6 Target Evolution
2.6.1 General Strategy
The purpose o f this thesis is to improve on the existing set o f wall units used in the modular set. Specifically, the goals are to obtain macrocycles in high yield, to ehminate the regio and stereochemical problems, and to broaden the understanding o f ion transport via the addition o f a wide variety o f functional groups. The basic design o f the new wall units is to replace the m aleate esters thus avoiding the formation o f multiple regio and stereoisomers. In principle, w e could work fi'om chiral precursors to ensure single isomers result in the target. Rather w e chose to use a plane o f symmetry at all connection points to make achiral wall units. Thus, 1,3,5-trisubstituted benzenes are placed at either end o f the macrocycle w here tw o o f th e substituents are the same (A o r B ) and the other is
Scheme 5: General Design o f N ew Wall U nits
different (X o r Y) (Scheme 5). T he sides o f the macrocycle (G ) could be the same, o r could incorporate different functionality. The phenyl ring a t each end additionally imposes some rigidity on th e system. This chapter discusses the evolution o f the
44
targ ets in general, although many o f th e decisions w ere based on th e detailed chem istry to b e presented in chapter 3. Som e o f th o se results are briefly stated in this section simply to indicate the rationale. A detailed discussion o f the results is deferred to ch ap ter 3.
2.6.2 Design of Macrocycles Based on a-Heteroamides*^
Nature uses peptide linkages to impose structural order using the rigidity o f the amide link and we will use amide bonds fo r this same reason, to ensure rigidity o f the system. CPK models and molecular m echanics were used to determine the overall shapes and energies o f the structures, ie. w hether the macrocycle had the desired length and rigidity. Secondary amides would be an ideal functional group in that they not only have a high barrier to rotation, but they possess a hydrogen atom that is capable o f hydrogen bonding to another member o f the ring system^^"^^
22 a, X =Y =H
Figure 17; T a rg e t 1, M acrocycle 22
A structure that seemed favourable fi'om both structural and synthetic points o f view is 22 (Figure 17). Substituents X and Y could be different and could be any one o f a variety o f functional groups, such as -N O2 o r -NH2. At this stage, the exact chemical
functional groups are not particularly important, except that they allow a convenient chemical link between a head group at one end and a linker at the other end. For reasons including cost and simplicity, the model compound 22a was targeted to examine the chemistry and the overall structure o f the macrocycle. M olecular modelling predicted the length between the 5 and 5' positions o f the two benzene rings to be approximately 13.5 A.
A retrosynthetic analysis o f 22a (Scheme 6) shows that the
o X o 22a, X =Y =R=H 25a, X=Y=H, R = C H3 V o R R - S , o 24 a, R=H, X=H b, R=CH3, X=H 23, V =H Schem e 6; Retrosynthetic Analysis o f Targets 22 and 25
structure could result from the reaction o f diol 23 and the bis chloroamide 24a, which is believed to behave as a rigid group*^^. The bis chloroamides supposedly form a rigid, claw-like type structure’^ ^ (Scheme 7), conducive to macrocyclization. For this reason, low dilution techniques may be used. Bradshaw and co-w orkers report excellent yields for
