Design, synthesis, conformational analysis and nucleic acid hybridisation properties of thymidyl pyrrolidine-amide oligonucleotide mimics (POM)

View Article Online / Journal Homepage / Table of Contents for this issue

Design, synthesis, conformational analysis and nucleic acid

hybridisation properties of thymidyl pyrrolidine-amide

oligonucleotide mimics (POM)†

David T. Hickman,^aT. H. Samuel Tan,^aJordi Morral,^aPaul M. King,^bMatthew A. Cooper^c

and Jason Mickleﬁeld*^a

^aDepartment of Chemistry, University of Manchester Institute of Science and Technology,

Faraday Building, PO Box 88, Manchester, UK M60 1QD

^bDepartment of Chemistry, Birkbeck College, University of London, 29 Gordon Square,

London, UK WC1H 0PP

^cUniversity Chemical Laboratory, Lensﬁeld Road, Cambridge, UK CB2 1EW

Received 30th May 2003, Accepted 6th August 2003

First published as an Advance Article on the web 28th August 2003

Pyrrolidine-amide oligonucleotide mimics (POM) 1 were designed to be stereochemically and conformationally

similar to natural nucleic acids, but with an oppositely charged, cationic backbone. Molecular modelling reveals

that the lowest energy conformation of a thymidyl-POM monomer is similar to the conformation adopted by

ribonucleosides. An eﬃcient solution phase synthesis of the thymidyl POM oligomers has been developed, using

both N-alkylation and acylation coupling strategies. ¹H NMR spectroscopy conﬁrmed that the highly water soluble

thymidyl-dimer, T₂-POM, preferentially adopts both a conﬁguration about the pyrrolidine N-atom and an overall

conformation in D₂O that are very similar to a typical C3Ј-endo nucleotide in RNA. In addition the nucleic acid

hybridisation properties of a thymidyl-pentamer, T₅-POM, with an N-terminal phthalimide group were evaluated

using both UV spectroscopy and surface plasmon resonance (SPR). It was found that T₅-POM exhibits very high

aﬃnity for complementary ssDNA and RNA, similar to that of a T₅-PNA oligomer. SPR experiments also showed

that T₅-POM binds with high sequence ﬁdelity to ssDNA under near physiological conditions. In addition, it was

found possible to attenuate the binding aﬃnity of T₅-POM to ssDNA and RNA by varying both the ionic strength

and pH. However, the most striking feature exhibited by T₅-POM is an unprecedented kinetic binding selectivity for

ssRNA over DNA.

Introduction

Modiﬁed nucleic acids and oligonucleotide mimics are of con-

siderable interest as agents for down-regulating gene expression

through hybridisation either with target mRNA (antisense) or

dsDNA (antigene).^1–6This approach has not only led to new

therapeutic treatments, but is widely used in target validation to

identify new therapeutic targets. In addition, modiﬁed oligo-

nucleotides are of value as diagnostic and bio-analytical probes,

or as tools in molecular biology.^6–11These molecules are also of

fundamental importance as models to explore the structure–

function relationship of nucleic acids.^12,13In particular, by

studying the biophysical properties of chemically modiﬁed

nucleic acids, it is argued that we can gain insight as to why

DNA and RNA were selected as the ubiquitous genetic

Fig. 1 Structure of POM and other related oligonucleotide mimics.

materials on this planet.¹⁴

B = nucleobase (A, T, C or G).

Despite the large number of backbone modiﬁcations that

have been developed, surprisingly few confer enhanced recog-

the best of all neutral linkages, for example amide 3¹⁹or 3Ј-N-

sulfamate modiﬁed DNA,²⁰result in only modest increases in

nition properties, aﬃnity and sequence selectivity, for DNA and

RNA relative to native nucleic acids.^1,15A general strategy to

aﬃnity for complementary nucleic acids. Of all the neutral

enhance aﬃnity has been to develop modiﬁed oligonucleotides

that are better pre-organised by restricting the conformational

freedom of the backbone.^1,16For example, sugar modiﬁcations

which retain the native phosphodiester linkage, such as LNA

2 (Fig. 1),^17,18have been used to particularly good eﬀect.^5,6,11

Several groups have also replaced the phosphodiester linkage in

nucleic acids with alternative neutral linkages based on the

premise of favourable enthalpy changes resulting from dimin-

ished electrostatic repulsion between strands.¹However even

mimics that have been developed, probably PNA 4, where the

whole backbone is replaced by a pseudo peptide structure,^21–23

is the most eﬀective and widely used.^7–10Nevertheless PNA suf-

fers from poor aqueous solubility and a tendency to aggregate.

In addition PNA is achiral and can bind in both a parallel and

antiparallel fashion with nucleic acid, which can jeopardise

sequence speciﬁcity.

The introduction of positive charges into the backbone of

oligonucleotide mimics is an alternative approach for increasing

the aﬃnity for complementary DNA and RNA. One of the

earliest examples reported was the deoxynucleic guanidine

(DNG) 5 oligomers, where the phosphodiester linkages in

DNA are replaced by cationic guanidinium groups.²⁴This

† Electronic supplementary information (ESI) available: experimental

details (16 Figures, 3 Tables). See http://ww w .rsc.org/suppdata/ob/b3/

b306156f/

T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 3

O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 2 7 7 – 3 2 9 2

3277

View Article Online

modiﬁcation was shown to vastly increase aﬃnity for comple-

mentary ssDNA and RNA. However, it remains to be seen if

such strong electrostatic attraction between oppositely charged

backbones results in non-speciﬁc binding, via salt bridges,

which may not necessitate base pairing and therefore may

not be evident from standard UV thermal denaturation

experiments.

As part of current research in our laboratory into the devel-

opment of backbone modiﬁed oligonucleotides, we recently

introduced the novel cationic pyrrolidine-amide oligonucleotide

mimic (POM).²⁵In this paper we describe the design rationale,

solution phase synthesis and conformational analysis of thym-

idyl T₂-POM. Finally, the nucleic acid hybridisation properties

of a phthalimide capped T₅-POM are discussed in detail.

Results and discussion

Fig. 2 X-ray crystal structures of pyrrolidine derivative 6²⁹and

uridine 7.³²

Design and molecular modelling

Pyrrolidine-amide Oligonucleotide Mimics (POM) 1 (Fig. 1)

are derived by replacing the furanose ring and phosphodiester

backbone in DNA with a (2R,4R)-conﬁgured pyrrolidine ring

and amide linkage, respectively. The pyrrolidine ring was

chosen because it will be substantially protonated at physio-

logical pH. Due to nearest neighbour interactions^26,27it is anti-

cipated that only every other pyrrolidine would be protonated

at physiological pH. Nevertheless oligomers will possess high

water solubility and exhibit electrostatic attraction for comple-

mentary nucleic acids, possibly increasing aﬃnity. In this

respect it is similar to DNG 5.²⁴It was also envisaged that the

binding aﬃnity to DNA/RNA might be attenuated by changing

the ionic strength or pH, which may be an attractive property if

these molecules are to be used as probes, for example in DNA-

microarray type devices. In addition there is some evidence to

suggest that positively charged peptides and peptidomimetics

(like POM) could bind through electrostatic attraction to phos-

phate groups of the cell wall, which can aid transport through

the cell membranes.²⁸The amide linkage was selected to con-

nect the pyrrolidine units, since this group has been shown to be

a viable replacement for phosphodiester in the amide-linked

DNA mimics.¹⁹Furthermore the amide linkage should facili-

tate the synthesis of POM using peptide chemistry and also

allow easy conjugation with peptides, in a similar fashion to

PNA.

designed as conformationally restricted analogues of PNA,

unlike POM which we envisage shares little conformational or

electrostatic similarity with PNA.

In order to determine the relative energies of the diﬀerent

conﬁgurations and conformations of the pyrrolidine ring in

POM, semi-empirical quantum mechanical calculations⁴¹were

performed on the model thymidyl-pyrrolidine cis- and trans-

dimethyl diastereoisomers 8 and 9 (Fig. 3). Each conformer,

generated for every 10 degrees of pseudorotation, was ﬁrst

energy minimised allowing free rotation of the thymine base

about the C4Ј-N1 bond in order to optimise the pseudo-glyco-

sidic torsion angle (χ). The standard enthalpies of formation

for each conformer were then calculated [see Fig. 3 and ESI

Table 1]. These reveal that, as predicted, the conformers gener-

ated for the trans-diastereoisomer 9 are on the whole lower

in energy than the cis. Moreover the lowest energy trans-

conformer A has a phase angle of 48Њ which is similar to the

crystal structure of pyrrolidine 6 and also just outside that

region of conformational space occupied by C3Ј-endo ribose

typical for RNA.³¹The trans-conformer B (P = 198Њ) is 0.57

kcal mol^Ϫ1higher in energy and analogous to a C2Ј-endo deoxy-

ribose typically highly populated in DNA structures.

The (2R,4R) stereochemistry of the pyrrolidine ring was

carefully selected because this should ensure that the

N-acetamido group will adopt the sterically less demanding

trans relative conﬁguration making the system a stereochemical

match with native nucleic acids. Of course the pyrrolidine

N-atom can invert conﬁguration via the free amine at a rate that

will depend on pH. However, X-ray crystallographic data²⁹for

a protonated pyrrolidine ring 6, which is stereochemically iden-

tical and both electronically and sterically similar to the

pyrrolidine ring in POM, clearly show that the N-alkyl substi-

tuent adopts the trans relative stereochemistry, at least in the

crystalline state (Fig. 2). In addition the overall conformation

of 6, which is described by a pseudorotation phase angle (P)^30,31

of 43.7Њ and a maximum torsional angle (ν_max) of 40.7Њ, is simi-

lar to uridine 7 in the crystalline state (P = 14Њ and ν_max= 42Њ).³²

In fact pyrrolidine 6 is only a few degrees of pseudorotation

out from the preferred C3Ј-endo conformation of ribose in

RNA which typically falls within the range P = 0

36Њ.³¹As

such it was envisaged that the pyrrolidine unit of POM would

bear a closer conformational resemblance to native RNA and

more conservative mimics such as amide linked DNA 3 and

LNA 2 which similarly adopt a C3Ј-endo sugar-conformation.

As a result of these critical stereochemical and conform-

ational features POM is predicted to bind in an antiparallel

fashion and diﬀers conceptually from other oligonucleotide

systems incorporating pyrrolidine units.^33–40These systems were

Fig. 3 Variation in the standard enthalpies of formation (∆H Њ) with

pseudorotation phase angles (P) for model cis- and trans-dimethyl

pyrrolidines 8 and 9.

Whilst these ﬁndings are encouraging and suggest the most

highly populated conﬁguration and conformation will match

closely the RNA ribose structure, it should be noted these

O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 2 7 7 – 3 2 9 2

3278

View Article Online

studies do not account for intramolecular forces that may arise

in longer oligomers. For example, one could envisage a possible

H-bond extending from the protonated pyrrolidine N-atom to

an adjacent O-atom of the amide linkage in oligomeric POM 1.

Nevertheless intrastrand electrostatic repulsion should favour

the formation of an extended linear structure, with trans-

relative stereochemistry, and disfavour the formation of tightly

folded secondary structures, which might jeopardise base

pairing with complementary DNA and RNA.

Synthesis of a thymidyl-POM dimer

In order to investigate the conformational properties of POM,

using ¹H NMR spectroscopy, the synthesis of a thymidyl-dimer,

T₂-POM, was embarked upon ﬁrst. Accordingly the ester 10

(Scheme 1) was prepared from trans-4-hydroxy--proline⁴²and

further transformed to the thymidyl-pyrrolidine 11 in an analo-

gous fashion to the corresponding methyl ester, which was pre-

pared earlier.³³Reduction of the ester 11 proceeded in 69%

yield with concomitant cleavage of the benzoyl-protecting

group to give the alcohol 12. Whilst the cleavage of the benzoyl

group was unexpected, it was nevertheless advantageous since

thymine-N3 protection was employed to ensure regioselectivity

in the previous Mitsunobu reaction,³³and was not envisaged to

be necessary for subsequent transformations. Initially we

sought to change the alcohol 12 to the corresponding amine 14

via the tosylate 15 and the azide 16. The tosylate 15 was thus

prepared in essentially quantitative yield, but upon heating to

80 ЊC with LiN₃in DMF for 16 h surprisingly none of the

expected azide 16 was obtained. Instead the 6-enamine-bridged

tricyclic compound 18 was isolated as the major product. Pre-

sumably under the thermal conditions of substitution reaction

the azide 16 undergoes an intramolecular 1,3-dipolar cyclo-

addition with the C5–C6 double bond of the thymine ring.^43,44

The resulting fused triazoline 17 presumably spontaneously

rearranges to give 18. In order to circumvent this, the alcohol 12

was converted into the corresponding phthalimide derivative

13 under standard Mitsunobu conditions. Removal of the

phthaloyl group from 13 proceeded in 92% yield on treatment

with 40% aqueous methylamine, at 40 ЊC for 1 h, to furnish

amine 14.

Scheme 2

Accordingly a solution of the primary amine 14 in dry di-

chloromethane was added slowly to a 65% (w/v) solution of

bromoacetic anhydride in acetonitrile at Ϫ15 ЊC, which resulted

in a near quantitative yield of the bromoacetamide 20. Removal

of the Boc protecting group from 13 under standard conditions

gave the secondary amine 19 as the TFA salt in essentially

quantitative yield. Coupling bromoacetamide 20 with this

amine salt 19 in the presence of Et₃N and DMF, however, gave

the dinucleotide analogue 21 in a disappointingly low 58%

yield. A major side product from this reaction was the primary

alcohol 23. However, when this coupling was attempted using

DIPEA instead of Et₃N, the yield of dimer 21 rose dramatically

to 98% with no observed side products. In both cases the reac-

tions were carried out under totally anhydrous conditions. It

seems most plausible that the less sterically hindered base tri-

ethylamine can deprotonate the acidic proton on thymine (N3H

pK_a= 9.7)³¹to liberate a nucleophilic neighbouring group,

which could participate in an intramolecular S_N2 displacement

of the bromide to give the lactam 22. This intermediate 22

could not be isolated probably because it undergoes facile

hydrolysis, either on TLC or during the aqueous workup, to

give the side-product 23.

Whilst the N-alkylation approach (19 ϩ 20

21) using

DIPEA in DMF was excellent in terms of yield, the reaction

proceeded slowly at room temperature (18 h) and would there-

fore not be ideal for future application in the solid-phase syn-

thesis of POM. Dimer formation using an acylation approach

was therefore also investigated (Scheme 3). Thus amine 19

was treated with tert-butyl bromoacetate to give the ester 24,

which upon treatment with 20% TFA in dichloromethane gave

acid 25 as a TFA salt. This acid 25 was then transformed into

the pentaﬂuorophenyl ester 26 in 85% yield using pentaﬂuoro-

phenyl triﬂuoroacetate. Acylation of primary amine 14 with the

Pfp-ester 26 proceeded smoothly at room temperature in di-

chloromethane and after 3 h gave the protected dinucleotide

analogue 21 in quantitative yield.

In order to obtain the fully deprotected T₂-POM 30 for

NMR-conformational studies, dinucleotide mimic 21 was

treated with methylamine, as before, in order to remove the

phthaloyl protecting group. However, three products were

isolated: the desired amine-dimer 27 as well as bicyclic lactam

28 and the amine-monomer 14. Clearly the basic conditions

required for removal of the phthaloyl group facilitate cyclis-

ation via the attack of the N-terminal amine on the central

amide linkage. This problem was overcome, to an extent, by

lowering the reaction temperature to room temperature and

stopping the reaction after only 40 min. This led to a mixture of

the desired product 27 and an intermediate, presumably the

ring-opened phthaloyl derivative 29. The latter was separated

Scheme 1

Two diﬀerent approaches were envisaged for chain extension

of POM. In the ﬁrst approach N-alkylation of the pyrrolidine

19 with bromoacetamide 20 was investigated (Scheme 2).

O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 2 7 7 – 3 2 9 2

3279

View Article Online

Table 1 ¹H NMR conformational analysis of T₂-POMؒ3HCl 30

Calculated ³J_HH/Hz^b

(dihedral angle ꢀ_HH/deg)^c

for conformer

Karplus parameters [r]^d

A

C

B and D

P = 198Њ

Vicinal H–H

Observed ³J_HH/Hz for 30^a

P = 48Њ

P = 38Њ

a_j

b_j

c_j

2Ј–3Ј

2Ј–3Љ

3Ј–4Ј

3Љ–4Ј

4Ј–5Ј

4Ј–5Љ

9.9^a

7.4

6.2

9.6

3.7

8.8

10.2^b

(157)^c

5.8

(42)

8.0

(143)

8.2

(25)

3.4

10.0

(155)

5.9

(41)

6.7

(136)

8.7

(20)

2.5

1.3

(92)

8.4

(23)

2.9

(113)

9.8

(2)

7.3

(140)

8.1

(20)

9.6

9.7

9.2

8.6

Ϫ0.99

1.2

1.0

1.1

1.4

(117)

9.0

(1)

(110)

8.9

(7)

3

^aThe observed J_HHcoupling constants for the tri-substituted “upper” pyrrolidine ring of T₂-POMؒ3HCl 30. ^bThe calculated J_HHcoupling

constants and ^cdihedral angles for the vicinal protons of the lowest energy conformers A–D of the model cis- and trans-dimethyl pyrrolidines 8 and 9

(Fig. 3). ^dThe Karplus parameters a_j, b_jand c_jderived previously for -4-hydroxyproline⁴⁵were used in the generalised Karplus equation to calculate

³J_HH(see Experimental section).

Noticeably the H2Ј–H3Ј coupling constant (³J_2Ј3Ј) of 9.9 Hz

observed for the upper ring of dimer 30 (Fig. 4) corresponds

well to the calculated values for both conformers A and C where

the H2Ј–H3Ј dihedral angles (ꢀ_2Ј3Ј) are 157Њ and 155Њ respect-

ively. In contrast both conformers B and D have identical

pucker (P = 198Њ) and thus equivalent dihedral angles (ꢀ_2Ј3Ј

=

3

92Њ), which correspond to a typically small calculated J_2Ј3Јof

3

1.3 Hz. Similarly, the observed J_3Ј4Јand J_4Ј5Јcoupling con-

stants of 6.2 and 3.7 Hz correspond more closely with the calcu-

lated values for either structure A or C. Thus, the observed

³J values are in best overall agreement with the structures A

(P = 48Њ) or C (P = 38Њ), both of which are analogous to the

northern or C3Ј-endo ribose ring conformations.

Scheme 3

and then treated with more methylamine to give amine 27 in a

combined yield of 75%. Treatment of the amine 27 with HCl

gave the totally deprotected hydrochloride salt T₂-POM 30 that

was used in the following conformational studies.

NMR conformational analysis of T₂-POM

In order to analyse the conformation of POM in aqueous solu-

Fig. 4 Numbering system for the tri-substituted “upper” pyrrolidine

ring of T₂-POMؒ3HCl 30 and 3D structures of the lowest energy

conformers A and B of trans-dimethyl pyrrolidine 9 from Fig. 3.

tion, the tri-cationic T₂-POM 30 was subjected to extensive 2D

1

¹H NMR analysis. The H NMR spectrum of 30, in D₂O, was

fully assigned with the aid of COSY-45, TOCSY, ROESY

and J-resolved experiments. Vicinal coupling constants were

obtained for every proton on the trisubstituted “upper”

pyrrolidine ring and these are shown in Table 1. Theoretical

coupling constants were also calculated, using a generalised

Karplus equation,⁴⁵and the dihedral angles obtained for the

four lowest energy conformers A, B, C and D, that were derived

previously from calculations performed on the model cis- and

trans-diastereoisomers 8 and 9 (Fig. 3). The experimentally

3

The analysis of J values provides no information regarding

the preferred conﬁguration of the pyrrolidine N-atom. In this

respect through space H–H correlations, as determined by

ROESY spectra, proved most useful. Notably H7Ј/7Љ–H2Ј and

H7Ј/7Љ–H5Љ ROE cross-peaks are observed for 30 (Fig. 4)

whereas no cross-peaks exist between H7Ј/7Љ and H5Ј or H3Ј.

This is most consistent with the trans N-conﬁguration of con-

formers A and B where the distances between H7Ј/7Љ and H2Ј or

H5Љ are in the order of 2.5 Å. In addition, H7Ј/7Љ–H6Ј/6Љ cross-

peaks are evident. This is most consistent with conformer A,

3

observed J values from T₂-POM 30 were then compared with

3

the calculated J_HHvalues of the model conformers (Table 1).

O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 2 7 7 – 3 2 9 2

3280

View Article Online

where the C7Ј–C6Ј dihedral angle is 76Њ which places the corre-

sponding protons within 2.5 Å. However, in conformer B, the

corresponding dihedral angle is much larger, 155Њ, which places

these protons more than 3.8 Å apart (Fig. 4). Interestingly,

H6–H3Ј and H6–H5Ј cross-peaks are also detected as well as a

H6–H4Ј cross-peak, which suggests that the thymine ring is in

equilibrium between the anti and syn conformations, relative to

the pyrrolidine ring. A similar conformational equilibrium is

often observed with natural nucleosides.³¹In summary both

case of the pentamer 34 than the dimer 21 (Scheme 3). Thus the

ﬁnal step in the synthesis was achieved by treating Phth-T₅-Boc

POM 34 with aqueous hydrochloric acid. Evaporation and

lyophilisation gave Phth-T₅-POM 35, as its HCl salt, which was

greater than 95% pure as determined by analytical reverse

phase HPLC (see ESI Fig. 4). All subsequent nucleic acid bind-

ing studies were carried out with this oligomer 35, retaining the

phthaloyl group. Subsequent as yet unpublished results have

indicated that the N-terminal phthaloyl moiety has little eﬀect

on DNA/RNA binding properties compared with N-acetyl

capped oligomers.

3

observed J_HHcoupling constants and ROEs for the tri-

substituted “upper” pyrrolidine ring in 30 are consistent with

the conformer A (Fig. 4). This conﬁguration and conformation

was also the lowest energy structure determined by modelling.

Whilst it is possible that the tri-substituted pyrrolidine ring in

30 is in equilibrium between several conformations, in solution,

the fact that the NMR and modelling studies are consistent

suggests that the most highly populated, lowest energy structure

is that which most closely matches a typical nucleotide in

RNA.

UV spectroscopic analysis of T₅-POM–nucleic acid binding

properties

Phth-T₅-POM binding poly(rA). UV thermal denaturation

experiments were performed in order to evaluate the aﬃnity of

Phth-T₅-POM 35 upon hybridisation with RNA. Initially the

variation of absorption (A₂₆₀) of an equimolar mixture of Phth-

T₅-POM 35 and poly(rA) (42 µM each in bases) in 10 mM

K₂HPO₄buﬀer adjusted to 0.12 M K^ϩ, pH 7.0, was recorded.

A complete cycle of the thermal induced denaturation and

renaturation involves fast heating (5 ЊC min^Ϫ1), slow cooling

(0.2 ЊC min^Ϫ1), followed by slow heating (0.2 ЊC min^Ϫ1). The

slow heating curve resulted in a single 26% hyperchromic shift

with a melting temperature (T_m) of 48.5 ЊC (ca. 10 ЊC T_m/base)

(Fig. 5a). Under identical conditions corresponding native

d(T)₅did not show any hyperchromic shift with poly(rA) above

8 ЊC, whilst the denaturation of d(T)₂₀and poly(rA) has a T_mof

42.0 ЊC (ca. 2 ЊC T_m/base). A T₅-PNA containing a Lys residue

at either terminus exhibits similar aﬃnity for poly(rA), as Phth-

T₅-POM 35, with a T_mof 56 ЊC. While it is often hazardous to

extrapolate a T_mvalue obtained with homo-oligomers into an

accurate value of T_mper modiﬁcation found in a mixed

sequence, it is clear that Phth-T₅-POM 35 binds with extremely

high aﬃnity to poly(rA). Thus it would seem that the pyrroli-

dine ring imposes a favourable conformation for hybridisation

with RNA. Based on results from modelling and NMR studies

(vide supra) this is likely to be, at least in part, a result of the

favourable northern C3Ј-endo type conformation adopted by

the pyrrolidine ring.

Interestingly, substantial hysteresis was evident between the

cooling and heating curves in this experiment (see ESI Fig. 6)

indicating that the rate of heating/cooling employed (0.2 ЊC

min^Ϫ1) is faster than the rate of association/dissociation of

Phth-T₅-POM 35 and poly(rA) such that a true equilibrium is

not attained. This is in contrast to native duplex forming nucleic

acids which show essentially no hysteresis at this rate of tem-

perature change, implying Phth-T₅-POM 35 hybridises more

slowly to poly(rA) than do native nucleic acids. Since it is

impossible to retrieve thermodynamic data unless the heating/

cooling curves are coincident, an even slower rate of heating/

cooling was employed. However, even at a heating/cooling rate

of 0.1 ЊC min^Ϫ1, equilibrium conditions were still not realised.

Thus in order to obtain a true equilibrium T_mvalue, an equi-

molar sample of Phth-T₅-POM 35 and poly(rA) was subjected

to a series of cooling/heating ramps performed at ﬁve diﬀerent

rates (5, 2, 1, 0.5 and 0.1 ЊC min^Ϫ1). It was evident from these

curves that the slower rates of annealing result in higher T_m

values and greater hyperchromic shifts (see ESI Figs. 7 and 8).

By extrapolating to an inﬁnitely slow rate of heating/cooling,

the true equilibrium T_mwas determined to be 49 ЊC and is 0.5

ЊC higher than that observed at 0.2 ЊC min^Ϫ1. All further heat-

ing/cooling experiments were performed at 0.2 ЊC min^Ϫ1and

so the quoted T_mvalues are slightly lower than the true

equilibrium values and stand uncorrected.

Synthesis of T₅-POM

For our preliminary investigations into the nucleic acid hybrid-

isation properties of POM we decided to synthesise and evalu-

ate T₅-POM. The synthesis toward T₅-POM proceeded using

iterative alkylation of the pyrrolidine ring with the bromo-

acetamide 20 already in hand. Accordingly, Boc protected T₂-

POM 21 was treated with triﬂuoroacetic acid to give the

secondary amine 31, as the TFA salt. Coupling of 31 with bro-

moacetamide 20 proceeds in DMF with excess DIPEA to give

Boc protected T₃-POM 32, in near quantitative yield (Scheme

4). Repeating this Boc deprotection/coupling cycle leads to the

Boc protected T₄- and T₅-POM 33 and 34 respectively, in

similarly satisfactory yields after puriﬁcation by silica gel

chromatography. Complete assignment of the ¹H and ¹³C NMR

spectra of these protected oligomers 32–34 was signiﬁcantly

complicated due to spectral overlap. Nevertheless the signals of

the thymine CH₃and C4 carbon atoms are clearly resolved in

the ¹³C NMR spectra recorded in CD₃OD (see ESI Fig. 2),

whilst the thymine NH signals can be distinguished from the ¹H

NMR in DMSO-d₆(see ESI Fig. 3). Attempts to remove the

phthaloyl group of Phth-T₅-Boc POM 34, again using aqueous

methylamine, proved unsuccessful and were ultimately aban-

doned. It appears that base catalysed decomposition, involving

attack of the N-terminal amine on the central amide linkage to

give bicyclic lactam 28, is considerably more signiﬁcant in the

Fidelity and stoichiometry of T₅-POM binding RNA. In order

to assess the ﬁdelity of Phth-T₅-POM 35 binding to RNA, UV

denaturation experiments were carried out with 35 and the

Scheme 4

O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 2 7 7 – 3 2 9 2

3281

View Article Online

Fig. 5 (a) UV thermal denaturation curves of Phth-T₅-POM 35, Lys-T₅-LysNH₂-PNA (N- and C-terminal Lys, with C-terminal amide) and d(T)₂₀

vs. poly(rA) (42 µM each in bases) in buﬀer A (10 mM K₂HPO₄) adjusted to 0.12 M K^ϩ, pH 7.0. (b) UV thermal denaturation curves for Phth-T₅-

POM 35 and an equimolar amount (42 µM each strand in bases) of poly(rA), (rU), (rC) or (rG) in buﬀer A adjusted to 0.12 M K^ϩ, pH 7.0. (c) A Job

plot of the % change in A₂₆₀vs. molar ratio Phth-T₅-POM 35 and poly(rA).

non-complementary homopolymers poly(rG), (rC) and (U)

(Fig. 5b). In all three cases no signiﬁcant hyperchromicity is

observed, hence it can be concluded that Phth-T₅-POM 35 does

not bind to the non-complementary homopolymers by a typical

base-pairing/stacking mechanism. These results do not, how-

ever, rule out non-speciﬁc binding, via electrostatic interactions

between the oppositely charged backbones.

A single hyperchromic shift was observed in the thermal

denaturation of Phth-T₅-POM 35 and poly(rA) which is con-

sistent with the melting of a Watson and Crick base paired

duplex or the simultaneous dissociation of the three strands of

a TA.T triple helix. In order to determine the binding stoichio-

metry and discriminate between these two possibilities the

method of UV continuous variation was employed.⁴⁶Here UV

absorbance, over the range 220–300 nm, was measured as a

function of the molar ratio of bases, for mixtures of Phth-T₅-

POM 35 and poly(rA). An incubation period of 24 h prior to

measurements was found to give the clearest results, presum-

ably due to the relatively slow rate of association between 35

and poly(rA). A Job plot⁴⁶of the percentage change in A₂₆₀

against mole ratio of T : A shows a well-deﬁned minimum at

1 : 1 consistent with duplex formation. In addition, the mini-

mum in the Job plot occurs at the intersection of two straight

lines, indicating a reversibly formed complex with few or no

vacant sites.⁴⁷

Eﬀect of ionic strength and pH on T_mfor Phth-T₅-POM bind-

ing RNA. Bruice and co-workers showed that decrease in ionic

strength results in an increase in the binding aﬃnity of the

positively charged oligonucleotide DNG 5 (Fig. 1) to poly-

(dA) and poly(rA).^48,49This is due the “intimate” association

between the more “naked” oppositely charged backbones at

lower salt concentration. In order to determine whether electro-

static attraction could contribute to the stability of the duplex

formed between Phth-T₅-POM 35 and poly(rA), melting

Fig. 6 (a) Transition melting temperatures per base (T_m/base) of Phth-

T₅-POM 35, Lys-T₅-LysNH₂-PNA and d(T)₂₀with poly(rA) (42 µM

experiments were carried out at diﬀerent ionic strength over a

ten-fold concentration range. Surprisingly, there is actually a

slight increase in duplex stability with increasing ionic strength

between 0.12 and 1.2 M K^ϩ(T_m= 48.5 and 55.0 ЊC respect-

ively), which equates to an increase in T_m/base of ca. 1 ЊC

(Fig. 6a). This implies that electrostatic attraction is not a major

contributing factor to the duplex stability. This may be a con-

sequence of either there being a relatively large distance

between the positive charge in POM and the negative charge of

the phosphodiester group in a duplex with RNA, or that Phth-

T₅-POM 35 is only partially protonated at pH 7.0. In addition,

it may be that the observed increased aﬃnity of 35 for poly(rA)

at higher ionic strength is due to 35 adopting a slightly diﬀer-

ent conformation at higher salt concentration. It was also

noticeable from the UV melting curves, recorded at diﬀerent

salt concentrations, that the hyperchromicity decreases signiﬁ-

cantly, whilst hysteresis between the heating and cooling curves

each in bases) in buﬀer A (total volume 1 mL) adjusted to the

appropriate ionic strength and pH. UV absorbance (A₂₆₀) was recorded

with heating at 5 ЊC min^Ϫ1from 25 ЊC to 93 ЊC, cooling at 0.2 ЊC min^Ϫ1

to 15 ЊC and heating at 0.2 ЊC min^Ϫ1to 93 ЊC. The T_mwas determined

from the ﬁrst derivative of the ﬁnal slow heating curve. (b) T_m/base for

Phth-T₅-POM 35, Lys-T₅-LysNH₂-PNA and d(T)₂₀with poly(dA). As

above except the mixture of 35 and poly(dA) (210 µM each in bases)

was incubated for 48–96 h at 25 ЊC in buﬀer A (total volume 0.2 mL),

diluted to 1 mL adjusting to the appropriate ionic strength and pH,

cooled to 15 ЊC at 1 ЊC min^Ϫ1, heated at 0.2 ЊC min^Ϫ1to 93 ЊC from

which the T_mwas measured.

increases, at higher ionic strengths (see ESI Fig. 9). This is

probably due to slower rates of association and dissociation at

higher ionic strength.

In order to assess the eﬀect of pH on the aﬃnity of Phth-T₅-

POM 35 binding to poly(rA), melting experiments were carried

out at several diﬀerent pH values over the range 6–8 (10 mM

O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 2 7 7 – 3 2 9 2

3282

View Article Online

phosphate buﬀer adjusted to 0.12 M K^ϩ). This revealed that

there is a substantial increase in T_mfrom 45 to 57 ЊC (∆T_m/base

= 2.1 ЊC) on lowering the pH from 8.0 to 6.0 (Fig. 6a). Thus,

protonation of the nitrogen atom of the pyrrolidine ring is an

important factor in the formation of more stable duplexes. The

fact that duplex stability increases with salt concentration sug-

gests that conformational changes brought about by proton-

ation are more likely to be the cause of the observed increase in

duplex stability at lower pH, than electrostatic attraction. From

the melting curves (see ESI Fig. 10) it was again noticeable that

greater hyperchromic shifts and reduced hysteresis are observed

at lower pH, suggesting that binding occurs faster at lower pH.

The fact that there was signiﬁcant hysteresis between the heat-

ing and cooling curves even at pH 6.0 indicates binding is still

slow relative to native duplex hybridisation.

Phth-T₅-POM binding DNA. In contrast to the distinct melt-

ing curves obtained with poly(rA), repeated UV heating/

cooling experiments with an equimolar mixture (42 µM each in

bases) of Phth-T₅-POM 35 and poly(dA) revealed no hyper-

chromic shifts between 15 and 93 ЊC. Surprisingly, a modest

cooperative hyperchromic shift could be detected after a ﬁve-

fold increased concentration of both 35 and poly(dA) (210 µM

each in bases, total volume 200 µL) was incubated at 25 ЊC for

48 h, before the volume was readjusted to 1.0 mL (42 µM each

in bases) (Fig. 7a). However, the T_mof 57.0 ЊC (11.4 ЊC/base)

was only just detectable from the sloping baselines, indicating

only a fraction of the single strands had fully annealed. Thus

Phth-T₅-POM 35 can hybridise with poly(rA) in the order of

minutes (at 42 µM each in bases), but remarkably only partially

hybridises with poly(dA) at a ﬁve-fold increased concentration

after an extended period of 2 days incubation. It is also of note

that although Phth-T₅-POM 35 binds extremely slowly to

poly(dA), it does so with a higher aﬃnity than for poly(rA)

(increase in T_mof 8 ЊC at 0.12 M K^ϩ, pH 7.0). In addition, 35

binds to poly(dA) with higher aﬃnity compared to Lys-T₅-Lys-

PNA, which gave rise to a T_mof 48 ЊC (9.6 ЊC/base). All sub-

sequent T_mvalues for equimolar mixtures of 35 and poly(dA)

were obtained by incubating the two complementary strands at

a ﬁve-fold increased concentration (210 µM each in bases) for at

least 48 h prior to diluting the sample ﬁve-fold. While obtaining

T_mvalues with poly(dA) proved more diﬃcult than with

poly(rA), the trends appear the same for both series (Fig. 6b).

Phth-T₅-POM 35 binds to poly(dA) with higher aﬃnity at

higher ionic strength and lower pH. In some cases two hyper-

chromic shifts could be detected, suggesting triplex formation.

However, due to the extremely slow binding kinetics, the stoi-

chiometry of binding could not be satisfactorily determined,

using the method of UV continuous variation. Indeed,

although the Job plot for 35 and poly(dA), obtained at an ionic

strength of 0.12 M and pH 7.0, gave a minimum at approx-

imately 30% T indicative of a TA.T triplex, this was poorly

deﬁned.

Fig. 7 (a) UV thermal denaturation curves of Phth-T₅-POM 35 vs.

poly(dA) following incubation for 48–96 h at 25 ЊC in buﬀer A (210 µM

each in bases, 0.2 mL), ﬁve-fold dilution and adjustment of ionic

strength and pH. (b) Normalised UV absorbance (A₂₆₀) of Phth-T₅-

POM 35 with poly(rA) and (dA) vs. time at 25 ЊC. 35 and poly(dA)

(42 µM each in bases), 0.12 M K^ϩ, pH 7 (᭺); 35 and poly(dA) (210 µM),

0.12 M K^ϩ, pH 7 (ᮀ); 35 and poly(rA) (42 µM), 0.22 M K^ϩ, pH 7 (ꢀ);

35 and poly(rA) (42 µM), 0.12 M K^ϩ, pH 7 (᭹); 35 and poly(rA)

(42 µM), 0.12 M K^ϩ, pH 6 (᭿).

poly(rA) over poly(dA), we decided to monitor the binding

events as a function of time. Initially absorbance changes at 260

nm were monitored at 25 ЊC immediately upon mixing an

equimolar amount of 35 into a buﬀered solution containing

either poly(rA) or poly(dA) (42 µM each in bases, 0.12 M K^ϩ,

pH 7.0). Under these conditions an hypochromic shift of 29%

was observed after 200 min upon mixing 35 with poly(rA)

(Fig. 7b). However when 35 was similarly mixed with poly(dA)

no drop in A₂₆₀was observed even after 16 h. It was only after

increasing the concentration of each strand ﬁve-fold to 210 µM

(in bases), that hybridisation between Phth-T₅-POM and

poly(dA) can be observed. At this elevated concentration a

moderate absorbance drop of only ca. 6% was observed after

200 min, remarkably A₂₆₀was still dropping even after 12 h.

These experiments clearly conﬁrm that Phth-T₅-POM 35 binds

much faster to RNA than DNA. Overall, however, it is also

evident that binding of Phth-T₅-POM to poly(rA) is slow when

compared with Lys-T₅-Lys-PNA. In an identical UV kinetic

experiment PNA hybridisation with poly(rA) (42 µM each in

bases) is essentially complete after 10 minutes, whilst Phth-T₅-

Following our initial communication of POM,²⁵it was

reported that incorporation of either one or two thymidyl-

pyrrolidine units, with the same stereochemistry as POM

(2ЈR,4ЈR), into ssDNA results in substantially destabilised

duplexes and triplexes, relative to wild-type, upon hybridisation

with complementary ssDNA and dsDNA, respectively.³⁹Our

results using a fully modiﬁed Phth-T₅-POM 35 thus highlight

that it is clearly very hazardous, and unadvisable, to extrapolate

T_mvalues based on such a simple chimeric approach as a means

for understanding the nucleic acid binding properties of fully

modiﬁed novel nucleic acid mimics.

UV kinetic analysis of Phth-T₅-POM binding DNA and RNA

Phth-T₅-POM binds faster to RNA than DNA. Since UV

thermal denaturation experiments suggested there to be a sub-

stantial kinetic selectivity for Phth-T₅-POM 35 binding to

O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 2 7 7 – 3 2 9 2

3283

View Article Online

POM fails to reach equilibrium even after 100 minutes (see ESI

Fig. 11).

and thermodynamics and thus binding speciﬁcity.⁵⁶It was also

chosen as a complementary technique to UV spectroscopy in

order to observe more readily the slow interaction between

Phth-T₅-POM 35 and poly(dA) since SPR permits the use of

higher analyte concentrations than is possible with UV spectro-

scopy. Accordingly 5Ј-biotinylated targets, d(A)₂₀and r(A)₂₀,

were immobilised onto a streptavidin loaded carboxymethyl-

dextran matrix of a standard SPR sensor chip (Biacore,

Stevenage). To ensure an equivalent number of binding sites,

d(A)₂₀and r(A)₂₀were injected across ﬂow cells 2 and 3 of the

four compartment sensor chip to give values of 1400 and 1450

response units (RU), respectively. In order to ensure against

non-speciﬁc electrostatic interactions that may have arisen from

the interaction between positively charged Phth-T₅-POM 35

and the negatively charged carboxylate surface of the matrix,

one surface (ﬂow cell 1) was left underivatised. In addition, as a

control of Phth-T₅-POM 35 binding speciﬁcity, the surface of

ﬂow cell 4 was derivatised with an ssDNA of mixed sequence,

5Ј-biotin-d(AGC TTC AGA GAT CGA TCG GAG AGA

GTA CTG) (NFκB), to a value of 1000 RU. Having immobil-

ised the native oligonucleotides, Phth-T₅-POM 35 was injected

at diﬀerent concentrations across all four ﬂow cells of the sen-

sor surface using a 10 mM K₂HPO₄buﬀer adjusted to 0.12 M

K^ϩand pH 7.0. Typically, Phth-T₅-POM 35 was injected across

the surfaces for 300 s followed by buﬀer alone. After this dis-

sociation phase, the surface was regenerated and any excess 35

was removed, by injecting a pulse of 10 mM HCl across the

surfaces.

Fig. 8a shows the sensogram obtained at a concentration of

40 µM and shows clearly that while Phth-T₅-POM 35 associates

with d(A)₂₀, it does so considerably slower than to r(A)₂₀. Phth-

T₅-POM 35 also shows no signiﬁcant binding to the negatively

charged carboxymethyl dextran matrix, conﬁrming the feasibil-

ity of using this type of sensor surface. Crucially, no binding

was observed to the mixed sequence control (NFκB), above

background, demonstrating that Phth-T₅-POM 35 retains the

high ﬁdelity of base pair recognition displayed by native nucleic

acids and does not exhibit any non-speciﬁc binding through

electrostatic interactions. Furthermore when the concentration

of Phth-T₅-POM 35 injected was increased two-fold (80 µM,

Fig. 8b) there was still no binding to the mixed sequence con-

trol, above the background. These results are also signiﬁcant in

that they prove that the hypochromic shifts observed upon

hybridising Phth-T₅-POM 35 with poly(dA) compared to

poly(rA) are not an artefact of poor base stacking within the

Phth-T₅-POM/poly(dA) complex relative to that of the Phth-

T₅-POM/poly(rA) complex, but are due to a diﬀerence in

relative rates of association.

To the best of our knowledge, there have been no reports of a

kinetically selective sequence speciﬁc association of a ligand for

RNA over ssDNA (or vice-versa). However other systems have

been reported to display weak but thermodynamically selective

binding for single stranded RNA over DNA. For example,

2Ј,5Ј-linked RNA,⁵⁰2Ј,5Ј-linked DNA⁵¹and 2Ј,5Ј-linked thio-

formacetal-modiﬁed⁵²oligonucleotides all show substantial

thermodynamic selectivity for hetero-duplex formation with

RNA over ssDNA, albeit they bind with lower aﬃnity to com-

plementary RNA than do the corresponding isosequential

DNA oligonucleotides. At present we have no satisfactory

model to explain this kinetic selectivity of Phth-T₅-POM for

poly(rA) over poly(dA) and further studies will be required to

delineate the mechanism of binding. Perhaps binding of

Phth-T₅-POM to poly(dA) induces a major conformational

change to poly(dA) (or vice-versa), which retards further

binding. It is noteworthy that a chiral PNA thymidyl decamer

with a (2R,4R)-N-aminoethyl--proline backbone reported by

Vilaivan et al.,³⁷which diﬀers only in the position of the amide

group in the backbone, was also found to bind selectively to

poly(rA) and not poly(dA). However, no investigation into

binding kinetics was reported and it is thus tempting to postu-

late that longer incubation times would likewise reveal some

binding to poly(dA). Ganesh and co-workers found no such

selectivity with the stereoisomers (2S,4S) and (2R,4S) of this

N-aminoethyl proline based oligomer.³⁸However, these oligo-

mers³⁸were chimeras with just one pyrrolidine unit incorpor-

ated in a standard PNA strand. Thus any conclusions that were

drawn from this work³⁸are, at best, fragile.

Ionic strength and pH also aﬀect the rates of hybridisation.

Similar UV kinetic experiments were also carried out to deter-

mine the eﬀects of changing ionic strength and pH on the kinet-

ics of Phth-T₅-POM 35 binding poly(rA) (see Fig. 7b and ESI

Figs. 12 and 13). This reveals that binding is much faster at

lower ionic strengths. This conﬁrms that greater hyperchromic

shifts observed in the UV thermal denaturation of Phth-T₅-

POM 35/poly(rA) duplexes (vide supra) at lower ionic strengths

must arise from a higher percentage of duplex formed on

annealing due to faster rates of association. Furthermore,

these results suggest that electrostatic interactions are import-

ant in determining the kinetics of duplex formation. The results

for Phth-T₅-POM 35 binding poly(rA) at diﬀerent pHs also

conﬁrm that binding is faster at pH 6 than at pH 7, which

in turn, is considerably faster than at pH 8. It is also notice-

able that the rates of binding are much less sensitive to pH

than to ionic strength. Unfortunately, the hybridisation of 35

with poly(rA) did not obey either ﬁrst order or pseudo ﬁrst-

order kinetics. Indeed no quantitative kinetic analysis was per-

formed using this data since as far as we are aware, no model

exists to account for the binding of a short ligand containing

several identical binding domains to a target bearing many

potential binding sites of which the degree of occupancy is

unknown.⁵³

As a comparison, d(T)₅was injected across the sensor

surfaces. As expected, no duplex formation was observed

with d(A)₂₀or r(A)₂₀at concentrations even as high as 160

µM. This clearly reﬂects the much higher binding aﬃnity

of Phth-T₅-POM 35 compared with its DNA 5mer counter-

part.

The eﬀect of ionic strength on the rates of association and

dissociation of Phth-T₅-POM with d(A)₂₀and r(A)₂₀was

investigated using a buﬀer with an increase in KCl concen-

tration. This reveals that the rates of association to both r(A)₂₀

and d(A)₂₀are very much reduced at higher ionic strength (see

ESI Fig. 14), consistent with observations obtained by UV

spectroscopy. As postulated previously, this is most likely due to

diminished inter-strand electrostatic attraction since the oppos-

itely charged backbones will eﬀectively be shielded from one

another by the presence of more counter ions in the solution.

However, it is also possible that conformational diﬀerences

arise at higher ionic strength which could be responsible for this

diﬀerence. Again, faster binding is observed to r(A)₂₀than to

d(A)₂₀at higher ionic strength, albeit the diﬀerence is slightly

diminished compared with a buﬀer of low ionic strength.

Nonetheless, this shows that the factors responsible for the

SPR analysis

Since UV spectroscopy can only detect changes in base stacking

interactions, it cannot be used to unequivocally ascertain the

speciﬁcity of binding of modiﬁed oligonucleotides to nucleic

acids. This was a particular concern since any non-speciﬁc

electrostatic backbone interactions between Phth-T₅-POM 35

and complementary or non-complementary nucleic acids need

not involve a change in the extent of base-stacking. Surface

Plasmon Resonance (SPR) is a technique which is gaining

popularity for studying biomolecular interactions.^54–56Whilst it

has rarely been used for the study of modiﬁed oligonucleotide–

nucleic acid interactions, it can give rapid insight into kinetics

O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 2 7 7 – 3 2 9 2

3284

View Article Online

NFκB is also observed, implying that the binding is weak. Thus

at pH 5.0, weak, non-speciﬁc interactions occur between Phth-

T₅-POM 35 and ssDNA, presumably arising from electrostatic

attraction, such that the speciﬁc Watson–Crick type base-

pairing is disrupted. This also represents the ﬁrst example of a

modiﬁed oligonucleotide capable of switching from sequence

speciﬁc to sequence non-speciﬁc binding of a nucleic acid under

the control of an external stimulus (pH). Finally it is note-

worthy that the rate of dissociation of Phth-T₅-POM 35 from

the mixed sequence ssDNA (NFκB) at pH 5.0 is much faster

than from d(A)₂₀, suggesting that, in the presence of a comple-

mentary sequence, Phth-T₅-POM binds to d(A)₂₀preferentially

via Watson–Crick type base-pairing.

Summary

An eﬃcient synthetic route to the highly water-soluble cationic

dinucleotide analogue T₂-POM 30, has been developed. ¹H

NMR spectroscopy has shown that the N-acetamido group of

the protonated pyrrolidine ring in POM adopts a preferred

trans-conﬁguration relative to the base and thus the backbone

is stereochemically equivalent to native nucleic acids. Semi-

empirical quantum mechanical calculations carried out on di-

methylpyrrolidine nucleoside analogues also support this

observation. In addition, it was found that the pyrrolidine ring

preferentially adopts a conformation similar to the northern-

type conformation adopted by ribose in RNA. Moreover, since

the thymine base was found to be suitably oriented for

intermolecular hydrogen bonding, thymidyl POM satisﬁes the

necessary conﬁgurational and conformational requirements

for hybridisation with DNA and RNA.

Oligonucleotide analogues comprised of a pyrrolidine-

amide backbone are emerging as a new class of nucleic acid

mimics with desirable properties.^36,37,40Our preliminary studies

have found that the pyrrolidine-amide oligonucleotide mimics

(POM), in particular, exhibit many desirable, as well as

unforeseen properties. Solution-phase synthesis of the highly

water soluble thymidyl pentamer Phth-T₅-POM was achieved

using an eﬃcient iterative N-alkylation strategy. UV spectro-

scopy and surface plasmon resonance have shown that Phth-

T₅-POM binds with very high aﬃnity and speciﬁcity to com-

plementary ssDNA and RNA. Surprisingly, upon increasing

the ionic strength, slightly elevated T_m’s were observed, suggest-

ing that electrostatic attraction between oppositely charged

backbones does not play a signiﬁcant role in binding aﬃnity.

Decreasing the pH resulted in elevated T_m’s which is possibly a

consequence of conformational changes brought about by pro-

tonation of the pyrrolidine ring. By contrast, Phth-T₅-POM

was found to bind to complementary RNA faster at lower ionic

strength and pH, implying that electrostatic interactions are

important in determining the kinetics of hybridisation. SPR

analysis has also shown unequivocally that Phth-T₅-POM

maintains the ﬁdelity of sequence speciﬁc binding (base-

pairing) observed with native nucleic acids, at both pH 7.0 and

6.0. However, at pH 5.0 the cationic oligonucleotide was

attracted to a non-complementary ssDNA, presumably via

electrostatic contacts between oppositely charged backbones.

Intriguingly, Phth-T₅-POM was found to bind much faster to

RNA than to DNA. We are currently developing a solid-phase

synthesis of longer mixed sequence POM in order to explore in

greater depth this kinetically selective recognition, which may

provide new insight into the conformational diﬀerences

between the two native nucleic acid forms.

Fig. 8 SPR response (RU) vs. time for Phth-T₅-POM 35 injected

across r(A)₂₀, d(A)₂₀, d(AGC TTC AGA GAT CGA TCG GAG AGA

GTA GTG-3Ј) derivatised surfaces and an underivatised control

surface. Response proﬁles were determined using a BIAcore 2000

instrument (Biacore AB, UK) under standard conditions following

injection of Phth-T₅-POM 35 (a) 40 µM and (b) 80 µM strand

concentration (in buﬀer A, pH 7 adjusted to 0.12 M K^ϩ) across each

surface at a ﬂow rate of 20 µL min^Ϫ1for 300 s, followed by buﬀer to

allow dissociation. All oligonucleotides were biotinylated at the 5Ј-end

(Cruachem, UK) and immobilised on

a streptavidin derivatised

carboxymethyl dextran sensor chip (Biacore).

kinetic selectivity of Phth-T₅-POM 35 for binding RNA over

ssDNA are largely independent of ionic strength.

Next the eﬀect of pH on the rates of association, dissociation

and speciﬁcity of binding of Phth-T₅-POM 35 with immobil-

ised native oligonucleotides was investigated. This revealed that

the overall trends at pH 6.0 are very similar to those at pH 7.0,

with Phth-T₅-POM 35 binding much faster to r(A)₂₀relative to

d(A)₂₀(see ESI Fig. 15). Thus the factors responsible for the

kinetic selectivity prevail even at lower pH. Faster binding to

both targets was also observed at pH 6.0 compared with that at

pH 7.0, and is in agreement with the UV spectroscopic meas-

urements. This is likely to be a consequence of enhanced elec-

trostatic attraction between the more protonated Phth-T₅-POM

and target nucleic acids. Notably, lowering the pH to 6.0 does

not result in any non-speciﬁc binding to the mixed sequence

above the background response of the underivatised surface.

However, when the same experiments were performed at pH 5.0

(0.12 M K^ϩ), substantial binding to the mixed sequence ssDNA

(NFκB) was observed whereas very little binding to the dextran

was detected (see ESI Fig. 16). A fast rate of dissociation from

Experimental section

¹H NMR spectra were recorded at 270, 300, 400 or 500 MHz.

¹³C NMR spectra were recorded at 67.9, 75.5, or 100.6 MHz.

Chemical shifts are reported in parts per million relative to

Me₄Si or residual solvent signal. ¹⁹F NMR spectra were

O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 2 7 7 – 3 2 9 2

3285

View Article Online

recorded at 188.2 MHz unreferenced. FAB mass spectra were

obtained on a ZAB-SE VG Analytical Fisons Instrument. Elec-

tron Impact (EI) and Chemical Ionisation (CI) mass spectra

were recorded using a VG Analytical 70 70 EQ mass spectro-

meter using NH₃as carrier gas. Electrospray ionisation (ESI)

mass spectra were performed, in positive ion mode, using a Q-

Tof (Micromass, Manchester). Infra-red spectra were acquired

on a Perkin-Elmer 783 as KBr discs or CHCl₃solutions. Melt-

ing points were determined using a Cambridge Instrumental

microscope with a Reichert-Jung heating mantle and are un-

corrected. Optical rotations were measured at 20 ЊC with an

Optical Activity AA-1000 polarimeter. Flash silica column

chromatography was carried out over Merck Kieselgel 60 (230–

400 mesh) and Merck Kieselgel 60 F254 0.25 mm plates were

used for analytical TLC. Reactions involving anhydrous condi-

tions were carried out in ﬂame-dried glassware under a posi-

tive pressure of argon. All solvents were distilled before use.

Reagents were puriﬁed and solvents dried using standard

procedures.

N-(tert-Butoxycarbonyl)-trans-4-formyloxy-D-proline

ethyl

ester. Ph₃P (9.47 g, 36.0 mmol), followed by dry formic acid

(1.4 mL, 37.0 mmol), was added to a stirred solution of

N-(tert-butoxycarbonyl)-cis-4-hydroxy--proline ethyl ester

10⁴²(7.6 g, 29.0 mmol) in dry THF (125 mL) under argon. The

mixture was then cooled to Ϫ20 ЊC and diisopropyl azodicarb-

oxylate (DIAD) (7.3 mL, 37.0 mmol) was added dropwise over

30 min. The solution was held at Ϫ20 ЊC for a further 15 min

and then allowed to warm to room temperature and stirred for

a further 18 h. The resulting yellow solution was evaporated

under reduced pressure, and Et₂O (ca. 50 mL) was added to

precipitate triphenylphosphine oxide which was removed by ﬁl-

tration. The ﬁltrate was evaporated under reduced pressure and

the residue was puriﬁed by ﬂash chromatography, eluting with

5% acetone in CH₂Cl₂, to give the title compound (7.62 g, 90%)

as a pale yellow oil. R_f0.28 (5% acetone in CH₂Cl₂); [α]_Dϩ46.5

(c = 2.0, EtOH); δ_H(300 MHz, CDCl₃) 1.28–1.38 (3H, m,

CH₂CH₃), 1.44 and 1.47 (total 9H, 2 × s, C(CH₃)₃rotamers),

2.20–2.29 (1H, m, H_a3), 2.39–2.49 (1H, m, H_b3), 3.58–3.74 (2H,

m, H5), 4.22 (2H, q, J 7.0, CH₂CH₃), 4.34–4.47 (1H, 2 × t, each

J 7.9, H2 rotamers), 5.43 (1H, br s, H4), 8.02 and 9.21 (total

1H, 2 × s, HCO rotamers); δ_C(100.6 MHz, CDCl₃) 14.5

(CH₂CH₃), 28.6 and 28.7 (C(CH₃)₃rotamers), 35.8 and 36.9

(C3 rotamers), 52.2 and 52.5 (C5 rotamers), 57.9 and 58.2 (C2

rotamers), 61.6 (CH₂CH₃), 71.1 and 72.0 (C4 rotamers), 81.0

Molecular modelling

Co-ordinates of X-ray crystal structures of uridine 7³²and a

pyrrolidine HCl salt 6²⁹(Fig. 2) were obtained from the Cam-

bridge Crystallographic Database. The pseudorotation phase

angle (P) and degree of pucker or maximum torsional angle

(ν_max) are calculated by substituting the published torsion

t

(C(CH₃)₃), 153.9 and 154.4 (CO₂Bu rotamers), 160.4 and 160.5

(HCO₂rotamers), 172.6 and 172.8 (CO₂Et rotamers); m/z

(FAB^ϩ) 288 ([M ϩ H]^ϩ, 53%), 260 (45), 232 (47), 188 (63);

HRMS m/z (CI) 288.1451 ([M ϩ H]^ϩ, C₁₃H₂₂NO₆requires m/z,

288.1447).

angles (ν₀

ν₄) into the equations: Tan P = {(ν₄ϩ ν₁) Ϫ (ν₃ϩ

ν₀)}/{2 ؒ ν₂ؒ (sin 36Њ ϩ sin 72Њ)}and ν_max= ν₂/cos P.^30,31Models

of the cis- and trans-dimethylpyrrolidines 8 and 9 (Fig. 3) were

built from the co-ordinates of the reported crystal structure of

pyrrolidine 6. For both diastereoisomers 35 conformations, for

every 10 degrees of pseudorotation, were energy minimised,

allowing only free rotation of the thymine base about the C4Ј–

N1 bond, in order to optimise the C3Ј–C4Ј–N1–O2 torsion

angle which is analogous to the glycosyl torsion angle in native

nucleic acids (χ).³¹The enthalpies of formation for each con-

former were then calculated by semi-empirical quantum mech-

anical calculations using MOPAC 6.0 software⁴¹on a Silicon

Graphics workstation. The maximum torsional angle (ν_max) was

optimised to 40Њ. The standard enthalpies of formation, phase

N-(tert-Butoxycarbonyl)-trans-4-hydroxy-D-proline

ethyl

ester. 25% aq. NH₃(1.8 mL) was added to a solution of N-(tert-

butoxycarbonyl)-trans-4-formyloxy--proline ethyl ester (3.50

g, 13.4 mmol) in CH₃OH (30 mL) and the mixture was then

stirred at room temperature for 5 h. Evaporation of solvent

under reduced pressure and subsequent puriﬁcation by ﬂash

chromatography, using ethyl acetate as eluant, gave the title

compound (2.90 g, 91%) as a viscous transparent oil. R_f0.35

(3 : 1 ethyl acetate–hexane); [α]_Dϩ67.4 (c = 2.0, EtOH) [lit.⁴²

[α]_Dϩ70.4 (c = 2.0, EtOH)]; δ_H(300 MHz, CDCl₃) 1.28 (3H, t,

J 7.2, CH₂CH₃), 1.41 and 1.46 (total 9H, 2 × s, C(CH₃)₃

rotamers), 2.00–2.17 (1H, m, H_a3), 2.18–2.65 (2H, m, H_b3 and

OH), 3.40–3.71 (2H, m, H5), 4.17 and 4.18 (total 2H, 2 × q,

J 7.2, CH₂CH₃rotamers), 4.32–4.44 (1H, m, H2), 4.45–4.53

(1H, m, H4); δ_C(100.6 MHz, CDCl₃) 14.6 (CH₂CH₃), 28.6

(C(CH₃)₃), 38.8 and 39.5 (C3 rotamers), 55.1 (C5), 58.1 and

58.4 (C2 rotamers), 61.4 (CH₂CH₃), 69.7 and 70.5 (C4 rota-

angles (P), cyclic torsion angles ν₀

ν₄, and the torsion angle

C3Ј–C4Ј–N1–C2 (χ for conformers A–D) are shown in ESI

(Table 1). The structures of conformers A and B are shown in

Fig. 4.

NMR conformational analysis

t

NMR spectra of T₂-POMؒ3HCl 30 (ca. 4 mg) dissolved in

99.9% D₂O (0.5 ml) were recorded at 500 MHz on a Bruker

DRX500 at 30 ЊC and referenced to the residual HDO peak at

mers), 80.7 (C(CH₃)₃), 154.5 (CO₂Bu), 173.3 and 173.5 (CO₂Et

rotamers); m/z (FAB^ϩ) 260 ([M ϩ H]^ϩ, 50%), 204 (96), 160

(100); HRMS m/z (CI) 260.1501 ([M ϩ H]^ϩ, C₁₂H₂₂NO₅

requires m/z, 260.1498).

1

4.70 ppm. A H spectral width of 5000 Hz was used through-

out. 2D TOCSY, COSY 45, ROESY and J-resolved spectra

were collected using 8k data points in t₂and 256 t₁increments,

typically zero-ﬁlled to 8k × 512 data points, with a relaxation

delay of 2 seconds. Suppression of residual solvent signal

(HDO) was achieved using presaturation during the relaxation

delay. ROESY and TOCSY spectra had typical mixing times

of 300 and 60 ms respectively. The assignment of all proton

chemical shifts, vicinal coupling constants, and ROESY cross

cis-4Ј-(N³-Benzoylthymin-1-yl)-N-(tert-butoxycarbonyl)-D-

proline ethyl ester (11). Diethyl azodicarboxylate (DEAD)

(1.0 mL, 6.37 mmol) was added dropwise over 1 h to a solution

of N-(tert-butoxycarbonyl)-trans-4-hydroxy--proline ethyl

ester (1.50 g, 5.79 mmol), N ³-benzoylthymine³³(1.33 g, 5.79

mmol) and Ph₃P (1.72 g, 6.37 mmol) in THF (50 mL) at

Ϫ15 ЊC. The mixture was warmed to room temperature and

stirred for 18 h. Evaporation under reduced pressure and puriﬁ-

cation by column chromatography, eluting with 5% acetone in

CH₂Cl₂and then 5% CH₃OH in CH₂Cl₂, gave the ester 11 (1.69

g, 62%) as a white foam. R_f0.70 (5 : 1 CH₂Cl₂–acetone); [α]_D

ϩ9.9 (c = 2.0, CH₂Cl₂); ν_max(KBr)/cm^Ϫ11731 and 1692 (CO);

λ_max(CH₃OH)/nm 270 (ε/dm³mol^Ϫ1cm^Ϫ111000); δ_H(300 MHz,

CDCl₃) 1.24 (3H, t, J 7.1, CH₂CH₃), 1.35 and 1.39 (total 9H,

2 × s, C(CH₃)₃), 1.92 (3H, d, J 0.8, thymine CH₃), 1.97–2.14

(1H, m, H_a3Ј), 2.65–2.83 (1H, m, H_b3Ј), 3.38–3.71 (1H, m,

H_a5Ј), 3.90 (1H, dd, J 11.9, 7.7, H_b5Ј), 4.17 (2H, q, J 7.1,

3

peaks are given in ESI. The observed J_HHcoupling constants

for the tri-substituted (upper) pyrrolidine ring of T₂-POMؒ

3

3HCl 30 (Table 1) were compared with the calculated J_HH

coupling constants for the corresponding vicinal protons for

the lowest energy conformers A–D of the model cis- and

trans-dimethylpyrrolidines 8 and 9 (Fig. 3). The calculated ³J_HH

coupling constants are derived from the general Karplus equa-

tion J = a_jcos²ꢀ ϩ b_jcos ꢀ ϩ c_j.⁴⁵The Karplus parameters

j

used in the calculations, a_j, b_jand c_j, are listed in Table 1 and are

the same as those derived for -4-hydroxyproline.⁴⁵

O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 2 7 7 – 3 2 9 2

3286

View Article Online

CH₂CH₃), 4.22–4.43 (1H, m, H2Ј), 5.18 and 5.35 (total 1H,

2 × br s, H4Ј rotamers), 7.31 (1H, s, H6), 7.42 (2H, t, J 7.5, Bz

m-H), 7.58 (1H, t, J 7.5, Bz p-H), 7.84 (2H, d, J 7.1, Bz o-H);

δ_C(100.6 MHz, CDCl₃) 12.5 and 13.0 (thymine CH₃rotamers),

14.6 (CH₂CH₃), 28.6 (C(CH₃)₃), 36.1 and 36.5 (C3Ј rotamers),

49.9 (C5Ј), 52.8 (C4Ј), 57.9 (C2Ј), 61.6 and 62.0 (CH₂CH₃

rotamers), 81.6 (C(CH₃)₃), 112.1 (C5), 129.1 and 129.5 (Bz m-C

rotamers), 130.8 (Bz o-C), 132.0 (Bz ipso-C), 135.4 (C6), 136.4

mmol) in dry DMF (1 mL) and the mixture was stirred at 80 ЊC

for 16 h. The solution was allowed to cool to room temperature

whereupon saturated brine (10 mL) was added. The mixture

was extracted with ethyl acetate (4 × 10 mL), dried over

anhydrous MgSO₄, ﬁltered and evaporated under reduced

pressure. Puriﬁcation by column chromatography (10%

CH₃OH in CH₂Cl₂) gave the enamine 18 (38 mg, 72% based on

unrecovered starting material) as a white crystalline solid. R_f

0.31 (10% CH₃OH in CH₂Cl₂); δ_H(300 MHz, CDCl₃) 1.47 and

1.49 (total 9H, 2 × s, C(CH₃)₃rotamers), 1.76 (1H, d, J 13.6,

H_a3Ј), 1.90 (3H, s, thymine CH₃), 2.26–2.48 (1H, m, H_b3Ј), 3.03

(1H, t, J 13.2, NHCH_aH_b), 3.46 (1H, t, J 12.4, H_a5Ј), 3.62–3.87

(2H, m, H_b5Ј and NHCH_aH_b), 4.20–4.42 (1H, m, H2Ј

rotamers), 5.96 (1H, t, J 7.4, H4Ј), 9.37 (1H, br s, NH);

δ_C(100.6 MHz, CDCl₃) 9.4 (thymine CH₃), 28.8 (C(CH₃)₃),

37.7 and 38.3 (C3Ј rotamers), 50.0 and 50.9 (NHCH₂rotamers),

51.6 and 51.9 (C5Ј rotamers), 53.1 and 54.1 (C4Ј rotamers), 54.2

and 54.6 (C2Ј rotamers), 80.5 (C(CH₃)₃), 91.3 and 91.6 (C5

rotamers), 151.2 (C2), 153.6 and 153.9 (C6 rotamers), 154.3

and 154.5 (Boc CO rotamers), 163.2 (C4); m/z (FAB^ϩ) 345

([M ϩ Na]^ϩ, 17%), 323 ([M ϩ H]^ϩ, 100), 267 (92); HRMS

m/z (FAB^ϩ) 323.1702 ([M ϩ H]^ϩ, C₁₅H₂₃N₄O₄requires m/z,

323.1719).

t

(Bz p-C), 150.3 (C2), 153.8 and 154.1 (CO₂Bu rotamers), 162.2

and 162.8 (C4 rotamers), 169.2 (COPh), 172.8 (CO₂Et); m/z

(FAB^ϩ) 472 ([M ϩ H]^ϩ, 20%), 416 (36), 372 (20); HRMS m/z

(CI) 472.2081 ([M ϩ H]^ϩ, C₂₄H₃₀N₃O₇requires m/z, 472.2083).

N-(tert-Butoxycarbonyl)-cis-4Ј-(thymin-1-yl)-D-prolinol (12).

Lithium borohydride (49 mg, 2.24 mmol) was added portion-

wise over 1 h to a solution of ethyl ester 11 (0.53 g, 1.12 mmol)

in dry THF (10 mL) at 0 ЊC. Stirring was continued at 0 ЊC for a

further 30 min, then the mixture was allowed to warm to room

temperature and stirred overnight. The reaction was cooled to

0 ЊC, quenched by the addition of methanol (20 mL), and evap-

orated under reduced pressure. Saturated aq. NaCl (20 mL) was

then added and the mixture was extracted with CHCl₃–EtOH

(2 : 1 v/v, 6 × 20 mL), dried over anhydrous MgSO₄, ﬁltered and

evaporated under reduced pressure. Puriﬁcation by column

chromatography, eluting with 5% CH₃OH in CH₂Cl₂, gave the

alcohol 12 (0.250 g, 69%) as a white solid. R_f0.12 (5% CH₃OH

in CH₂Cl₂); mp 79–81 ЊC (from hexane–ethyl acetate); [α]_D

ϩ10.6 (c = 2.0, CH₂Cl₂); ν_max(KBr)/cm^Ϫ13412 and 3188 (OH),

1689 (CO); λ_max(CH₃OH)/nm 270 (ε/dm³mol^Ϫ1cm^Ϫ18800);

δ_H(300 MHz, CDCl₃) 1.49 (9H, s, C(CH₃)₃), 1.96–2.00 (4H, m,

thymine CH₃and H_a3Ј), 2.40–2.49 (1H, m, H_b3Ј), 3.27 (1H, dd,

J 10.9, 9.8, H_a5Ј), 3.69 (1H, dd, J 11.3, 5.6, CH_aH_bOH), 3.92

(1H, d, J 10.9, H_a5Ј), 3.98–4.10 (2H, m, CH_aH_bOH and H2Ј),

5.05–5.11 (1H, m, H4Ј), 7.13 (1H, s, H6), 8.25 (1H, s, H3);

δ_C(67.9 MHz, CDCl₃) 12.5 (thymine CH₃), 28.3 (C(CH₃)₃),

32.2 (C3Ј), 49.8 (C5Ј), 51.8 (C4Ј), 58.6 (C2Ј), 65.5 (CH₂OH),

81.2 (C(CH₃)₃), 111.7 (C5), 135.9 (C6), 151.1 (C2), 155.9

N-(tert-Butoxycarbonyl)-(2ЈR,4ЈR)-2Ј-[(phthalimido)methyl]-

4Ј-(thymin-1-yl)pyrrolidine 13. PPh₃(833 mg, 3.18 mmol) and

phthalimide (470 mg, 3.18 mmol) were added to a solution of

alcohol 12 (794 mg, 2.44 mmol) in dry THF (20 mL). The

mixture was stirred at Ϫ15 ЊC and DEAD (500 µL, 3.18 mmol)

was added dropwise over 20 min. After a further 30 min the

reaction was warmed to room temperature and stirred for 18 h.

Evaporation under reduced pressure and subsequent puriﬁ-

cation by column chromatography (ethyl acetate) gave the

phthalimide derivative 13 (690 mg, 63%) as a white solid. R_f

0.3 (ethyl acetate); mp 102–105 ЊC (from hexane–ethyl acetate);

[α]_DϪ54.0 (c = 1.56, CH₂Cl₂); ν_max(CHCl₃)/cm^Ϫ13686, 3396,

1774 (Phth CO), 1718 and 1692 (CO); δ_H(270 MHz, CDCl₃)

1.33 (9H, br s, C(CH₃)₃), 1.96 (3H, d, J 0.6, thymine CH₃),

1.92–2.05 (1H, m, H_a3Ј), 2.43–2.51 (1H, m, H_b3Ј), 3.41 (1H,

dd, J 11.9, 7.4, H_a5Ј), 3.71–3.77 (1H, m, H_b5Ј), 3.88–

4.09 (2H, m, Phth-CH₂), 4.22–4.42 (1H, m, H2Ј), 4.93–4.99

(1H, m, H4Ј), 7.28 (1H, s, H6), 7.66–7.74 (2H, m, Ar), 7.79–

7.86 (2H, m, Ar), 9.26 (1H, s, NH); δ_C(67.9 MHz, CDCl₃)

12.4 (thymine CH₃), 28.2 and 28.3 (C(CH₃)₃rotamers), 33.9

(C3Ј), 41.6 (Phth-CH₂), 48.5 and 48.6 (C5Ј rotamers), 53.4

(C4Ј), 54.4 (C2Ј), 80.8 (C(CH₃)₃), 111.6 (C5), 123.3 and 134.0

(ArCH), 132.1 (ArC ipso), 136.4 (C6), 150.9 (C2), 154.4

(Boc CO), 163.5 (C4), 168.4 (Phth CO); m/z (FAB^ϩ) 477

([M ϩ Na]^ϩ, 5%), 455 ([M ϩ H]^ϩ, 8), 400 (7), 355(100); HRMS

m/z (FAB^ϩ) 455.1924 ([M ϩ H]^ϩ, C₂₃H₂₇N₄O₆requires m/z,

455.1931).

(CO₂Bu), 163.4 (C4); m/z (FAB^ϩ) 348 ([M ϩ Na]^ϩ, 25%), 326

t

([M ϩ H]^ϩ, 43), 270 (64), 226 (100); HRMS m/z (FAB^ϩ)

326.1717 ([M ϩ H]^ϩ, C₁₅H₂₄N₃O₅requires m/z, 326.1716).

N-tert-Butoxycarbonyl-cis-4Ј-(thymin-1-yl)-D-prolinol

p-toluenesulfonate (15). Tosyl chloride (121 mg, 0.63 mmol) was

added to a solution of alcohol 12 (103 mg, 0.32 mmol) in dry

pyridine (1.0 mL) at 0 ЊC and the mixture was stirred for 15

min, allowed to warm to room temperature and stirred for a

further 18 h. CH₃OH (10 mL) and water (5 mL) were slowly

added and the mixture was then evaporated under reduced

pressure. Puriﬁcation by ﬂash chromatography (0 10%

CH₃OH in CH₂Cl₂) gave the tosyl derivative 15 (150 mg, 99%)

as a white foam. [α]_Dϩ5.8 (c = 0.12, CH₂Cl₂); ν_max(KBr)/cm^Ϫ1

1700 br (C᎐O), 1400 (SO and C(CH ) ), 1385 (C(CH ) ), 1165

᎐

2

3

3 3

(SO₂), 807 (Ar-H); δ_H(270 MHz, CDCl₃) 1.32 (9H, s, C(CH₃)₃),

1.91 (3H, s, thymine-CH₃), 1.93–2.14 (1H, m, H_a3Ј), 2.38 (3H, s,

Ar-CH₃), 2.45–2.58 (1H, m, H_b3Ј), 3.05–3.20 (1H, m, H_a5Ј),

3.86–3.94 (1H, m, H_b5Ј), 4.00–4.15 (2H, m, CH₂OSO₂), 4.22–

4.55 (1H, m, H2Ј), 5.00–5.23 (1H, m, H4Ј), 7.19 (1H, br s, H6),

7.36 (2H, d, J 8.3, Ar), 7.78 (2H, d, J 8.3, Ar), 9.34 (1H, s, NH);

δ_C(67.9 MHz, CDCl₃) 12.5 (thymine CH₃), 21.6 (Ar CH₃), 28.4

(C(CH₃)₃), 31.5 (C3Ј), 49.4 (C5Ј), 51.4 (C4Ј), 54.5 (C2Ј), 70.3

(CH₂OSO₂), 80.9 (C(CH₃)₃), 112.3 (C5), 127.8 and 130.1 (Ar

CH), 132.7 (Ar CCH₃), 135.7 (C6), 145.3 (Ar CSO₂), 150.9

(C2), 153.5 (Boc CO), 163.1 (C4); m/z (FAB^ϩ) 612 ([M ϩ Cs]^ϩ,

27%), 394 (100), 345 (43), 312 (47), 286 (90); HRMS m/z

(FAB^ϩ) 612.0792 ([M ϩ Cs]^ϩ, C₂₂H₂₉N₃O₇SCs requires m/z,

612.0781).

N-(tert-Butoxycarbonyl)-(2ЈR,4ЈR)-2Ј-[(amino)methyl]-4Ј-

(thymin-1-yl)pyrrolidine (14). A solution of phthalimide deriv-

ative 13 (340 mg, 0.88 mmol) in a 25–30% aqueous solution of

methylamine (2 mL) was heated at 40 ЊC for 1 h. Evaporation

under reduced pressure followed by column chromatography

(20% CH₃OH in CH₂Cl₂) gave the amine 14 (253 mg, 89%) as a

white solid. R_f0.28 (30% CH₃OH in CH₂Cl₂); mp 116–118 ЊC

(from CH₃OH–CH₂Cl₂); [α]_Dϩ22.1 (c = 0.54, CH₃OH); ν_max

(CHCl₃)/cm^Ϫ13680, 3390, 3020, 1690 (CO), 1470; δ_H(270 MHz,

CD₃OD) 1.47 (9H, s, C(CH₃)₃), 1.89 (3H, d, J 1.1, thymine

CH₃), 2.09–2.23 (1H, m, H_a3Ј), 2.33–2.46 (1H, m, H_b3Ј), 2.94

(2H, d, J 4.9, CH₂NH₂), 3.26–3.37 (1H, m, H_a5Ј), 3.81–4.00

(2H, m, H_b5Ј and CH2Ј), 4.68–4.92 (1H, m, CH4Ј), 7.57 (1H, d,

J 1.1, H6); δ_C(67.9 MHz, CD₃OD) 12.3 (thymine CH₃), 28.7

(C(CH₃)₃), 33.7 (C3Ј), 45.3 (CH₂NH₂), 50.1 (C5Ј), 54.4 (C4Ј),

59.1 (C2Ј), 81.7 (C(CH₃)₃), 111.8 (C5), 139.5 (C6), 153.0 (C2),

156.4 (Boc CO), 166.3 (C4); m/z (FAB^ϩ) 455 ([M ϩ Cs]^ϩ, 13%),

347 ([M ϩ Na]^ϩ, 20), 325 ([M ϩ H]^ϩ, 100), 225 (96), 175 (43);

N-(tert-Butoxycarbonyl)-(2ЈR,4ЈR)-6,2Ј-imino(methyl)-4Ј-

(thymin-1-yl)pyrrolidine (18). Lithium azide (45 mg, 0.93 mmol)

was added to a solution of tosyl derivative 15 (111 mg, 0.23

O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 2 7 7 – 3 2 9 2

3287

View Article Online

HRMS m/z (FAB^ϩ) 325.1864 ([M ϩ H]^ϩ, C₁₅H₂₅N₄O₄requires

m/z, 325.1876).

C(CH₃)₃), 1.71 and 1.75 (both 3H, s, thymine CH₃), 1.83–1.93

(1H, m, H_a3Ј), 2.09–2.25 (1H, m, H_b3Ј), 2.36–2.46 (2H, m,

H3Ј), 2.60 (1H, dd, J 11.1, 6.1, H5Ј), 2.86–2.94 (2H, m, 2 ×

H2Ј), 3.19–3.27 (1H, m, H5Ј), 3.33 (1H, t, J 9.9, H5Ј), 3.47–3.54

(2H, m, CH₂CONH), 3.71 (1H, dd, J 14.4, 2.3, Phth-CH_aH_b),

3.81 (1H, dd, J 14.6, 4.8, Phth-CH_aH_b), 3.85 (1H, dd, J 10.6,

3.5, CONHCH_aH_b), 3.89 (1H, s, H5Ј), 3.92–3.99 (1H, m,

CONHCH_aH_b), 4.59–4.66 and 4.69–4.74 (each 1H, m, H4Ј),

7.45 (1H, s, H6), 7.67 (4H, br, Ar), 7.70 (1H, s, H6); δ_C(100.6

MHz, CD₃OD) 12.9 (2 × thymine CH₃, coincident), 29.2

(C(CH₃)₃), 34.2 and 37.6 (2 × C3Ј), 40.3 (Phth-CH₂), 43.3

(CH₂CONH), 50.0 (C5Ј), 55.5 (C2Ј), 57.6 (C4Ј), 57.9 (C5Ј),

59.1 (C4Ј), 64.0 (C2Ј), 82.1 (C(CH₃)₃), 111.4 and 112.0 (2 ×

C5), 124.7, 133.5 and 136.1 (Ar), 140.3 and 140.6 (2 × C6),

153.2 and 153.3 (2 × C2), 156.8 (Boc CO), 166.5 and 166.6 (2 ×

C4), 170.8 (Phth CO), 173.8 (CH₂CONH); m/z (FAB^ϩ) 741 ([M

ϩ Na]^ϩ, 9%), 447 ([M ϩ H]^ϩ, 10), 719 ([M ϩ H]^ϩ, 100), 619

(24), 460 (86); HRMS m/z (FAB^ϩ) 719.3176 ([M ϩ H]^ϩ,

C₃₅H₄₃N₈O₉requires m/z, 719.3153).

(2ЈR,4ЈR)-2Ј-[(Phthalimido)methyl]-4Ј-(thymin-1-yl)pyrrol-

idine triﬂuoroacetic acid salt (19). CF₃CO₂H (2.0 mL) was

added to a solution of phthalimide derivative 13 (830 mg, 1.83

mmol) in dichloromethane (4.0 mL). After stirring at room

temperature for 4 h the solution was reduced in volume under a

stream of argon to give a thick oil. On addition of diethyl ether

(5.0 mL) and cooling to 0 ЊC, a precipitate formed which was

ﬁltered and dried to give the desired amine salt 19 (740 mg,

86%) as a white solid. [α]_DϪ46.3 (c = 0.19, 10% CH₃OH in

CH₂Cl₂); λ_max(EtOH)/nm 219 (ε/dm³mol^Ϫ1cm^Ϫ145050); δ_H

(400 MHz, DMSO-d6) 1.79 (3H, s, thymine CH₃), 1.99–2.03

(1H, m, H_a3Ј), 2.54–2.64 (1H, m, H_b3Ј), 3.37–3.59 (2H, m,

H5Ј), 3.80–3.86 (1H, m, H2Ј), 3.93–4.04 (2H, m, Phth-CH₂),

5.14–5.22 (1H, br m, H4Ј), 7.66 (1H, d, J 1.0, H6), 7.85–7.95

(4H, m, Ar), 8.90 and 9.80 (each 1H, br s, NH1Ј), 11.42 (1H, s,

thymine NH); δ_C(100.6 MHz, DMSO-d6) 12.1 (thymine CH₃),

32.9 (C3Ј), 37.7 (C5Ј), 48.0 (Phth-CH₂), 53.6 (C4Ј), 58.2 (C2Ј),

109.7 (C5), 117.1 (q, J 299.2, CF₃), 123.2, 131.7 and 134.5 (3 ×

Ar), 138.6 (C6), 151.2 (C2), 158.4 (q, J 31.3, CF₃CO), 163.7

(C4), 167.9 (Phth CO); m/z (FAB^ϩ) 355 (M^ϩ, 100%), 307 (12),

176 (14); HRMS m/z (FAB^ϩ) 355.1393 (M^ϩ, C₁₈H₁₉N₄O₄

requires m/z, 355.1406).

N-tert-Butoxycarbonyl-(2ЈR,4ЈR)-2Ј-[(hydroxyacetamido)-

methyl]-4Ј-(thymin-1-yl)pyrrolidine (23). δ_H(400 MHz, CD₃OD)

1.40 (9H, s, C(CH₃)₃), 1.80 (3H, d, J 1.3, thymine CH₃), 1.95–

2.09 (1H, m, H_a3Ј), 2.28–2.39 (1H, m, H_b3Ј), 3.26–3.32 (1H, m,

H_a5Ј), 3.48 (2H, br s, NHCH₂), 3.86 (1H, dd, J 10.9, 8.1, H_b5Ј),

3.90 (2H, s, CH₂OH), 3.93–3.98 (1H, m, H2Ј), 4.72 (1H, m,

H4Ј), 7.39 (1H, d, J 1.0, H6), 8.07 (1H, br s, CH₂CONH ); δ_C

(100.6 MHz, CD₃OD) 12.8 (thymine CH₃), 29.1 (C(CH₃)₃),

34.1 (C3Ј), 43.4 (C5Ј), 43.6 (Phth-CH₂), 55.2 (C4Ј), 57.4 (C2Ј),

63.1 (CH₂OH), 82.3 (C(CH₃)₃), 112.2 (C5), 139.9 (C6), 153.3

(C2), 156.8 (Boc CO), 166.7 (C4), 176.1 and 176.1 (CH₂CONH

rotamers); m/z (FAB^ϩ) 383 ([M ϩ H]^ϩ, 12%), 283 (100); HRMS

m/z (CI) 383.1934 ([M ϩ H]^ϩ, C₁₇H₂₇N₄O₆requires m/z,

383.1930).

N-(tert-Butoxycarbonyl)-(2ЈR,4ЈR)-2Ј-[(bromoacetamido)-

methyl]-4Ј-(thymin-1-yl)pyrrolidine (20). To a 65% (w/v) solu-

tion of bromoacetic anhydride (13 µL, 0.031 mmol) in

acetonitrile was added dichloromethane (500 µL). This was

cooled to Ϫ8 ЊC and a solution of amine 14 (10 mg, 0.031

mmol) in dichloromethane (1 mL) was added. After 5 min, the

ﬂask was warmed to room temperature and evaporated under

reduced pressure. Column chromatography (5 10% CH₃OH

in CH₂Cl₂) gave the bromoacetamide 17 (13.4 mg, 97%) as a

white solid. R_f0.29 (5% CH₃OH in CH₂Cl₂); mp 90–92 ЊC (from

CH₃OH–CH₂Cl₂); [α]_DϪ38.1 (c = 0.16, CH₃OH); λ_max(EtOH)/

nm 270.7 (ε/dm³mol^Ϫ1cm^Ϫ120852); δ_H(400 MHz, CD₃OD)

1.49 (9H, s, C(CH₃)₃), 1.88 (3H, d, J 1.1, thymine CH₃), 2.04–

2.22 (1H, m, H_a3Ј), 2.34–2.48 (1H, m, H_b3Ј), 3.27–3.41 (1H, m,

H_a5Ј), 3.53–3.64 (2H, m, CH₂Br), 3.87 (2H, d, J 2.4, NHCH₂),

3.88–4.12 (2H, m, H_b5Ј and H2Ј), 4.72–4.96 (1H, m, H4Ј), 7.45

(1H, d, J 1.1, H6); δ_C(67.9 MHz, CD₃OD) 12.5 (thymine CH₃),

28.7 (C(CH₃)₃), 29.1 (CH₂Br), 33.4 (C3Ј), 43.3 (CH₂NH), 50.0

C(5Ј), 54.7 (C4Ј), 57.1 (C2Ј), 81.8 and 81.9 (C(CH₃)₃rotamers),

111.7 (C5), 139.4 and 139.5 (C6 rotamers), 152.8 (C2), 156.2

(Boc CO), 166.2 (C4), 170.1 (BrCH₂CO); m/z (FAB^ϩ) 469 ([M

ϩNa]^ϩ,5%)‡,467([MϩNa]^ϩ,4)§,447([MϩH]^ϩ,10)‡,445([M

ϩ H]^ϩ, 10)§, 347 (100), 345 (98), 314 (84); HRMS m/z (FAB^ϩ)

445.1090 ([M ϩ H]^ϩ, C₁₇H₂₆N₄O₅⁷⁹Br requires m/z, 445.1087).

Phth-T₂-Boc-POM (21) via optimised N-alkylation. DIPEA

(0.33 mL, 1.89 mmol) was added to a solution of amine 19 (300

mg, 0.63 mmol) and bromoacetamide 20 (280 mg, 0.63 mmol)

in dry DMF (2 mL) and the mixture stirred under argon at

room temperature for 18 h. Evaporation under reduced pres-

sure gave a residue to which was added brine (10 mL). The

mixture was extracted with CH₂Cl₂(4 × 10 mL), dried over

anhydrous MgSO₄, ﬁltered, evaporated and then puriﬁed by

column chromatography (10% CH₃OH in CH₂Cl₂) to give

Phth-T₂-Boc-POM 21 (440 mg, 98%) as a white crystalline

solid.

Phth-T₂-Boc-POM (21) via N-acylation. Amine 14 (18 mg,

0.056 mmol) was added to a solution of pentaﬂuorophenyl

ester 26 vide infra (32 mg, 0.056 mmol) in CH₂Cl₂(2 mL) and

the mixture stirred at room temperature for 3 h. Evaporation

under reduced pressure and puriﬁcation by column chromato-

graphy (5 10% CH₃OH in CH₂Cl₂) gave Phth-T₂-Boc-POM

21 (40 mg, 100%) as a white crystalline solid.

Phth-T₂-Boc-POM (21)

Triethylamine (24 µL, 0.172 mmol) was added to a solution of

amine 19 (20 mg, 0.043 mmol) and bromoacetamide 20 (19 mg,

0.043 mmol) in dry DMF (0.5 mL). This mixture was stirred at

room temperature for 16 h, then evaporated under reduced

pressure. A 10% (w/v) K₂CO₃solution in saturated brine (5 mL)

was added and the mixture was extracted with ethyl acetate

(4 × 5 ml). The combined extracts were dried over anhydrous

MgSO₄, ﬁltered and evaporated under reduced pressure. Col-

umn chromatography (10% CH₃OH in CH₂Cl₂) gave hydroxy-

acetamide 23 (8 mg, 49%) and Phth-T₂-Boc-POM 21 (15 mg,

49%) as a white crystalline solid. R_f0.46 (10% CH₃OH in

CH₂Cl₂); mp 122–125 ЊC (from CH₃OH–CH₂Cl₂); [α]_DϪ70.83

(c = 0.12, CH₃OH); ν_max(CHCl₃)/cm^Ϫ13690, 3410, 3024, 1780

(Phth CO), 1700 (CO); δ_H(400 MHz, CD₃OD) 1.37 (9H, s,

(2ЈR,4ЈR)-N-[(tert-Butoxycarbonyl)methyl]-2Ј-[(phthal-

imido)methyl]-4Ј-(thymin-1-yl)pyrrolidine (24). DIPEA (74 µL,

0.428 mmol) and then tert-butyl bromoacetate (35 µL, 0.214

mmol) was added to a solution of amine 19 (50 mg, 0.141

mmol) in dry DMF (0.5 mL) at 0 ЊC. After 5 min the reaction

was warmed to room temperature and stirred for a further 18 h.

Evaporation under reduced pressure followed by column

chromatography (10% CH₃OH in CH₂Cl₂) gave the ester 24 (46

mg, 92%) as a white powder, mp 80–82 ЊC (from CH₃OH–

CH₂Cl₂); [α]_DϪ9.56 (c = 0.46, CH₂Cl₂); ν_max(CHCl₃)/cm^Ϫ1

3690, 1770 (Pht CO), 1734, 1715, 1698 and 1684 (C᎐O); δ (300

᎐

H

MHz, CDCl₃) 1.34 (3H, s, thymine CH₃), 1.42 (9H, s, C(CH₃)₃),

1.80–1.93 (1H, m, H_a3Ј), 2.43–2.57 (1H, m, H_b3Ј), 2.61–2.73

(1H, m, H_a5Ј), 2.90–2.96 (1H, m, H2Ј), 3.01 (1H, d, J 17.0,

‡ As a consequence of ⁸¹Br.

§ As a consequence of ⁷⁹Br.

t

CH_aH_bCO₂Bu), 3.29 (1H, d, J 10.4, H_b5Ј), 3.72–3.79 (2H, m,

O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 2 7 7 – 3 2 9 2

3288

View Article Online

t

Phth-CH₂), 3.83 (1H, d, J 17.0, CH_aH_bCO₂Bu), 4.78–4.88 (1H,

NCH_aH_bCO), 5.30 (1H, m, H4Ј), 6.78 (1H, br s, CH₂CONH ),

m, H4Ј), 7.62–7.70 and 7.71–7.79 (each 2H, m, Ar), 7.81 (1H, s,

H6), 8.97 (1H, s, thymine NH); δ_C(67.9 MHz, CDCl₃) 12.0

(thymine CH₃), 28.1 (C(CH₃)₃), 37.0 (C3Ј), 39.0 (Phth-CH₂),

52.4 (C4Ј), 54.0 (CH₂CO₂Bu), 57.8 (C5Ј), 61.2 (C2Ј), 81.3

(C(CH₃)₃), 110.6 (C5), 123.3, 131.8 and 134.2 (3 × Ar), 137.9

7.48 (1H, d, J 1.5, H6); δ_C(75.5 MHz, CDCl₃) 13.1 (thymine

CH₃), 37.0 (C3Ј), 46.8 (NHCH₂), 52.5 (C4Ј), 56.4 (NCH₂), 59.1

(C2Ј), 60.3 (C5Ј), 112.7 (C5), 137.1 (C6), 151.4 (C2), 164.2 (C4),

169.5 (CH₂CONH); m/z (FAB^ϩ) 265 ([M ϩ H]^ϩ, 100%), 176

(20); HRMS m/z (EI) 264.1225 (M^ϩ, C₁₂H₁₆N₄O₃requires m/z,

264.1222).

t

(C6), 150.9 (C2), 163.5 (C4), 168.4 (Phth CO) and 169.6

t

(CO₂Bu); m/z (FAB^ϩ) 469 ([M ϩ H]^ϩ, 25%), 413 (44), 314

(100), 254 (85); HRMS m/z (FAB^ϩ) 469.2103 ([M ϩ H]^ϩ,

C₂₄H₂₉N₄O₆requires m/z, 469.2087).

T₂-Boc-POM (27)

A solution of Phth-T₂-Boc-POM 21 (83 mg, 0.116 mmol) in

40% aqueous methylamine (1 mL) was stirred at room temper-

ature for 40 min. Evaporation under reduced pressure then col-

umn chromatography (10 15% CH₃OH in CH₂Cl₂with 0.5%

DIPEA) separated the desired amine 27 at lower R_ffrom the

intermediate product at higher R_f(assumed to be 29). The latter

29 was treated with 40% aqueous methylamine (1 mL) for a

further 30 min; evaporation and puriﬁcation as before gave

H₂N-T₂-Boc-POM 27 as a white foam (combined yield of 51

mg, 75%). δ_H(300 MHz, CD₃OD) 1.50 (9H, br s, C(CH₃)₃), 1.90

and 1.91 (each 3H, s, thymine CH₃), 2.27–2.65 (4H, m), 2.73–

2.94 (2H, m), 2.94–3.18 (3H, m), 3.19–3.34 (1H, m), 3.44–3.79

(4H, m), 3.85–4.15 (2H, m), 4.63–4.82 and 4.96–5.11 (each 1H,

m, H4Ј), 7.48 and 7.83 (each 1H, s, H6); δ_C(75 MHz, CD₃OD)

12.7 and 12.9 (2 × thymine CH₃), 24.4 (C3Ј), 29.1 (C(CH₃)₃),

33.5, 36.3, 41.6, 43.5, 45.4, 54.4, 57.6, 57.8, 60.0, 64.2 (2 × C1Ј,

2 × C2Ј, C3Ј, 2 × C4Ј, 2 × C5Ј and NCH₂CONH), 82.2 and

82.5 (C(CH₃)₃rotamers), 112.0 and 112.4 (2 × C5), 140.1 and

140.4 (2 × C6), 153.2 and 153.3 (2 × C2), 166.8 and 166.8 (2 ×

C4), 174.2 (CH₂CONH); m/z (FAB^ϩ) 627 ([M ϩ K]^ϩ, 3%), 511

(5), 345 (60); HRMS m/z (FAB^ϩ) 627.2893 ([M ϩ K]^ϩ,

C₂₇H₄₀N₈O₇K requires m/z, 627.2657).

N-Carboxymethyl-(2ЈR,4ЈR)-2Ј-[(phthalimido)methyl]-4Ј-

(thymin-1-yl)pyrrolidine triﬂuoroacetic acid salt (25). Tri-

ﬂuoroacetic acid (0.5 mL) was added to a solution of tert-butyl

ester 24 (72 mg, 0.154 mmol) in dichloromethane (2.0 mL) and

the mixture was stirred for 3 h at room temperature. Evapor-

ation under reduced pressure followed by column chromato-

graphy (5 15% CH₃OH in CH₂Cl₂) gave the acid 25 (81 mg,

100%) as a white foam. R_f0.62 (20% CH₃OH in CH₂Cl₂); [α]_D

Ϫ9.0 (c = 0.5, CH₃OH); δ_H(300 MHz, CD₃OD) 1.90–2.10 (1H,

m, H_a3Ј), 2.42–2.55 (1H, m, H_b3Ј), 2.60–2.76 (1H, m, H_a5Ј),

2.93–3.15 (1H, m, H2Ј), 3.04 (1H, d, J 17.0, CH_aH_bCO₂H),

3.50–3.60 (1H, m, H_b5Ј), 3.65–3.75 (1H, m, Phth-CH_aH_b), 3.85

(1H, dd, J 14.7, 4.9, Phth-CH_aH_b), 3.96 (1H, d, J 17.0, CH_aH_b-

CO₂H), 4.48–4.57 (1H, m, H4Ј), 7.67 (4H, s, Ar), 7.95 (1H, s,

H6); δ_C(75.5 MHz, CD₃OD) 12.5 (thymine CH₃), 37.6 (C3Ј),

39.6 (Phth-CH₂), 56.0 (CH₂CO₂H), 56.2 (C4Ј), 58.5 (C5Ј), 63.6

(C2Ј), 111.2 (C5), 117.0 (q, J 244.1, CF₃), 124.6, 133.6 and

135.9 (Ar), 141.5 (C6), 153.4 (C2), 163.1 (q, J 88.6, CF₃CO₂),

166.6 (C4), 170.6 (Phth CO), 176.7 (CO₂H); m/z (FAB^ϩ) 545

([M ϩ Cs]^ϩ, 6%), 435 ([M ϩ Na]^ϩ, 10), 423 ([M ϩ H]^ϩ, 24), 392

(20), 344 (100); HRMS m/z (FAB^ϩ) 413.1443 ([M ϩ H]^ϩ,

C₂₀H₂₁N₄O₆requires m/z, 413.1461).

T₂-POMؒ3HCl (30)

N-(Pentaﬂuorophenoxycarbonyl)methyl-(2ЈR,4ЈR)-2Ј-

T₂-Boc-POM 27 (51 mg, 0.087 mmol) was dissolved in a solu-

tion of 1 : 1 THF–CH₃OH (1 mL) saturated with HCl gas.

After stirring for 15 minutes a white precipitate formed. The

mixture was then cooled to 0 ЊC for 10 min and ﬁltered. The

ﬁltrate was re-cooled to 0 ЊC and a second batch of precipitate

was collected. The combined precipitate was dried under

reduced presure to give T₂-POM hydrochloride salt 30 (49 mg,

94%) as a white solid. δ_H(500 MHz, D₂O) 1.88 (3H, s, thymine

CH₃), 2.00 (3H, s, thymine* CH₃), 2.19 (1H, ddd, J 14.0, 9.9,

6.2, H3Ј), 2.26 (1H, ddd, J 14.3, 11.5, 6.5, H*3Ј), 2.77 (1H, ddd,

J 14.3, 9.6, 7.4, H*3Љ), 2.93 (1H, ddd, J 14.0, 9.6, 7.4, H3Љ), 3.38

(1H, dd, J 13.6, 7.8, H6Ј), 3.47 (1H, dd, J 12.5, 8.8, H5Љ), 3.51

(2H, dd, J 13.6, 3.4, H6Љ), 3.70 (1H, dd, J 13.4, 7.8, H*5Љ), 3.79

(1H, d, J 16.1, H7Ј), 3.81 (1H, dd, J 13.4, 2.8, H*5Ј), 3.86 (2H,

dd, J 15.3, 4.1, H*6Ј and H*6Љ), 3.91 (1H, dd, J 12.5, 3.7, H5Ј),

3.93 (1H, dddd, J 11.1, 7.4, 41, 4.1, H*2Ј), 4.11 (1H, d, J 16.1,

H7Љ), 4.86 (1H, dddd, J 9.6, 7.8, 6.5, 2.8, H*4Ј), 5.03 (1H, dddd,

J 9.6, 8.8, 6.2, 3.7, H4Ј), 7.48 (1H, s, H*6), 7.61 (1H, s, H6).

*Denotes lower pyrrolidine ring (see ESI for numbering sys-

tem). δ_C(100.6 MHz, D₂O) 11.6 and 11.7 (2 × thymine CH₃),

31.6 and 33.4 (2 × C3Ј), 39.1 and 39.3 (2 × C5Ј), 49.5, 55.1, 59.3

(2 × C6Ј, NCH₂CO), 57.1, 60.1, 60.8, 63.3 (2 × C2Ј, 2 × C4Ј),

110.8 and 111.5 (2 × C5), 142.1 and 143.5 (2 × C6), 152.3 and

152.7 (2 × C2), 166.8 and 167.0 (2 × C4), 171.5 (CH₂CONH);

m/z (FAB^ϩ) 511 ([M Ϫ 3HCl ϩ Na]^ϩ, 98%), 489 ([M Ϫ 3HCl ϩ

H]^ϩ, 100); HRMS m/z (FAB^ϩ) 489.2565 ([M Ϫ 3HCl ϩ H]^ϩ,

C₂₂H₃₃N₈O₅requires m/z, 489.2574).

[(phthalimido)methyl]-4Ј-(thymin-1-yl)pyrrolidine (26). Dry

pyridine (5 µL, 0.063 mmol) and pentaﬂuorophenyl tri-

ﬂuoroacetate (6 µL, 0.035 mmol) were added to a solution of

acid 25 (15 mg, 0.029 mmol) in dry DMF (100 µL). The mixture

was stirred at room temperature for 2 h and then evaporated

under reduced pressure. Ethyl acetate (2 mL) and 5% aqueous

NaHCO₃(2 mL) were then added and the mixture was

extracted with ethyl acetate (5 × 1 mL). The extracts were dried

over anhydrous MgSO₄, ﬁltered, evaporated under reduced

pressure and then puriﬁed by column chromatography (ethyl

acetate) to give the ester 26 (13 mg, 78%) as a white foam.

δ_H(300 MHz, CDCl₃) 1.17 (3H, s, thymine CH₃), 1.88–1.99

(1H, m, H_a3Ј), 2.54–2.68 (1H, m, H_b3Ј), 2.77 (1H, dd, J 11.3,

6.8, H_a5Ј), 3.00–3.09 (1H, m, H2Ј), 3.30 (1H, d, J 10.9, H_b5Ј),

3.52 (1H, d, J 17.7, NCH_aH_bCO₂), 3.80 (1H, dd, J 14.7, 1.9,

Phth-CH_aH_b), 3.90 (1H, dd, J 14.7, 4.5, Phth-CH_aH_b), 4.43

(1H, d, J 17.7, NCH_aH_bCO₂), 4.83–4.92 (1H, m, H4Ј), 7.61

(1H, d, J 1.14, H6), 7.68 (2H, dd, J 5.6, 3.03, Ar), 7.77 (2H, dd,

J 5.5, 3.03, Ar), 7.95 (1H, s, thymine NH); δ_F(188.2 MHz,

CDCl₃) Ϫ84.04 (2F, dd, J 21.8, 17.3, m-F), Ϫ79.47 (1F, t, J 21.8,

p-F), Ϫ74.45 (2F, d, J 17.3, o-F); m/z (FAB^ϩ) 579 ([M ϩ H]^ϩ,

100%), 292 (73); HRMS m/z (EI) 578.1217 (M^ϩ, C₂₆H₁₉N₄O₆F₅

requires m/z, 578.1225).

(7R,8aR)-7-(Thymin-1Ј-yl)hexahydropyrrolo[1,2-a]pyrazin-3-

one (28). A solution of Phth-T₂-Boc-POM 21 (71 mg, 0.099

mmol) in 40% (w/v) aqueous methylamine (1 mL) was warmed

to 50 ЊC for 5 h. On cooling to room temperature the solution

was evaporated under reduced pressure and puriﬁed by column

chromatography (10 15% CH₃OH in CH₂Cl₂) to give the

lactam 28 (18 mg, 69%) as a white foam. δ_H(300 MHz, CDCl₃)

1.33–1.44 (1H, m, H_a3Ј), 1.89 (3H, d, J 1.1, thymine CH₃), 2.40–

2.68 (3H, m, H_b3Ј, H2Ј and H_a5Ј), 2.86 (1H, d, J 16.2, NCH_a-

H_bCO), 3.03 (1H, d, J 11.3, H_b5Ј), 3.23 (1H, dd, J 10.6 and 10.9,

NHCH_aH_b), 3.38–3.47 (1H, m, NHCH_aH_b), 3.65 (1H, d, J 16.2,

Phth-T₃-Boc-POM (32)

Phth-T₂-Boc-POM 21 (370 mg, 0.515 mmol) was added to a

20% solution of CF₃CO₂H in CH₂Cl₂(10 mL) and stirred at

room temperature for 4 h. The solvent was removed under a

stream of argon followed by evaporation under reduced pres-

sure. Dry DMF (2 mL), bromoacetamide 20 (229 mg, 0.515

mmol) and DIPEA (450 µL, 2.58 mmol) were then added and

O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 2 7 7 – 3 2 9 2

3289

View Article Online

the solution stirred at room temperature for 16 h. Evaporation

under reduced pressure and puriﬁcation by column chromato-

graphy (10 20% CH₃OH in CH₂Cl₂) gave Phth-T₃-Boc-POM

32 (0.490 g, 97%) as a white solid. R_f0.27 (9% CH₃OH and 1%

Et₃N in CH₂Cl₂); δ_H(300 MHz, CD₃OD) 1.21 (3H, s, thymine

CH₃), 1.51 (9H, s, C(CH₃)₃), 1.87 and 1.98 (each 3H, s, thymine

CH₃), 1.98–3.20 (12H, m), 3.30–4.40 (13H, m), 4.58–4.67, 4.67–

4.82 and 4.97–5.08 (each 1H, m, CH4Ј), 7.51 (1H, s, H6), 7.73–

7.79 (4H, m, Ar), 8.07 and 8.09 (each 1H, s, H6); δ_C(75.5 MHz,

CD₃OD) 12.5, 12.8 and 13.2 (3 × thymine CH₃), 33.9, 36.5,

37.6, 39.6, 43.4, 43.5, 54.3, 55.2, 55.7, 56.6, 57.5, 57.9, 58.5,

60.2, 63.0, 63.6 and 65.3 (3 × C2Ј, 3 × C3Ј, 3 × C4Ј, 3 × C5Ј, 2 ×

CONHCH₂, Phth-CH₂and 2 × NCH₂CONH), 82.2 (C(CH₃)),

111.0, 111.8 and 112.0 (3 × C5), 124.7, 133.7, 133.5 and 135.9

(Ar C), 140.8, 141.0 and 141.1 (3 × C6), 153.1, 153.2 and 153.6

(3 × C2), 156.7 (Boc CO), 166.5, 166.7 and 167.2 (3 × C4),

(each 3H, s, 3 × thymine CH₃), 1.77 (6H, s, 2 × thymine CH₃),

1.40 (9H, s, C(CH₃)₃), 1.45–1.60 (3H, m, 3 × H_a3Ј), 1.92–2.08

(2H, m, 2 × H3Ј), 2.20–3.90 (38H, 4 × NHCH₂CH2Ј, Phth-

CH₂, 5 × H2Ј, 5 × H3Ј, 5 × H5Ј and 4 × NCH₂CONH), 4.60–

4.95 (5H, m, 5 × H4Ј), 7.54–8.05 (13H, 5 × H6, 4 × Ar and 4 ×

CH₂CONH ), 11.05, 11.12,|| 11.17 and 11.28 (5H, 5 × thymine

NH); m/z (ESI^ϩ) 1511.7 ([M ϩ H]^ϩ, C₇₁H₉₁N₂₀O₁₈requires m/z,

1511.7).

Phth-T₅-POMؒ5HCl (35)

Phth-T₅-Boc-POM 34 (7 mg, 4.6 µmol) was treated with a 1 : 1

CH₃OH–H₂O mixture saturated with HCl gas (2 mL). The

reaction was stirred at 40 ЊC for 5 min and then at room tem-

perature for 2 h. The methanol was then removed under

reduced pressure and the remaining water was lyophilised to

give the desired Phth-T₅-POM 35 HCl salt (7 mg, 95%) as a

white powder. A sample was subsequently lyophilised twice

from D₂O for NMR purposes. δ_H(400 MHz, D₂O) 1.57, 1.65,

1.72, 1.75 and 1.78 (total 15H, 5 × thymine CH₃), 2.00–4.40

(43H, 4 × NHCH₂CH2Ј, 1 × Phth-CH₂, 5 × H2Ј, 5 × H3Ј, 5 ×

H5Ј and 4 × NCH₂CONH), 4.38–4.67 (5H, m, 5 × H4Ј), 7.33–

7.55 (5H, m, 5 × H6), 7.81 and 7.82 (total 4H, 2 × s, Ar); m/z

(ESI^ϩ) 706.3 ([M Ϫ 3H Ϫ 5Cl]^2ϩ, C₆₆H₈₄N₂₀O₁₆requires m/z,

706.3) and 471.2 ([M Ϫ 2H Ϫ 5Cl]^3ϩ, C₆₆H₈₅N₂₀O₁₆requires

m/z, 471.2).

170.5 (Phth CO), 173.6 and 174.2 (2

× CH₂CONH);

m/z (FAB^ϩ) 1005 ([M ϩ Na]^ϩ, 50%), 983 ([M ϩ H]^ϩ, 100), 883

(42), 466 (48); HRMS m/z (FAB^ϩ) 983.4334 ([M ϩ H]^ϩ,

C₄₇H₅₉N₁₂O₁₂requires m/z, 983.4375).

Phth-T₄-Boc-POM (33)

Phth-T₃-Boc-POM 32 (410 mg, 0.418 mmol) was added to a

20% solution of CF₃CO₂H in CH₂Cl₂(10 mL) and the solution

stirred at room temperature for 4 h. The solvent was removed

under a stream of argon followed by evaporation under reduced

pressure. Dry DMF (2 mL), bromoacetamide 20 (185 mg, 0.418

mmol) and then DIPEA (440 µL, 2.53 mmol) were added and

the solution stirred at room temperature for 16 h. Evaporation

under reduced pressure and then puriﬁcation by column chro-

matography (15% CH₃OH in CH₂Cl₂) gave Phth-T₄-Boc-POM

33 (500 mg, 96%) as a white solid. δ_H(400 MHz, CD₃OD) 1.08,

1.76, 1.83 and 1.85 (each 3H, s, thymine CH₃), 1.38 (9H, s,

C(CH₃)₃), 1.55–1.64 and 1.78–1.82 (each 1H, m, H_a3Ј), 1.87–

1.95 (2H, m, 2 × H_a3Ј), 1.99–4.0 (30H, 3 × NHCH₂CH2Ј, Phth-

CH₂, 4 × H2Ј, 4 × H_b3Ј, 4 × H5Ј and 3 × NCH₂CONH), 4.48–

4.54 (1H, m, H_a4Ј), 4.54–4.68 (1H, m, H_b4Ј), 4.85–4.93 (1H, m,

H_A4Ј), 4.75–4.85 (probably 1H, H_B4Ј masked by residual

HDO), 7.36, 7.87, 7.92 and 8.00 (each 1H, s, H6), 7.64 and 7.65

(each 2H, s, Ar); δ_C(100.6 MHz, CD₃OD) 12.6, 12.8, 13.3 and

13.3 (4 × thymine CH₃), 29.2 (C(CH₃)₃), 33.7, 36.39, 37.4, 37.6,

38.8, 39.5, 54.3, 54.6, 55.2, 55.6, 56.2, 56.6, 56.9, 57.6, 57.8,

58.4, 59.8, 59.9, 63.6, 64.9 and 65.4 (4 × C2Ј, 4 × C3Ј, 4 × C4Ј, 4

× C5Ј, 3 × CONHCH₂, Phth-CH₂and 3 × NCH₂CONH)¶,

82.15 (C(CH₃)₃), 111.0, 111.6, 111.8 and 111.9 (4 × C5), 124.7,

133.5 and 136.0 (3 × Ar), 140.8, 140.9¶ and 141.1 (4 × C6),

153.1, 153.2, 153.3 and 153.6 (4 × C2), 156.7 (Boc CO), 166.4,

166.7, 166.8 and 167.3 (4 × C4), 170.52 (Phth CO), 173.5, 174.1

and 174.2 (3 × CH₂CONH); MS m/z (ESI^ϩ) 1247.4 ([M ϩ H]^ϩ,

C₅₉H₇₅N₁₆O₁₅requires m/z, 1247.6).

UV hybridisation studies

General. A Varian-Cary 1 UV-Visible spectrophotometer

equipped with Peltier heating block and six cell transport

mechanism was employed. Quartz cuvettes (24 × 5 × 10 mm)

ﬁtted with Teﬂon stoppers were used throughout. Reference

cuvettes always contained the same buﬀer used in the corre-

sponding sample cuvette. Data were analysed using software

provided by Varian-Cary. For thermal denaturation experi-

ments, absorbance readings were recorded at 260 nm (unless

stated otherwise) with data collection every 0.1 ЊC and an aver-

aging time of 1 s. Temperature monitoring was via a probe

placed inside a cuvette containing buﬀer and adjacent to the

sample cuvette in the cell holder. Concentrations of all oligo-

nucleotides are expressed in mol/base and for each of the oligo-

nucleotides the concentration was 42 µM in bases, unless stated

otherwise. The concentrations of stock solutions were obtained

via serial dilution and for polynucleotides and d(T)₂₀these were

determined spectrophotometrically at 80 ЊC using the Beer–

Lambert law and the known extinction coeﬃcients of the corre-

sponding nucleotides [ε/dm³mol^Ϫ1cm^Ϫ115000 for dA, 15000

for rA, 7600 for rC, 12160 for rG, 10210 for rU and 8500 for

dT]. For Phth-T₅-POM 35, the extinction coeﬃcient used was

8920 which represents the sum of the extinction coeﬃcients of

thymine (ε = 8500) plus 1/5 of the extinction coeﬃcient of

phthalimide (ε = 2100) at 260 nm. All stock solutions were

stored at Ϫ20 ЊC in between use. All sample vials and pipette

tips were sterile and the H₂O used was doubly distilled prior to

use. Silicon oil (Sigma) was used to prevent sample evaporation.

All polynucleotides were commercially available (Sigma) while

the 5-mer Lys-T₅-LysNH₂-PNA (N-terminal lysine and C-

terminal lysine amide) was purchased from PE/Applied

Biosystems, Warrington, UK.

Phth-T₅-Boc-POM (34)

Phth-T₄-Boc-POM 33 (240 mg, 0.193 mmol) was added to a

30% solution of CF₃CO₂H in CH₂Cl₂(6 mL) and stirred at

room temperature for 4 h. The solvent was removed under a

stream of argon followed by evaporation under reduced pres-

sure. Dry DMF (4 mL), bromoacetamide 20 (94 mg, 0.21

mmol) and then DIPEA (235 µL, 1.35 mmol) were added and

the solution stirred at room temperature for 16 h. Evaporation

under reduced pressure and puriﬁcation by column chromato-

graphy (10 20% CH₃OH in CH₂Cl₂then 10 20% CH₃OH in

CH₂Cl₂with 0.5% DIPEA), gave Phth-T₅-Boc-POM 34 (271

mg, ca. 91%) as a white solid containing some residual DIPEA–

TFA salt. Repeated recrystallisation via slow cooling from a

warm saturated methanolic solution aﬀorded an analytically

pure sample. δ_H(300 MHz, DMSO-d6) 1.32, 1.73 and 1.75

Sample preparation. Unless stated otherwise, thermal melt

experiments were carried out using equimolar amounts (in

bases) of both oligonucleotides taken from aqueous stock solu-

tions and adding 500 µL of double concentrated buﬀer. Water

was then added to give a total volume of 1 mL and the solution

then mixed gently using a pipette before adding it to the cuvette.

A thin layer of oil was then added and a Teﬂon stopper was

inserted to prevent evaporation.

¶ Probably two signals are coincident.

|| Two peaks are coincident.

O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 2 7 7 – 3 2 9 2

3290

View Article Online

Procedures for thermal denaturation experiments. For thermal

denaturation experiments involving Phth-T₅-POM 35 and

poly(rA), an equimolar mixture of the two strands (42 µM each

in bases) was initially heated at 5 ЊC min^Ϫ1from 25–93 ЊC in

order to completely dissociate the strands. After 1 min at 93 ЊC,

the sample was cooled at 0.2 ЊC min^Ϫ1to 15 ЊC and then after a

further 1 min the sample was heated at 0.2 ЊC min^Ϫ1to 93 ЊC.

For thermal melting experiments involving Phth-T₅-POM 35

and poly(dA), an equimolar mixture of the two strands was

incubated at a concentration of 210 µM each (in bases) at 25 ЊC

for at least 48 h. The sample was then diluted with the

appropriate buﬀer to 1 mL to give a concentration of 42 µM

each (in bases). After cooling from 25 15 ЊC at 1 ЊC min^Ϫ1, the

sample was allowed to equilibrate at 15 ЊC for at least 1 min

and then heated and cooled between 15 and 93 ЊC at 0.2 ЊC

min^Ϫ1. For thermal melting experiments involving d(T)₂₀or

Lys-T₅-LysNH₂-PNA with either poly(rA) or poly(dA), a

protocol identical to that used for the Phth-T₅-POM

35–poly(rA) thermal melting experiments was followed. For all

T_mdetermination, the temperature at the maximum of the ﬁrst

derivative of the slow melting curve was used. All melting

experiments were performed at least twice (except where stated

otherwise) and the values shown are the average of these

readings.

phosphoramidite and were purchased from Cruachem,

Glasgow, UK.

General procedures for binding studies. Samples were pre-

pared by serial dilution from aqueous stock solutions. All pro-

cedures for binding studies were automated as methods using

repetitive cycles of sample injection and regeneration. Samples

were injected at a ﬂow rate of 20 µL min^Ϫ1using the KINJECT

command from 7 mm plastic vials that were capped with

pierceable plastic crimp caps. In order to minimise carryover,

samples were injected in order of increasing concentration. For

all injections the running buﬀer was identical to the injection

buﬀer. Regeneration using 10 mM hydrochloric acid (20 µL,

20 µL min^Ϫ1) was performed between each injection.

Immobilisation of biotinylated oligonucleotides. Onto the

second ﬂow cell of a dextran chip containing streptavidin was

injected a solution of ssDNA 5Ј-biotin-d(A)₂₀at 1 µg mL^Ϫ1in

water for 3 min and at 10 µL min^Ϫ1until 1450 RU were

obtained. In a similar manner, ﬂow cell 3 was derivatised with

1400 RU of ssRNA 5Ј-biotin-r(A)₂₀and ﬂow cell 4 with 1000

RU of ssDNA 5Ј-biotin-d(AGC TTC AGA GAT CGA TCG

GAG AGA GTA CTG). After at least 100 s of buﬀer wash,

10 µL of 10 mM hydrochloric acid were passed over all the cells.

T₅-POM, T₅-PNA, d(T)₅and d(T)₂₀—oligonucleotide binding

assays. Serial two-fold dilutions (160–10 µM, 100 µL, 20 µL

min^Ϫ1) of Phth-T₅-POM 35 in 10 mM K₂HPO₄adjusted to 0.12

M K^ϩand pH 7.0, were passed serially over all four ﬂow cells of

the sensor chip. The complex was then washed with buﬀer for 6

min followed by regeneration using 10 mM hydrochloric acid

(20 µL, 20 µL min^Ϫ1). d(T)₅was assayed in an identical manner

using the same concentrations. PNA (Lys-T₅-LysNH₂) was

assayed in an identical manner except using serial two-fold dilu-

tions of 40–1.25 µM and d(T)₂₀using 5–0.31 µM dilutions.

Assays were performed at diﬀerent ionic strength and pH (as

indicated) following an identical protocol.

Procedures for continuous variation experiments. In order to

determine the stoichiometry for Phth-T₅-POM 35 binding to

poly(rA), diﬀerent mole ratios of the two strands at 10% incre-

ments and a total base concentration of 50 µM in 10 mM

K₂HPO₄buﬀer adjusted to 0.12 M K^ϩand pH 7.0, were mixed

by pipette and then allowed to anneal for 24 h at 25 ЊC. The

mixtures were then sequentially transferred into a cuvette and

the UV spectra recorded between 300 and 220 nm.

Procedures for kinetic experiments. Using the Varian-Cary

kinetic software, absorbance changes at 260 nm were monitored

with respect to time at 25 ЊC immediately upon adding and

mixing an equimolar amount of either Phth-T₅-POM 35 or

Lys-T₅-LysNH₂-PNA to a buﬀered solution containing either

poly(rA) or poly(dA). Experiments involving the hybridisation

between Phth-T₅-POM 35 and poly(rA) were conducted

using equimolar mixtures of each strand at base concentrations

of 42 µM using a cuvette of path length 1.0 cm. Experiments

involving the hybridisation between Phth-T₅-POM 35 and

poly(dA) were conducted using equimolar mixtures of each

strand at base concentrations of either 42 µM or 210 µM (as

indicated) and using cuvettes of path length 1.0 cm or 0.2 cm

respectively.

Acknowledgements

We thank the EPSRC for a studentship to DTH, the Paterson

Cancer Research Institute (Manchester), Dr Alan Tucker for

assistance with NMR experiments, the EPSRC National Mass

Spectrometry Service Centre (Swansea) and the Michael Baber

Centre for Mass Spectrometry.

References

1 J. Mickleﬁeld, Curr. Med. Chem., 2001, 8, 1157.

2 J. Kurreck, Eur. J. Biochem., 2003, 270, 1628.

3 D. A. Braasch and D. R. Corey, Biochemistry, 2002, 41, 4503.

4 A. Dove, Nat. Biotechnol., 2002, 20, 121.

5 M. Petersen and J. Wengel, Trends Biotechnol., 2003, 21, 74.

6 D. A. Braasch and D. R. Corey, Chem. Biol., 2001, 8, 1.

7 P. E. Nielsen, Curr. Opin. Biotechnol., 1999, 10, 71.

8 P. E. Nielsen, Curr. Opin. Biotechnol., 2001, 12, 16.

9 C. Carlsson, M. Jonsson, B. Norden, M. T. Dulay, R. N. Zare,

J. Noolandi, P. E. Nielsen, L.-P. Tsui and J. Zielenski, Nature, 1996,

380, 207.

10 F. A. Rogers, K. M. Vasquez, M. Egholm and P. M. Glazer,

Proc. Natl. Acad. Sci. USA, 2002, 99, 16695.

11 H. Parekh-Olmedo, M. Drury and E. B. Kmeic, Chem. Biol., 2002,

9, 1073.

12 C. Richert, A. L. Roughton and S. A. Benner, J. Am. Chem. Soc.,

1996, 118, 4518.

13 Z. Huang and S. A. Benner, J. Org. Chem., 2002, 67, 3996.

14 A. Eschenmoser, Science, 1999, 284, 2118.

SPR hybridisation studies. The SPR measurements were per-

formed using a BIAcore 2000 instrument (Biacore AB, UK).

Streptavidin derivatised carboxymethyl dextran sensor chips

(SA) were used as obtained from Biacore. Each sensor chip

contained four ﬂow cells (2.4 × 0.5 × 0.05 mm) with a probing

spot for the SPR signal of ca. 0.26 mm²for each ﬂow cell.

Buﬀers used were as for UV melting experiments (0.12 M K^ϩ,

pH 5.0, 6.0 or 7.0, and 1.2 M K^ϩ, pH 7.0) and were degassed

immediately prior to use. Milli-Q grade water was used

throughout. All assays were carried out at 25 ЊC with data

points collected every 0.5 s. Absence of stem loops or dimers in

the biotinylated mixed sequence oligonucleotides was con-

ﬁrmed using the computer programmes AMPLIFY and GCG

respectively. Data were prepared for analysis and presentation

by subtracting the average response recorded 20 s prior to injec-

tion and adjusting the time of injection to zero. Data from ﬂow

cell 1 (underivatised dextran) were subtracted from the corre-

sponding data obtained from the ﬂow cells (2, 3 and 4) that

contained biotinylated oligonucleotides, in order to correct for

bulk refractive index changes and signal drift. All oligonucleo-

tides were biotinylated at the 5Ј-end using multi biotin-III-CE

15 S. M. Freier and K.-H. Altmann, Nucleic Acid Res., 1997, 25, 4429.

16 E. T. Kool, Chem. Rev., 1997, 97, 1473.

17 J. Wengel, Acc. Chem. Res., 1999, 32, 301.

18 S. K. Singh and J. Wengel, Chem. Commun., 1998, 1247.

19 A. De Mesmaeker, A. Waldner, J. Lebreton, P. Hoﬀmann, V. Fritsch,

R. M. Wolf and S. M. Freier, Angew. Chem., Int. Ed. Engl., 1994, 33,

226.

O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 2 7 7 – 3 2 9 2

3291

View Article Online

20 K. J. Fettes, N. Howard, D. T. Hickman, S. A. Adah, M. R. Player,

P. F. Torrence and J. Mickleﬁeld, Chem. Commun., 2000, 765.

21 B. Hyrup and P. E. Nielsen, Bioorg. Med. Chem., 1996, 4, 5.

22 P. E. Nielsen, Acc. Chem. Res., 1999, 32, 624.

23 E. Uhlmann, A. Peyman, G. Breipohl and D. W. Will,

Angew. Chem., Int. Ed., 1998, 37, 2796.

24 P. M. Reddy and T. C. Bruice, Bioorg. Med. Chem. Lett., 2003, 13,

1281.

25 D. T. Hickman, P. M. King, M. A. Cooper, J. M. Slater and

J. Mickleﬁeld, Chem. Commun., 2000, 2251.

26 G. J. M. Koper, R. C. van Duijvenbode, D. D. P. W. Stam,

U. Steuerle and M. Borkovec, Macromolecules, 2003, 36, 2500.

27 G. J. M. Koper, M. H. P. van Genderen, C. Elissen-Román, M. W. P.

L. Baars, E. W. Meijer and M. Borkovec, J. Am. Chem. Soc., 1997,

119, 6512.

28 N. Umezawa, M. A. Gelman, M. C. Haigis, R. T. Raines and

S. H. Gellman, J. Am. Chem. Soc., 2002, 124, 368.

29 T. Furuya, S. Fujita, S. Iwanami, A. Takenka and Y. Sasada,

Acta Crystallogr., Sect. C, 1986, 42, 1345.

37 T. Vilaivan, C. Khongdeesameor, P. Harnyuttanakorn, M. S.

Westwell and G. Lowe, Bioorg. Med. Chem. Lett., 2000, 10, 2541.

38 M. D’Costa, V. Kumar and K. N. Ganesh, Org. Lett., 2001, 3, 1281.

39 V. Kumar, P. S. Pallan, Meena and K. N. Ganesh, Org. Lett., 2001, 3,

1269.

40 T. Vilaivan and G. Lowe, J. Am. Chem. Soc., 2002, 124, 9326.

41 J. J. P. Stewart, J. Comput. Aided Mol. Des., 1990, 4, 1.

42 G. L. Baker, S. J. Fritschel, J. R. Stille and J. K. Stille, J. Org. Chem.,

1981, 46, 2954.

43 T. Sasaki, K. Minamoto, T. Suzuki and T. Sugiura, J. Am. Chem.

Soc., 1978, 100, 2248.

44 T. Sasaki, K. Minamoto, T. Suzuki and T. Sugiura, J. Org. Chem.,

1979, 44, 1424.

45 Z. Dzakula, M. L. DeRider and J. L. Markley, J. Am. Chem. Soc.,

1996, 118, 12796.

46 P. Job, Ann. Chim., 1928, 9, 113.

47 G. Felsenfeld, Biochim. Biophys. Acta, 1958, 29, 133.

48 R. O. Dempcy, K. A. Browne and T. C. Bruice, Proc. Natl. Acad. Sci.

USA, 1995, 92, 6097.

30 C. Altona and M. Sundaralingam, J. Am. Chem. Soc., 1972, 94,

8205.

31 W. Saenger, Principles of Nucleic Acid Structure (Ed. C. R. Cantor),

Springer-Verlag, New York, 1984.

32 E. A. Green, R. D. Rosenstein, R. Shiono, D. J. Abraham, B. L. Trus

and R. E. Marsh, Acta Crystallogr., Sect. B, 1975, 31, 102.

33 G. Lowe and T. Vilaivan, J. Chem. Soc., Perkin Trans. 1, 1997, 539–

560.

49 R. O. Dempcy, K. A. Browne and T. C. Bruice, Proc. Natl. Acad. Sci.

USA, 1995, 92, 7051.

50 M. Wasner, D. Arion, G. Borkow, A. Noronha, A. H. Uddin,

M. A. Parniak and M. J. Damha, Biochemistry, 1998, 37, 7478.

51 T. L. Sheppard and R. C. Breslow, J. Am. Chem. Soc., 1996, 118,

9810.

52 J. S. Pudlo, X. Cao, S. Swaminathan and M. D. Matteucci,

Tetrahedron Lett., 1994, 35, 9315.

53 J. D. McGhee and P. H von Hippel, J. Mol. Biol., 1974, 86, 469.

54 M. Malmqvist, Nature, 1993, 361, 186.

34 G. Lowe, T. Vilaivan and M. S. Westwell, Bioorg. Chem., 1997, 25,

321.

35 M. D’Costa, V. Kumar and K. N. Ganesh, Org. Lett., 1999, 1,

1513.

55 M. Hendrix, S. E. Priestley, G. F. Joyce and C.-H. Wong, J. Am.

Chem. Soc., 1997, 119, 3641.

56 K. K. Jensen, H. Ørum, P. E. Nielsen and B. Nordén, Biochemistry,

1997, 36, 5072.

36 A. Püschl, T. Tedeschi and P. E. Nielsen, Org. Lett., 2000, 2,

4161.

O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 2 7 7 – 3 2 9 2

3292

Article Doi

Design, synthesis, conformational analysis and nucleic acid hybridisation properties of thymidyl pyrrolidine-amide oligonucleotide mimics (POM)

DOI: 10.1039/b306156f

Source and publish data:

Authors:

Article abstract of DOI:10.1039/b306156f

Full text of DOI:10.1039/b306156f

Products guided by the article

R&D Labs maybe for 664967-01-5

Relevant to this article

Hot Product