A new method for the polymer-supported synthesis of cyclic oligoesters for potential applications in macrocyclic lactone synthesis and combinatorial chemistry

Article abstract of DOI:10.1039/b110528k

Attachment of ω-hydroxyalkanecarboxylic acids to Merrifield beads followed by treatment with a catalytic amount of di-n-butyltin oxide in chlorobenzene at 133 °C for 4-18 h brought about the formation of the corresponding cyclic oligomers (COs) as the mai

Full text of DOI:10.1039/b110528k

View Article Online / Journal Homepage / Table of Contents for this issue
A new method for the polymer-supported synthesis of cyclic
oligoesters for potential applications in macrocyclic lactone
synthesis and combinatorial chemistry
1
Clare L. Ruddick,† Philip Hodge,* Anthony Cook and Andrew J. McRiner
Chemistry Department, University of Manchester, Oxford Road, Manchester, UK M13 9PL
Received (in Cambridge, UK) 16th November 2001, Accepted 7th January 2002
First published as an Advance Article on the web 4th February 2002
Attachment of ω-hydroxyalkanecarboxylic acids to Merrifield beads followed by treatment with a catalytic amount
of di-n-butyltin oxide in chlorobenzene at 133 ЊC for 4–18 h brought about the formation of the corresponding
cyclic oligomers (COs) as the main (>92%) soluble products in yields of 56–83%. Similar results were obtained with
polymer-supported (PS) ricinoleic acid, which has a secondary hydroxy group, but attempts to carry out analogous
reactions with PS lithocholic acid failed. PS 20-hydroxyicos-10-enoic acid and N-(11-hydroxyundecanoyl)-
11-aminoundecanoic acid reacted well to give COs, the cyclic monomers being the major products. Samples of
these cyclic monomers were isolated. The latter two hydroxy acids were also successfully assembled on the beads
(the former using olefin metathesis) and then cyclo-oligomerised. There are several possible applications for this PS
synthetic method. Thus, the COs might be used, via transesterifications, for the preparation of soluble combinatorial
libraries of macrocycles, or used as starting materials for the small-scale combinatorial synthesis of copolyesters.
For both of these applications the fact that the PS syntheses afford families of COs is not a problem because the
applications involve the establishment of equilibria into which all of the COs can feed. Finally, these PS reactions
may provide a useful approach to the small-scale synthesis of macrocyclic lactones with 15 or more ring atoms,
since with these ring sizes the cyclised monomers are formed in >67% yield.
yields of cyclic oligomers (COs) essentially free of linear oligo-
mers or other species. It was not specifically intended to achieve
site isolation. There are three potential applications for such PS
syntheses. Firstly, in appropriate cases the COs may be used, via
Introduction
There is considerable interest in the synthesis of macrocyclic
lactones,1,2 not least because there are many important natur-
ally occurring compounds of this general type. For example,
transesterifications (TEs) at high dilution, for the preparation
erythromycin A, whose structure includes a 14-membered lac-
of soluble combinatorial libraries of macrocycles.19 Secondly,
tone, has significant antibacterial activity,3 and valinomycin,
the COs may be used as starting materials for a novel type of
whose structure includes a 36-membered ring incorporating six
entropically driven ring-opening polymerisation that simply
ester and six amide linkages, is important in the transport of Kϩ
involves heating the neat COs with a TE catalyst.20 Such poly-
across biological membranes.4 The classical approach to the
merisations offer the opportunity to synthesise combinatorial
synthesis of such macrocycles is to cyclise an appropriate α,ω-
bifunctional compound at high dilution so that first order
intramolecular reactions are heavily favoured over second order
libraries of copolyesters.21 Note that for both of these applica-
tions the fact that a family of COs is produced is not a problem
because the applications involve the establishment of equilibria
into which all the COs can feed. Finally, in certain cases the
reactions may provide a useful approach to the synthesis of
macrocyclic lactones.
To carry out this new type of PS macrocycle synthesis, an
unprotected ω-hydroxy acid is bound, at loadings as high as
conveniently possible, to Merrifield beads 1 via ester linkages
(see Scheme 1). TE is then brought about by heating the beads
in the presence of a suitable catalyst. Both chain growth (for
example, “a” in Scheme 1) and cyclisation reactions (for
example, “b” in Scheme 1) occur, but only the cyclic products are
released into solution. The linear products remain bound to
the beads via their end groups. It was initially anticipated that
the cyclic products in solution would then equilibrate, again via
intermolecular reactions, i.e. to carry out the cyclisation at a
concentration significantly below the “effective molarity”.5
Polymer-supported (PS) reactions have been considered as an
alternative to high dilution syntheses6 because, under appropri-
ate conditions, they can provide some “site isolation”.7–10 Con-
sequently several PS syntheses of macrocyclic lactones have
been reported where the aim was to obtain the cyclised mono-
mer in high yield with the help of site isolation.11–18 In most
of these cases, however, either the loadings on the beads were
low, the yields of the cyclised monomers were modest, and/or
chromatography was needed to isolate the desired products.
In only a few of these reaction systems was any site isolation
actually demonstrated.12,14–17
This paper presents a further method for the PS synthesis of
TEs, so that eventually, whatever COs were formed initially, a
macrocyclic lactones. The aim in this case was simply to obtain
family of COs would eventually be present in their equilibrium
in solution, by an experimentally convenient procedure, good
proportions. However, the results presented below indicate
that, with the most successful TE catalyst used, little or no
equilibration occurs, i.e. the composition of the CO products is
largely kinetically controlled.
† Current address: Mettler-Toledo Myriad Ltd., 2, Saxon Way, Mel-
bourn, Hertfordshire, UK.
DOI: 10.1039/b110528k
J. Chem. Soc., Perkin Trans. 1, 2002, 629–637
This journal is © The Royal Society of Chemistry 2002
629

Table 1 Cyclo-oligomerisation of various polymer-supported hydroxyacids
Percentage composition of cyclic n-mers by weightc
Ring atoms
per repeat unit
Loading/
Reaction
time/h
Yield of
solubles (%)
Percentage
cyclic (%)b
Entry Hydroxyacid starting material
mmol gϪ1 a
n = 1
n = 2
n = 3
n = 4
n = 5
n = 6
n = 7d MALDI-TOF MSe
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
11-Hydroxyundecanoic acid (2)
11-Hydroxyundecanoic acid (2)
11-Hydroxyundecanoic acid (2)
12-Hydroxydodecanoic acid (3)
12-Hydroxydodecanoic acid (3)
8-Hydroxyoctanoic acid (4)
10-Hydroxydecanoic acid (5)
Ricinoleic acid (13)
12
12
12
13
13
9
11
13
14
16
17
21
21
21
24
24
1.15
0.39
0.10f
0.44
0.25f
0.47
0.65
0.37
0.63
0.96
0.42
0.46
0.14f
0.61
0.46
1.17
4
4
4
4
4
18
18
18
18
18
18
18
18
18
18
18
66
98
96
99
96
98
100
98
1.0
14
36
4.1
12
0
67
75
60
77
72
76
83
58
—
31
35
7
22
6.4
1.5
0.4
6.3
2.8
4.9
3.0
1.7
—
1.8
2.9
1.8
0.4
0.6
0.5
0
2.2
0.2
0
1.5
0.3
1.3
0.7
0.3
—
0.5
1.1
1.0
0.1
0.4
0
1.2
0
0
0.5
0.2
0.5
0.2
0.1
—
0
0.6
0
0
0.2
0
0
0.2
0
0
0.6
0.7
0.2
0.1
0
—
0
11
—
—
6
—
8
13
—
—
—
—
—
—
—
7
25 (74)g
3 (34)g
55
9.1
3.6
10
12
17
13
9.9
—
6.7
8.4
3.2
2.5
2.8
2.3
2.0
22 (46)g
67
56
69
14
81
83
70
61
0
90
30
—
60i
51i
87i
92
82
74
83i
h
Lithocholic acid (15)
—
15-Hydroxypentadecanoic acid (17)
16-Hydroxyhexadecanoic acid (9)
20-Hydroxyicos-10-enoic acid (19)
20-Hydroxyicos-10-enoic acid (19)
Undec-10-enoic acid (23)j
Hydroxyacid (25)
98
96
97
99
92
100
100
1.0
0
0
5
61
14
18
15
0
97f
58
5.2k
0
Hydroxyacid (30)l
0
—
a See Experimental section for loading procedure. b Calculated from the average degree of polymerisation of the soluble product as estimated from the SEC analysis, and the number of end groups per repeat unit as
1
estimated from the H NMR spectrum. c By SEC-analysis. d Except where indicated otherwise, these were higher linear and/or cyclic oligomers. e Samples of selected products were studied by MALDI-TOF mass
spectrometry. Peaks due to cyclics were clearly seen for cyclic n-mers from n = 3 up to the values given. f Sample used was the polymer beads recovered from the experiment summarised in the preceding entry. g First yield
is that based on the original starting material. Yield in parentheses is yield based on the starting material for the experiment summarised in that particular entry. h SEC trace showed many peaks. No assignment possible.
i A sample of the cyclic monomer was isolated by column chromatography. j Subsequently converted on the beads into PS hydroxyacid (19) at a loading of 0.21 mmol gϪ1 k A portion of this product (5.2%) consisted of
.
the products with n ester groups n ϩ 1 amide groups. Yields were 4.0, 1.1, 0.2 and 0.1% for n = 2 to 5 respectively. l Subsequently converted on the beads into PS hydroxyacid (25) at a loading of 0.82 mmol gϪ1
.

View Article Online
smaller a given ring is the more of it is present. Consistent with
this, when strained medium-sized rings could be formed, very
little of those cyclic products are present in the mixtures of
COs. When the cyclic monomer has 15 or more ring atoms it is
the major product, and, as there is a significant jump in ring size
on going from the monomer to the dimer to the trimer etc., it is
relatively easy to separate the cyclic monomer from the higher
oligomers.
Cyclo-oligomerisations of PS 11-hydroxyundecanoic acid (2), PS
12-hydroxydodecanoic acid (3), PS 8-hydroxyoctanoic acid (4)
and PS 10-hydroxydecanoic acid (5)
In view of the ready availability of 11-hydroxyundecanoic acid
2 and the corresponding COs 6 and linear oligomers,22 the PS
cyclo-oligomerisation of this acid was selected for detailed
study. Three possible catalyst–solvent systems for TE were
investigated. Treatment of the PS acid (0.64 mmol gϪ1) with 0.3
equivalents of potassium methoxide in toluene at 110 ЊC for 4 h
gave a 33% yield of soluble material. SEC and 1H NMR spectro-
scopic analysis, in comparison with authentic samples, indi-
cated22 that it consisted of 80% cyclic oligoundecanoates 6 and
20% linear oligoundecanoates. Lithium bis(trimethylsilyl)amide
(5 mol%) in chlorobenzene at 133 ЊC for 4 h gave a better yield
of soluble product (59%), but again it consisted of both cyclic
products 6 (74%) and linear oligomers (26%). Finally, treatment
with di-n-butyltin oxide (3 mol %) in chlorobenzene at 133 ЊC
for 4 h gave a 66% yield of soluble product that was 98% COs
(6). Clearly di-n-butyltin oxide in chlorobenzene was the most
successful catalyst–solvent combination and unless indicated
otherwise this combination was used as standard in all the
subsequent experiments.
Scheme 1
We have previously prepared some simple cyclic oligoesters
by cyclo-depolymerising appropriate polyesters,20,22,23 but this
approach first requires the synthesis of the corresponding
polymers and is clearly less convenient especially if, for
example, a natural product is to be synthesised.
Results and discussion
A range of ω-hydroxy acids, or other ω-substituted acids which
might subsequently be elaborated on the beads into ω-hydroxy
acids, were attached to chloromethylated (1.60 mmol of Cl per
gram) 1% cross-linked polystyrene beads 1 (see Scheme 1 and
Table 1). Except where indicated otherwise this was achieved by
treating the beads with the caesium salt of the hydroxy acid in
N,N-dimethylformamide (DMF) at 50 ЊC for 3 days.24 The load-
ings obtained, assessed by weight gains and the losses of chlor-
ine from the beads, were generally in the range 0.4–1.0 mmol
gϪ1 (see Table 1). Thus, assuming the beads swelled in the
reaction solvent by a factor of 5, the concentrations of the PS
hydroxy acids in the beads at the start of the reactions were
greater than 8 × 10Ϫ2 M. This is above the “effective molarity”
of the various hydroxy acids,5 so it was not expected that under
these initial reaction conditions there would be any evidence of
site isolation. To bring about cyclo-oligomerisation, the PS
ω-hydroxy acids were treated with a TE catalyst in a suitable
solvent (ca. 40 ml of solvent per gram of beads). At the end of
the reaction period the beads were filtered off and the soluble
products recovered and analysed by size exclusion chrom-
atography (SEC), by 1H NMR spectroscopy (mainly for seeking
evidence for the lack of end groups) and, in some cases, by
MALDI-TOF mass spectrometry. The results are summar-
ised in Table 1: note that the SEC traces give the percentage
compositions by weight.
The composition of the soluble product obtained from 11-
hydroxyundecanoic acid 2 under the standard conditions is
summarised in Table 1, entry 1. A typical SEC trace and
MALDI-TOF mass spectrum are shown in Fig. 1. As noted
above the yield was only 66% so, in an attempt to increase the
yield, the beads were retreated using the original conditions. This
produced COs equivalent to a further 25% yield. Repeating the
process a third time gave a further 3% yield, so that overall the
yield was 94%. The compositions of the second and third prod-
ucts are summarised in entries 2 and 3 of Table 1. It will be noted
that the compositions of the soluble products differed consider-
ably. This also proved to be the case in similar experiments with
12-hydroxydodecanoic acid 3 (see Table 1, entries 4 and 5).
Two important conclusions can be drawn from the results
with PS acids 2 and 3. First, it is evident that in each system the
COs 6 and 7 released into solution undergo little or no equili-
bration, otherwise for the experiments with each hydroxy acid
the CO fractions would all have the same composition. The
failure to equilibrate seems surprising given that various poly-
esters in solution undergo cyclodepolymerisation readily with
di-n-butyltin oxide as the catalyst.22,23 It can, however, be
rationalised on the basis of some other results we have obtained
with polymers.26 These strongly suggest that di-n-butyltin oxide
functions as a TE catalyst by first reacting with hydroxy end
groups to give species of the type 8 [see eqn. (1)].
The proportions of the various COs produced were essen-
tially as expected.25 Thus, when the rings are strainless, the
J. Chem. Soc., Perkin Trans. 1, 2002, 629–637
631

View Article Online
gave COs that were equilibrated because the reaction mixtures
were homogeneous and the reaction times and temperature suf-
ficient for the composition of the mixtures of COs to stabilise.22
Thus, it appears that the compositions of the kinetically and
thermodynamically generated mixtures of COs are very similar.
This is not surprising.5 Finally, it will be noted that with all the
four hydroxyalkanoic acids considered in this section, the pro-
portion of cyclised monomer in the soluble products from the
initial cyclisations is <4%. It is low because the cyclic mono-
mers are medium-sized rings. The main products (67–83% by
weight) are the relatively strainless cyclic dimers.
Cyclo-oligomerisations of PS hydroxy acids with secondary
hydroxy groups
All the hydroxy acids considered above had primary hydroxy
groups. Two with secondary hydroxy groups were also investi-
gated. The first was ricinoleic acid [(Z,R)-12-hydroxyoctadec-
9-enoic acid] 13. When the PS acid 13 was treated for 18 h
with di-n-butyltin oxide under the standard conditions cyclo-
oligomerisation occurred to afford COs 14 in 69% yield (see
Table 1, entry 8). Presumably the configuration of the carbon
atom bearing the hydroxy group is retained in these COs.
Although the cyclic dimer was the major product (58%), the
cyclic monomer, which has a 13-membered ring, was formed
in 30% yield. The latter compares with yields from PS 12-
hydroxydodecanoic acid 3 of <4% of dodecanolactone (7; n =
1), which also has a 13-membered ring. The difference arises
because the cis-olefinic linkage in the COs from ricinoleic acid
result in the ring in cyclic monomer 14 being less strained than
that in dodecanolactone 7 (n = 1).
Fig. 1 (a) SEC trace of cyclic undecanoates (6), and (b) MALDI-TOF
MS of the soluble product from the cyclo-oligomerisation of polymer-
supported 11-hydroxyundecanoic acid (2).
(1)
If there are no hydroxy end groups TE is extemely slow. Species
of the type 8 must then react further to achieve TE. Translated
to the present system, it suggests that TE is initiated by the di-n-
butyltin oxide reacting with the hydroxy groups of the bound
hydroxy acids. As the COs have no end groups, once liberated
into the solution surrounding the beads, they can effectively
only take part in further TEs if they re-enter the beads. Appar-
ently they do not do so to any great extent and so equilibration
of the COs is very slow. Consistent with this, cyclo-oligomer-
isations of PS acids 2 and 9 in the same vessel gave a soluble
product whose MALDI-TOF mass spectrum showed peaks due
to both cyclic oligoundecanoates 6 and cyclic oligohexadecano-
ates 10 but very little evidence for cyclic products containing
both types of repeat unit. The second important conclusion
to be drawn is that as the loadings prior to the cyclo-
oligomerisations decrease, the compositions of the CO prod-
ucts indicate that cyclisation becomes more important relative
to oligomerisation, i.e. there is some degree of site isolation at
the lower loadings. This trend is not unexpected.7–10
Cyclo-oligomerisations of PS 8-hydroxyoctanoic acid 4 and
PS 10-hydroxydecanoic acid 5 were carried out similarly to
those of PS acids 2 and 3 except that the reactions were carried
out once for 18 h rather than being carried out more than once
for 4 h. These reactions gave COs 11 and 12 in yields of 67 and
56%, respectively. The compositions of the CO fractions (see
Table 1, entries 6 and 7), like those from PS hydroxyundecanoic
acid 2 and PS 12-hydroxydodecanoic acid 3 at the higher load-
ings (i.e. entries 1 and 6, respectively), are very similar to the
compositions obtained by the cyclo-depolymerisation of the
corresponding polyesters.22 The latter reactions almost certainly
The second hydroxy acid studied, which contains a secondary
hydroxy group, was lithocholic acid 15. The loading achieved
with this acid was only 0.63 mmol gϪ1, but even at this loading
632
J. Chem. Soc., Perkin Trans. 1, 2002, 629–637

View Article Online
Scheme 2
Scheme 3
ca. 22% by weight of the beads is due to the steroid. We have
anoic acid 9 were studied next. PS acid 17 and PS acid 9 were
studied other reactions of this PS acid before.27 Surprisingly in
view of the results obtained with ricinoleic acid 13, treatment of
the PS acid for 18 h with di-n-butyltin oxide under the standard
conditions gave only a small soluble fraction which, by SEC,
was a complex mixture which did not show the pattern expected
for a mixture of COs. Use of the other catalyst systems tried
initially with 11-hydroxyundecanoic acid 2 produced no soluble
products other than, in the case of sodium methoxide, some
methyl lithocholate. It had been hoped the reactions would have
given some of the lithocholic acid COs 16 studied by Sanders’
group.28 These failures may well be due to steric congestion.27
cyclo-oligomerised for 18 h under the standard reaction con-
ditions to give high yields of COs 18 and 10, respectively (see
Table 1, entries 10 and 11). The cyclic monomers from these
acids, which contain 16 and 17 ring atoms, respectively, are
essentially strainless. Accordingly the cyclic monomers were the
main products: 60% and 51% by weight, respectively (corre-
sponding to 76 mol% and 71 mol%, and 59% and 57% overall
yields respectively) of the soluble products. Samples of the
cyclic monomers were isolated by chromatography. Clearly
these are satisfactory methods for the small-scale synthesis of
these macrocyclic lactones.
Attention was next turned to hydroxy acids which could form
even larger rings and which, additionally, might be assembled
on the beads. The first to be studied were the mixed geometrical
isomers of 20-hydroxyicos-10-enoic acid 19 (see Scheme 2).
These were prepared by a cross-metathesis reaction between
Cyclo-oligomerisations of larger PS hydroxy acids: synthesis of
hydroxy acids on the polymer beads
15-Hydroxypentadecanoic acid 17 and 16-hydroxyhexadec-
J. Chem. Soc., Perkin Trans. 1, 2002, 629–637
633

View Article Online
undec-10-enol and methyl undec-10-enoate using Grubbs’ cat-
alyst,29 followed by hydrolysis of the hydroxy ester product 20.
The PS hydroxy acid 19 was subjected to cyclo-oligomerisation
for 18 h under the standard conditions to give COs 21,
then the recovered beads were retreated using the original
cyclo-oligomerisation conditions (see Table 1, entries 12 and
13). The total yield of soluble product was 88%, the cyclic
monomer was the major product in each case (87% and
92% by weight, respectively, corresponding to 92 mol% and
96 mol%). As in the other cases discussed above, relative to
the first reaction, the second reaction showed some evidence
for site isolation.7–10 A sample of the mixed geometrical
isomers of cyclic monomer 22 was isolated using column
chromatography.
In order to assemble the hydroxy acid 19 on the beads, undec-
10-enoic acid 23 was bound to the beads using the usual pro-
cedure (see Table 1, entry 14, and Scheme 2). The PS acid was
then reacted with a large excess of undec-10-enol in the pres-
ence of Grubbs’ olefin metathesis catalyst.29 Any “dimeric”
product 24 formed by reaction between two undec-10-enol
molecules would be washed from the beads after the meta-
thesis reaction. The PS products were subjected to cyclo-
oligomerisation for 18 h using the standard conditions. Any PS
acid 23 residues that reacted but only with each other or did not
react at all would remain attached to the beads during the TE
reaction. Thus, the only molecules to be produced in solution
during the cyclo-oligomerisation would be the desired COs 21.
Analysis of the soluble product from the cyclo-oligomerisation
confirmed this. The yield of COs 21 was 61% and the com-
position of the product (entry 14) was very similar to that
summarised in entry 12 of Table 1.
A second hydroxy acid that could be assembled on the beads
is acid 25 (see Scheme 3). This acid was first prepared using a
literature procedure,30 and was attached to the beads using
the procedure first reported by Merrifield,31 i.e. treating the
chloromethylated resin beads with the acid and triethylamine
in ethyl acetate. The PS acid 25 was subjected to cyclo-
oligomerisation for 18 h using the usual procedures (see Table 1
entry 15). The reaction took place cleanly to give a very high
yield of soluble product 26, 74% of which was the cyclised
monomer 27. A sample of the cyclic monomer 27 was isolated.
Approximately 5% of the soluble product was a series of COs
28 formed by an ester-amide exchange reaction. The COs 26
have been prepared before by a related PS system where the PS
amino acid 29 was cyclo-oligomerised through reactions of
ester linkages with amines.18
coupled with 11-hydroxyundecanoic acid 2 using dicyclohexyl-
carbodiimide.31 When this product was subjected to cyclo-
oligomerisation, using the standard procedure, COs 26 were
formed in similar proportions to previously but in a somewhat
lower yield than before (see entry 16 of Table 1).
Conclusions
Attachment of ω-hydroxy acids 2–5, 9, and 17, all bearing
primary hydroxy groups, to Merrifield beads followed by treat-
ment with a catalytic amount of di-n-butyltin oxide in chloro-
benzene at 133 ЊC for between 4 and 18 h brought about the
formation of the corresponding COs as the main (>92%)
soluble products in yields of 56–83%. Similar results were
obtained with ricinoleic acid 13, which has a secondary hydroxy
group, but attempts to carry out analogous reactions with
lithocholic acid 15 failed. The hydroxy acids 19 and 25 also
reacted well to give COs and samples of the cyclic monomers 22
and 27 were isolated. The hydroxy acids 19 and 25 were also
successfully assembled on the beads.
There are several possible applications for COs synthesised
by the above approach. Thus, they might be used for the prep-
aration of soluble combinatorial libraries of macrocycles,19 or
as starting materials for the small-scale combinatorial synthesis
of a range of copolyesters.21 For both of these applications the
fact that the PS syntheses affords families of COs is not a prob-
lem because these applications involve the establishment of
equilibria into which all the COs can feed. Thirdly, the reactions
may provide a useful approach to the small-scale synthesis of
macrocyclic lactones with 15 or more ring atoms, since with
these ring sizes the cyclised monomer will be the main soluble
product.
Experimental
Unless indicated otherwise chemicals were purchased from
Aldrich, Sigma or Lancaster and were used without further
purification. Organic extracts were dried with magnesium
sulfate. Solid samples were dried in a vacuum oven at 1.0 mmHg
and the temperatures indicated. Melting points were deter-
mined using an Electrothermal apparatus and are uncorrected.
Infrared spectra were recorded using a Perkin Elmer 1720
instrument; unless indicated otherwise solid samples were pre-
pared as potassium bromide discs and liquid samples as thin
1
films between sodium chloride plates. H NMR spectra were
recorded for solutions in deuteriated chloroform on a Varian
Gemini 200 MHz NMR instrument using TMS as an internal
standard. Chemical shifts (δ) are quoted in ppm. relative to
tetramethylsilane. 13C NMR spectra were recorded for solutions
in deuteriated chloroform on a Bruker AC 300e (300 MHz)
instrument. MALDI mass spectrometry was carried out on
a Kratos Kompact 3 using dihydroxybenzoic acid as matrix
and applying the samples to the plate as a solution in chloro-
form. Elemental analyses were carried out by Butterworth
Laboratories, Teddington, Middlesex. SEC analyses were
carried out for solutions in chloroform using an instrument
equipped with a Gilson 307 pump operating at a flow rate of
0.3 ml minϪ1 through four Polymer Labs 3µ Mixed E columns
in tandem with a GBC LC 1240 differential refractometer for
detection.
Source of various hydroxy acids
Hydroxy acids 3, 4, 9, 13 and 15 were commercial samples.
10-Hydroxydecanoic acid 5 was available in the laboratory
from earlier studies.22 15-Hydroxypentadecanoic acid 17 was
prepared by the hydrolysis of a commercial sample of penta-
decanolactone, and N-(11-hydroxyundecanoyl)-11-amino-
undecanoic acid 25 was prepared according to a literature
procedure.30
Finally, hydroxy acid 25 was synthesised on the beads (see
Scheme 3) by first attaching the 11-(tert-butyloxycarbonyl-
amino)undecanoic acid 30 to the beads using the Merrifield
procedure31 (see Table 1, entry 16). The tBoc group was
removed by treatment with trifluoroacetic acid and the product
634
J. Chem. Soc., Perkin Trans. 1, 2002, 629–637

View Article Online
Synthesis of 11-hydroxyundecanoic acid (2)
to dryness and the residual water removed azeotropically with
toluene. The solid was partially dissolved in DMF (30 ml) and
Merrifield resin (7.52 g, 1.6 mmol of chloromethyl groups per
gram, 1% cross-linked) was added. The suspension was heated
at 50 ЊC for 3 days with stirring. The resin was then collected by
filtration and washed successively with DMF, DMF–water
(1 : 1), ethanol and acetone before being dried at 40 ЊC in vacuo.
The product (9.30 g, a weight gain of 1.78 g, corresponding to
10.76 mmol of acid, equal to 1.16 mmol of acid per gram) had
11-Bromoundecanoic acid (50.0 g) was dissolved in chloroform
(50 ml). Tetra-n-butylammonium hydroxide (40 wt% aqueous
solution) was added to achieve neutrality (phenolphthalein as
indicator). The two-phase mixture was stirred rapidly and
heated under reflux for 36 h. The mixture was then cooled and
poured into acidified methanol (50 ml concentrated sulfuric
acid in 500 ml methanol). The polyundecanoate precipitated
out, was collected, dissolved in the minimum amount of chloro-
form and re-precipitated from methanol (1 l). The white powder
was filtered off and dried at 40 ЊC under vacuum (32.5 g). The
polymer was hydrolysed by being stirred with aqueous potas-
sium hydroxide (20 g in 500 ml) for 18 h at reflux temperature.
The clear solution was cooled and acidified carefully with con-
centrated sulfuric acid. The white flakes that formed were
filtered off and dried at 40 ЊC under vacuum. Recrystallisation
from ethyl acetate–petroleum ether gave colourless needles
(28.0 g, 73%), m.p. 63–65 ЊC (lit.,32 68 ЊC); νmax(film) 3500–3300
νmax(KBr) 1734 cmϪ1 (C᎐O, ester). Elemental analysis for chlor-
᎐
ine gave 0.48%, which corresponds to a loading of 1.13 mmol
of acid per gram of resin.
(b) Table 1, entry 9. The following procedure27 is typical of
that used for the experiments summarised in Table 1, entries 9
and 15.
Lithocholic acid 15 (752 mg, 2.00 mmol), triethylamine
(200 mg, 2.00 mmol) and ethyl acetate (120 ml) were stirred
together at 20 ЊC for 30 min. Merrifield resin (1.25 g, 1.6 mmol
gϪ1, 1% cross-linked) was added and the mixture heated under
reflux for 90 h. The beads were filtered off and thoroughly
washed successively with ethyl acetate, ethanol and dichloro-
methane. The dried product (1.59 g) had νmax(KBr) 1726 cmϪ1
(br, OH), 1714 cmϪ1 (C᎐O, acid); δ (CDCl ): 5.0–4.2 (br s, 2H,
᎐
H
3
OH), 3.62 (t, J = 6.5 Hz, 2H, CH2OH), 2.33 (t, J = 6.5 Hz, 2H,
CH2CO2), 1.55 (quint, J = 6.5 Hz, 2H, CH2CH2OH), 1.34 ppm
(br m, 14H, 7 CH2); m/z (CI) 220 [M ϩ (NH4)ϩ], 203 [(Mϩ1)ϩ],
185 [(M Ϫ OH)ϩ].
(C᎐O, ester). The increase in weight of the beads corresponds
᎐
to a loading of 0.63 mmol gϪ1 of steroid.
Synthesis of 20-hydroxyicos-10-enoic acid 19
(a) Synthesis of methyl 20-hydroxyicos-10-enoate 20. A mix-
ture of Grubbs’ catalyst [bis(tricyclohexylphosphine)benzyl-
ideneruthenium dichloride purchased from Strem Chemicals]
(46 mg, 0.057 mmol), methyl undec-10-enoate (1.01 g, 5.43
mmol), undec-10-enol (0.95 g, 5.59 mmol) and dichlorometh-
ane (50 ml) was stirred at 20 ЊC under argon for 18 h. The
solvent was then evaporated off and the residue subjected to
column chromatography over silica gel with petroleum ether–
ethyl acetate (7 : 13 v/v) as the eluant. This gave methyl 20-
hydroxyicos-10-enoate (443 mg, 24%) as an oil with νmax(KBr)
General procedure for cyclo-oligomerisation of PS hydroxy acids
The following are typical of the procedures used to achieve
cyclo-oligomerisation. The results summarised in Table 1 were
achieved using procedure (iii) with the reaction times indicated
in the table.
(i) Using potassium methoxide. PS 11-hydroxyundecanoic
acid 2 (0.97 g, 1.12 mmol) was suspended in toluene (40 ml)
under a nitrogen atmosphere. Potassium methoxide (2.28 mg,
0.33 mmol) was added and the suspension heated under reflux
for 4 h. The beads were then filtered off and washed with
chloroform. The combined filtrate and washings were evapor-
ated to dryness. This left a colourless solid (67.0 mg, 33%)
which had νmax(KBr) 1734 cmϪ1. SEC analysis, using authentic
samples of cyclic and linear oligomers from previous studies,22
indicated the presence of methyl 11-hydroxyundecanoate (19%)
and both cyclic (65%) and linear oligomers (16%).
1
3600–3200 (br, OH), 1738 cmϪ1 (ester carbonyl); H NMR
δ(CDCl ) 5.35 (m, 2H, –CH᎐CH–), 3.61 (t, J = 6.5 Hz, 2H,
᎐
3
CH2OH), 3.65 (s, 3H, OCH3), 2.28 (t, J = 6.5 Hz, 2H,
CH COO), 1.95 (m, 4H, CH –CH᎐CH–CH ) 1.57 (m, 6H,
᎐
2
2
2
3 CH2) and 1.31 ppm (br s, 20H, 10 CH2); 13C NMR δ(CDCl3)
173.7 (C᎐O), 130.6 (trans CH᎐CH), 129.6 (cis CH᎐CH) (ratio
᎐
᎐
᎐
of trans and cis signals 0.80 : 0.20), 63.5 (CH3OCO), 51.9
(CH2OH), 34.6, 33.3, 33.2, 33.0, 30.1, 29.9, 29.8, 29.7, 29.6,
29.5, 26.2, 25.4 and 25.4 ppm; m/z (CI) 341 [(M ϩ 1)ϩ].
(ii) Using lithium bis(trimethylsilyl)amide. PS 11-hydroxy-
undecanoic acid 2 (1.50 g, 1.74 mmol) was suspended in
chlorobenzene (50 ml) under a nitrogen atmosphere. Lithium
bis(trimethylsilyl)amide (15 mg, 0.09 mmol) was added and the
suspension heated under reflux for 4 h. The beads were then
filtered off and washed with chloroform. The combined filtrate
and washings were evaporated to dryness. This left a colourless
solid (183 mg, 57%), which had νmax(KBr) 1734 cmϪ1; SEC
analysis indicated the presence of both cyclic (74%) and linear
oligomers (26%).
(b) Hydrolysis of the geometrical isomers of methyl 20-
hydroxyicos-10-enoate (20). The above ester (274 mg, 0.81
mmol) was suspended in 2 M aqueous sodium hydroxide
(20 ml) and the mixture was heated under reflux for 4 h. The
mixture, which was then homogeneous, was acidified with 2 M
hydrochloric acid and extracted with ethyl acetate. The solvent
was evaporated from the dried extract. The residue (244 mg,
93%) was a pale yellow solid, mp 65–70 ЊC, which had
νmax(KBr) 1702 cmϪ1 (acid carbonyl); δH(CDCl3) 5.45 (m,
2H –CH᎐CH–), 3.65 (t, J = 6.5 Hz, 2H, CH OH), 2.30 (t, J =
᎐
2
(iii) Using dibutyltin oxide. (a) PS 11-hydroxyundecanoic acid
(1.23 g, 1.41 mmol gϪ1) was suspended in chlorobenzene
(50 ml). Dibutyltin oxide (11.0 mg, 0.042 mmol) was added and
the mixture was heated under reflux for 4 h. The resin was
recovered and washed with chloroform. The combined filtrate
and washings were evaporated to give a white solid (171 mg,
0.93 mmol, 66%). The recovered resin had νmax(KBr) 1734 cmϪ1
(ester carbonyl). The white solid had νmax(KBr) 1730 cmϪ1 (ester
carbonyl); δH(CDCl3) 4.08 (t, J = 6.5 Hz, 2H, CH2OCO), 2.31
(t, J = 6.5 Hz, 2H, CH2CO2), 1.65 (br m, 2H, CH2) and 1.34
ppm (s, 14 H, 7 × CH2); MALDI-TOF MS showed peaks
due to COs from the cyclic dimer upto the cyclic undecamer, see
Fig. 1; SEC showed cyclic products from the monomer upto the
nonamer.
6.5 Hz, 2H, CH COO), 2.00 (m, 4H, CH –CH᎐CH–CH ) 1.57
᎐
2
2
2
(m, 6H, 3 CH2) and 1.31 ppm (br s, 20H, 10 CH2); m/z (CI) 344
[(M ϩ NH4)ϩ] and 327 [(M ϩ 1)ϩ].
General procedures for loading hydroxy acids onto Merrifield
resin
(a) Table 1, entry 1. The following procedure24 is typical of
that used in the experiments summarised in Table 1, entries 1, 4,
6–8, 10–12, 14 and 16.
11-Hydroxyundecanoic acid 2 (3.02 g, 14.95 mmol) was dis-
solved in 95% ethanol (20 ml) and water (5 ml) was added. The
solution was titrated to neutral (pH meter) using aqueous
cesium carbonate solution. The solution was then evaporated
J. Chem. Soc., Perkin Trans. 1, 2002, 629–637
635

View Article Online
(b) The recovered resin from experiment (a) was resuspended
in chlorobenzene (50 ml) and dibutyltin oxide (5.0 mg) was
added. The reaction was heated again under reflux for 4 h and
the work-up procedure conducted as before yielding a further
batch of colourless solid (65 mg, 0.35 mmol, 25%) which was
analysed as above.
(c) The recovered resin from experiment (b) was resuspended
in chlorobenzene (50 ml) and dibutyltin oxide (3.0 mg) was
added. The reaction was again heated under reflux for 4 h and
the work-up procedure conducted as before yielding a further
batch of colourless solid (8 mg, 0.042 mmol, 3%) which was
analysed as above.
Isolation of cyclic monomers 18, 10, 22 and 27
The experiments were carried out as indicated in Table 1. In
each case the crude soluble product was subjected to column
chromatography over silica gel with hexane–ether (first two and
final entries) or petroleum ether–ethyl acetate (third entry) as
the eluant. This afforded samples of the pure cyclic oligomers.
Cyclic monomer 18 (n = 1) had mp 33–34 ЊC (lit.,34 35–37 ЊC);
νmax(KBr) 1734 cmϪ1 (C᎐O, ester); δ (CDCl ) 4.10 (t, J =
᎐
H
3
6.5 Hz, 2H, CH2OCO), 2.33 (t, J = 6.5 Hz, 2H, CH2OCO), 1.67
(br m, 2H, CH2) and 1.32 ppm (s, 22H, CH2).
Cyclic monomer 10 (n = 1) had mp 30–33 ЊC (lit.,35 33–34 ЊC);
νmax(KBr) 1735 cmϪ1 (C᎐O, ester); δ (CDCl ) 4.05 (t, J =
᎐
H
3
6.5 Hz, 2H, CH2OCO), 2.30 (t, J = 6.5 Hz, 2H, CH2OCO), 1.67
(br m, 2H, CH2) and 1.36 ppm (s, 24H, CH2).
On-bead synthesis of PS 20-hydroxyicos-10-enoic acid (19)
Cyclic monomer 22 was obtained as a clear oil, νmax(KBr)
1737 cmϪ1 (C᎐O, ester); 1H NMR δ(CDCl ) 5.35 (m, 2H, –CH᎐
CH–), 4.16 (t, J = 6.5 Hz, 2H, CH2OCO), 2.30 (t, J = 6.5 Hz,
(a) Loading of undec-10-enoic acid (25) onto Merrifield beads.
Undec-10-enoic acid 25 was attached to the Merrifield beads
using the procedure described above for attaching hydroxy acid
2. The loading achieved was 0.61 mmol gϪ1.
᎐
᎐
3
2H, CH COO), 2.00 (br m, 4H, CH –CH᎐CH–CH ), 1.63 (m,
᎐
2
2
2
6H, 3 CH2) and 1.31 ppm (s, 20H, 10 × CH2); 13C NMR
δ(CDCl ) 176.3 (C᎐O), 131.4 and 131.1 (trans CH᎐CH), 130.6
᎐
᎐
3
(b) Synthesis of PS hydroxy acid 19. A mixture of the above
PS undec-10-enoic acid (7.80 g, 4.75 mmol), undec-10-enol
(5.65 g, 33 mmol), Grubb’s catalyst (40 mg) and dichlorometh-
ane (100 ml) was stirred at 20 ЊC under nitrogen for 48 h. The
beads were then filtered off, thoroughly washed with dichloro-
methane, and dried at 40 ЊC under vacuum. The beads (7.85 g)
of PS hydroxy acid 19 had νmax(KBr) 1736 cmϪ1. A small sample
of the acid was cleaved from the beads using sodium meth-
oxide in methanol. The amount of acid 19 recovered, identi-
(cis CH᎐CH) (ratio of trans and cis signals 0.80 : 0.20), 64.5
᎐
(CH OCO), 35.0 (CH CO), 32.5, 32.2 (CH CH᎐CHCH ),
᎐
2
2
2
2
29.9, 29.8, 29.7, 29.6, 29.5, 29.5, 29.3, 29.2, 29.1, 28.9, 28.8,
28.5, 26.3 and 25.7 ppm; m/z (CI) 326 [(M ϩ NH4)ϩ] and 309
[(M ϩ 1)ϩ]; one peak on SEC; accurate MS found 308.2720,
calculated for C20H36O2 308.2715.
Cyclic monomer 27 was obtained as crystals, mp 98–100 ЊC,
νmax(KBr) 1729 and 1644 cmϪ1 (ester and amide carbonyls);
δH(CDCl3) 5.35 (br s, 1H, NH), 4.03 (t, J = 6.5 Hz, 2H,
CH2OCO), 3.20 (m, 2H, CH2NHCO), 2.25 (t, J = 6.5 Hz, 2H,
CH2COO), 2.10 (t, J = 6.5 Hz, 2H, CH2CONH) and 1.40 ppm
(m, 32H, 16 CH2); m/z (CI) 368 [(M ϩ 1)ϩ]; C22H41O3N requires
C, 71.93, H, 11.17 and N, 3.81%, found C, 71.55, H, 11.05 and
N, 3.20%.
1
fied by H NMR spectroscopy, corresponded to a loading of
0.21 mmol gϪ1.
On-bead synthesis of PS N-(11-hydroxyundecanoyl)-11-amino-
undecanoic acid (25)
(a) Attachment of acid 30. 11-(tert-Butyloxycarbonylamino)-
undecanoic acid 30 was synthesised as described previously33
and attached to the Merrifield beads using the procedure
described above for attaching 11-hydroxyundecanoic acid 2.
The loading achieved was 1.17 mmol gϪ1.
Acknowledgements
We thank the EPSRC for financial support (Grant GR/K/
73305) and James Logan, Robert Thomas and Michael Doward
for experimental assistance.
(b) Deprotection of PS acid 30. The PS acid 30 (3.97 g,
4.66 mmol), prepared in (a) above, was stirred in a mixture of
dichloromethane and trifluoroacetic acid (1 : 1 v/v, 40 ml) at
20 ЊC for 4 h. The beads were then collected by filtration and
washed thoroughly with dichloromethane. The beads were
stirred with a mixture of diisopropylethylamine and dichloro-
methane (1 : 3 v/v, 40 ml) at 20 ЊC for 1 h. The beads were
collected by filtration and washed with dichloromethane. The
treatment with base was then repeated. Finally the beads were
filtered off and washed successively with dichloromethane,
methanol and ether. The dried beads (4.12 g) had νmax
(KBr) 1730 cmϪ1 and, by elemental analysis, N = 1.50%.
The latter corresponds to a loading of 1.07 mmol gϪ1 of
11-aminoundecanoic acid.
References
1 C. J. Roxburgh, Tetrahedron, 1995, 51, 9767.
2 W. H. Kruizinga and R. M. Kellogg, J. Am. Chem. Soc., 1981, 103,
5183 and references cited therein.
3 H. A. Kirst, Prog. Med. Chem., 1993, 30, 57.
4 M. Dobler, in Comprehensive Supramolecular Chemistry, eds. J. L.
Atwood, J. E. D. Davies, D. D. MacNicol and F. Vögtle, Pergamon,
Oxford, 1992, vol. 1, ch. 5, pp. 267–313.
5 G. Illuminati and L. Mandolini, Acc. Chem. Res., 1981, 14, 95.
6 J. I. Crowley and H. Rapoport, Acc. Chem. Res., 1976, 9, 135.
7 P. Hodge and E. Khoshdel, React. Polym., 1985, 3, 143.
8 P. Hodge and J. Waterhouse, J. Chem. Soc., Perkin Trans. 1, 1983,
2319.
9 Polymeric Reagents and Catalysts, ed. W. T. Ford, ACS Symposium
Series 308, Washington DC, 1986.
(c) Coupling of 11-hydroxyundecanoic acid (2) to PS 11-
aminoundecanoic acid. The total product from the experi-
ment described in (b) above was swollen in dichloromethane
(30 ml). Dicyclohexylcarbodiimide (0.40 g, 3.20 mmol) was
added and the mixture cooled to 0 ЊC. A solution of 11-
hydroxyundecanoic acid 2 (0.62 g, 3.05 mmol) in dichloro-
methane (10 ml) was added dropwise to the stirred mixture
over 30 min, then the mixture was stirred for 48 h. The beads
were collected by filtration, washed successively and extensively
with dichloromethane, methanol, ether, and ether. The dried
beads (2.58 g) had νmax(KBr) 1734 and 1655 cmϪ1, no bands
due to amine groups, and, by elemental analysis, N = 1.15%.
The latter corresponds to PS hydroxy acid 26 with a loading
of 0.82 mmol gϪ1.
10 P. Hodge, Chem. Soc. Rev., 1997, 26, 417.
11 L. T. Scott and J. O. Naples, Synthesis, 1976, 738.
12 S. L. Regen and Y. Kimura, J. Am. Chem. Soc., 1982, 104, 2064.
13 R. A. Amos, R. W. Emblidge and N. Havens, J. Org. Chem., 1983,
48, 3598.
14 B. M. Trost and R. W. Warner, J. Am. Chem. Soc., 1983, 105, 6940.
15 S. Mohanraj and W. T. Ford, J. Org. Chem., 1985, 50, 1616.
16 M. Tomoi, T. Watanabe, T. Suzuki and H. Kakiuchi, Makromol.
Chem., 1985, 186, 2473.
17 (a) P. Hodge, J.-L. Jiang, G. J. Owen and M. P. Houghton, Polymer,
1996, 37, 5059; (b) P. Hodge, M. P. Houghton and M. S. K. Lee,
J. Chem. Soc., Chem. Commun., 1993, 581.
18 P. Hodge and P. Peng, Polymer, 1999, 40, 1871.
19 P. Monvisade, P. Hodge and C. L. Ruddick, Chem. Commun., 1999,
1987.
20 P. Hodge, React. Funct. Polym., 2001, 48, 15.
636
J. Chem. Soc., Perkin Trans. 1, 2002, 629–637

View Article Online
21 P. Hodge and S. D. Kamau, unpublished results.
22 C. L. Ruddick, P. Hodge, Z. Yang, R. L. Beddoes and M. Helliwell,
J. Mater. Chem., 1999, 9, 2399.
23 P. Hodge, Z. Yang, A. Ben-Haida and C. S. McGrail, J. Mater.
Chem., 2000, 10, 1533.
24 B. F. Gisin, Helv. Chim. Acta, 1973, 56, 1476.
25 H. Jacobson and W. H. Stockmayer, J. Chem. Phys., 1950, 18, 1600.
26 S. Dad, A. J. Hall, P. Hodge, R. J. Kell and C. L. Ruddick,
unpublished results.
27 P. Hodge, J. Kemp, E. Khoshdel and G. M. Perry, React. Polym.,
1985, 3, 299.
28 P. A. Brady, R. P. Bonar-Law, S. J. Rowan, C. J. Suckling and
J. K. M. Sanders, Chem. Commun., 1996, 319.
29 R. H. Grubbs, S. J. Miller and G. C. Fu, Acc. Chem. Res., 1995, 28,
446.
30 B. Jasse, Bull. Chim. Soc. Fr., 1969, 3, 953.
31 R. B. Merrifield, J. Am. Chem. Soc., 1963, 85, 2149.
32 K. B. Wiberg and R. F. Waldron, J. Am. Chem. Soc., 1991, 113,
7697.
33 C. Galli and L. Mandolini, Org. Synth., 1978, 58, 98.
34 H. H. Mathur and S. C. Bhattacharyya, J. Chem. Soc., 1963, 3505.
35 P. Peng and P. Hodge, Polymer, 1998, 39, 981.
J. Chem. Soc., Perkin Trans. 1, 2002, 629–637
637