Fluorescent deazaflavin - Oligonucleotide probes for selective detection of DNA

Article abstract of DOI:10.1002/1521-3773(20020201)41:3<486::AID-ANIE486>3.0.CO;2-O

Dramatic quenching of fluorescence is observed both in solution and on a glass surface during hybridization of deazaflavin - oligonucleotide conjugates (see picture, N = nucleoside) to a complementary strand. This opens up the possibility of a new type of

Full text of DOI:10.1002/1521-3773(20020201)41:3<486::AID-ANIE486>3.0.CO;2-O

COMMUNICATIONS
[1] a) M. L. Gross, D. H. Blank, W. M. Welch, J. Org. Chem. 1993, 58,
2104; b) R. Lazny, J. Poplawski, J. Kˆbberrling, D. Enders, S. Br‰se,
Synlett 1999, 1304.
Fluorescent Deazaflavin Oligonucleotide
Probes for Selective Detection of DNA
[2] a) H. Heaney, J. M. Jablonski, C. T. McCarty, J. Chem. Soc. Perkin
¬
¬
¬
Cecile Dueymes, Jean Luc Decout,* Philippe Peltie,
and Marc Fontecave*
ƒ
Trans. 1 1972, 2903; b) F. Gavina, S. V. Luis, P. Ferrer, A. M. Costero, P.
Gil, Tetrahedron 1986, 42, 5641.
[3] P. C. Buxton, H. Heaney, Tetrahedron 1995, 51, 3929.
[4] a) E. B. Merlushev, Synthesis 1988, 923, and references therein; b) J. S.
Moore, E. J. Weinstein, Z. Wu, Tetrahedron Lett. 1991, 32, 2465.
[5] N. Satyamurthy, J. R. Barrio, G. T. Biba, M. E. Phelps, Tetrahedron
Lett. 1990, 31, 4409.
[6] K. C. Nicolaou, C. N. C. Boddy, S. Natarajan, T. Y. Yue, H. Li, S.
Br‰se, J. M. Ramanjulu, J. Am. Chem. Soc. 1997, 119, 3421.
[7] a) J. K. Young, J. C. Nelson, J. S. Moore, J. Am. Chem. Soc. 1994, 116,
10841; b) J. C. Nelson, J. K. Young, J. S. Moore, J. Org. Chem. 1996, 61,
Arrays of oligonucleotide probes (DNA chips) immobilized
on glass or silicon surfaces have emerged as powerful new
tools for the analysis of DNA and RNA.[1] These specific
DNA sensors operate through hybridization reactions which
are based on the mutual recognition of two complementary
nucleic acid strands that establish hydrogen bonds between
their nucleic bases. Hybridization is most commonly assayed
by fluorescence, and hence requires that fluorophore labels
are covalently attached to the DNA target fragments to be
analyzed.
ƒ
8160; c) F. Gavina, S. V. Luis, P. Ferrer, A. M. Costero, P. Gil,
Tetrahedron 1986, 42, 5641; d) S. Br‰se, D. Enders, J. Kˆbberling, F.
Avemaria, Angew. Chem. 1998, 110, 3614; Angew. Chem. Int. Ed.
1998, 37, 3413; e) S. Br‰se, M. Schroen, Angew. Chem. 1999, 111, 1139;
Angew. Chem. Int. Ed. 1999, 38, 1071; f) S. Br‰se, J. Kˆbberling, D.
Enders, R. Lazny, M. Wang, S. Brandtner, Tetrahedron Lett. 1999, 40,
2105; g) S. Br‰se, S. Dahmen, J. Heuts, Tetrahedron Lett. 1999 40,
6201; h) S. Dahmen, S. Br‰se, Angew. Chem. 2000, 112, 3827; Angew.
Chem. Int. Ed. 2000, 39, 3681; i) S. Br‰se, S. Dahmen, M. Pfefferkorn,
J. Comb. Chem. 2000, 2, 710.
This technology, however, is still fraught with a number of
drawbacks and requires new developments. For example, it is
still difficult to assess the quality of the oligonucleotides
attached to the surface. The homogeneity and the reprodu-
cibility of the procedures for preparing the DNA-functional-
ized surfaces, and consequently the DNA surface density, are
difficult to control. Furthermore the polymerase chain
reaction (PCR) step, which serves to amplify the DNA
targets, and the subsequent chemical step, for the attachment
of the fluorophore to the target DNA, are responsible for
significant modifications in the relative proportions of the
different populations of the nucleic acids to be analyzed.
We propose that some of these problems might be simply
solved if the detection label (here a fluorophore) is incorpo-
rated on the array capture DNA strand and not on the target
to be detected. This strategy might have many advantages.
First it greatly reduces the number of manipulations of the
targets. Second, it provides a way to control the quality of the
DNA array (by using standard fluorescence scanners) in the
absence of the targets. Such an array requires that: 1) the
fluorescence is not quenched by interaction of the fluoro-
phore with the surface; 2) the label on the probe does not
affect its affinity for its complementary oligonucleotide; and
3) the fluorescence is significantly changed as a specific
consequence of hybridization of the functionalized DNA
[8] J. Rademann, J. Smerdka, G. Jung, P. Grosche, D. Schmid, Angew.
Chem. 2001, 113, 390; Angew. Chem. Int. Ed. 2001, 40, 381.
[9] D. B. Kimball, A. G. Hayes, M. M. Haley, Org. Lett. 2000, 2, 3825.
[10] a) D. Seebach, D. Enders, Chem. Ber. 1975, 108, 1293; b) B. Renger,
H. O. Kalinowski, D. Seebach, Chem. Ber. 1977, 110, 1866; c) R. R.
Fraser, S. Passannanti, Synthesis 1976, 540; d) D. S. Grierson, J. Royer,
L. Guerrier, H. P. Husson, J. Org. Chem. 1986, 51, 4475; e) K. Rein, M.
Goicoechea-Pappas, T. V. Anklekar, G. C. Hart, G. A. Smitj, R. E.
Gawley, J. Am. Chem. Soc. 1989, 111, 2211; f) P. Beak, W. K. Lee, J.
Org. Chem. 1990, 55, 2578; g) D. B. Reitz, P. Beak, A. Tse, J. Org.
Chem. 1981, 46, 4316; h) A. I. Meyers, W. T. Hoeve, J. Am. Chem. Soc.
1980, 102, 7125.
[11] For the dearomatized compounds, see a) T. Bach, Angew. Chem. 1996,
108, 795; Angew. Chem. Int. Ed. 1996, 35, 729, and references therein;
b) H. A. J. Carless, Tetrahedron: Asymmetry 1992, 3, 7 95; c) W. D.
Harman, Chem. Rev. 1997, 97, 1953, and references therein; d) E. P.
K¸ndig, D. Amurrio, G. Anderson, D. Beruben, K. Khan, A. Ripa, L.
Ronggang, Pure Appl. Chem. 1997, 69, 543; e) S. Saito, K. Shimada, H.
Yamamoto, E. M. de Marigorta, I. Fleming, Chem. Commun. 1997,
1299; f) J. Boivin, M. Yousfi, S. Z. Zard, Tetrahedron Lett. 1997, 38,
5985; g) R. Berger, J. W. Ziller, D. L. Van Vranken, J. Am. Chem. Soc.
1998, 120, 841; h) A. Ahmed, J. Clayden, M. Rowley, Synlett 1999,
1954; i) A. Ahmed, J. Clayden, S. A. Yasin, Chem. Commun. 1999,
231; j) A. Ahmed, J. Clayden, M. Rowley, Chem. Commun. 1998, 297;
k) R. A. Bragg, J. Clayden, Tetrahedron Lett. 1999, 40, 8323; l) A.
Ahmed, J. Clayden, M. Rowley, Tetrahedron Lett. 1998, 39, 6103; m) J.
Clayden, K. Tchabanenko, Chem. Commun. 2000, 317; n) R. E.
Bolton, C. J. Moody, C. W. Rees, J. Chem. Soc. Perkin Trans. 1 1989,
2136.
¬
[*] Prof. J. L. Decout
¬
¬
Departement de Pharmacologie Moleculaire
¬
UMR Universite J. Fourier, CNRS no. 5063
Domaine de la Merci, 38706 La Tronche Cedex (France)
Fax : (‡33)4-76-51-86-67
E-mail: jean-luc.decout@ujf-grenoble.fr
Prof. M. Fontecave, C. Dueymes
DBMS-CB
¬
UMR Universite J. Fourier, CNRS, CEA no. 5047
CEA Grenoble
17Avenue des Martyrs, 38054 Grenoble Cedex 9 (France)
Fax : (‡33)4-38-78-91-24
E-mail: mfontecave@cea.fr
¬
Dr. P. Peltie
LETI, CEA Grenoble
17Avenue des Martyrs
38054 Grenoble Cedex 9 (France)
Supporting information for this article is available on the WWW under
http://www.angewandte.com or from the author.
486
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
1433-7851/02/4103-0486 $ 17.50+.50/0
Angew. Chem. Int. Ed. 2002, 41, No. 3

COMMUNICATIONS
probe to its complementary target, so that DNA recognition
can be detected.
O
NH
a)
b)
H
Here we report that flavins and deazaflavins provide the
first examples of such a label. Flavins and deazaflavins exhibit
a bright blue or green fluorescence at 520 540 and 450
460 nm, respectively, in aqueous solutions, with deazaflavins
giving larger fluorescence intensities. We have recently
reported an efficient method for the synthesis of stable
flavin oligonucleotide conjugates that is also used here for
the preparation of new deazaflavin oligonucleotide conju-
gates.[2] Furthermore, these conjugates are stable to UV
irradiation.[3] Preliminary results demonstrate here that these
molecules are able to hybridize to a complementary oligonu-
cleotide target both in solution and on solid support, and that
the hybridization can be detected by a dramatic quenching of
the probe fluorescence.
N
N
N
O
NH2
(CH2)6
OH
(CH2)6
OH
9
8
c)
O
d)
e)
NH
O
N
N
O
NH
f)
(CH2)6
O
11
N
N
O
PyH
O P
O
H
(CH2)6
OH
10
CN
g)
h)
O
P
N
O
N
O
O
HO
O
O
i)
j)
n -1
N: Nucleoside
The deazaflavin oligonucleotide conjugates
1 and 4
O
(Scheme 1 and 2) were synthesized from the corresponding
oligo-2'-deoxynucleotides prepared by the phosphoramidite
automated method with standard building phosphoramidites
NH
N
N
O
(CH2)6
O
P
N
O
O
OH
1, 4
O
Conjugates:
n
O
N
O
N
Scheme 2. Synthesis of the 5-deazaflavin oligonucleotide conjugates:
a) HO(CH2)6Cl, 1208C, 74%; b) 6-chlorouracil/H2O/EtOH, reflux, 65%;
c) Ac2O/Py, RT; d) DMF/POCl3, 1008C; e) KOH/EtOH, RT, 70% for the
three (c e) steps; f) H3PO3/TPSO2Cl/Py, 85%; g) adamantanecarboxylic
acid chloride/3'-acetylthymidine/Py; h) nucleoside, coupling step; i) I2/
THF/Et3N; j) NaOH/H2O/CH3OH. Py ˆ pyridine, TP ˆ 2,4,6-triisopropyl-
benzene.
N
N
NH
NH
N
O
O
Fl
dFl
O
P
5'
d(T)n
1: n = 11, X = dFl
2: n = 11, X = Fl
3: n = 16, X = Fl
X
(CH2)6
O
O
O
P
5'
The fluorescence emission spectrum of the deazaflavin
moiety of the conjugates was comparable to that of the free
deazaflavin, with a peak maximum at 450 nm (excitation at
396 nm) and a high quantum yield of 0.26, which is compa-
rable to or even better than that of standard fluorophores such
as Alexa and Cyanine dyes. Furthermore, the observation that
the duplex has the same melting temperature (Tm ˆ 628C,
NaCl 10 mm) whether the probe contained the tag or not[5]
indicates that the deazaflavin tag had no effect on the stability
of the duplex formed between the oligonucleotide probe and
a 42-mer oligonucleotide target, ICAM20, which contains the
20-mer complementary sequence (Scheme 1).
d(CCCCCACCACTTCCCCTCTC) 3'
X
(CH2)6
O
O
4: X = dFl, 5: X = Fl
Targets:
5' d(CTCATCGTGT(A)nGGCAGTACTGGAAGGGGCTAATTCT) 3'
6: n = 11, 7: n = 16
ICAM20: d(GCCTGATGAGAGGGGAAGTGGTGGGGGAGACATAGCCCACC)
ICAM17: d(GCCTGATGAGAGGGGAAGTGGTGGCCCAGACATAGCCCACC)
ICAM14: d(GCCTGATGAGAGGGGAAGTGGATTCCCAGACATAGCCCACC)
ICAM18: d(GCCTGATGAGAGGGGACTTGGTGGGGGAGACATAGCCCACC)
Scheme 1. Oligonucleotide conjugates of 5-deazaflavin (dFl) and flavin
(Fl), as well as their corresponding targets.
Figure 1 shows a representative experiment in which the
addition of increasing amounts of the target oligonucleotide
ICAM20 resulted in a decay of the fluorescence intensity of
compound 4 at 450 nm to less than 5% of the initial value
after the addition of one equivalent of ICAM20. Control
experiments, carried out either by addition of ICAM20 in the
presence of one equivalent of the free deazaflavin to an
oligonucleotide identical to 4 but lacking the deazaflavin
group or by addition of increasing amounts of a 45-mer
oligonucleotide target with no complementary sequence to a
solution of conjugate 4, resulted in no variation in the
fluorescence of the deazaflavin unit. In the case of conjugate
1, quenching also occured upon hybridization with target 6
(Tm ˆ 438C, NaCl 1m), and the fluorescence was shown to be
fully restored upon heating to a temperature that allowed
dissociation of the complex. Several cycles of hybridization/
(see Supporting Information). The 5'-hydroxyl group of the
oligonucleotide attached to the solid support was coupled to
the 5-deazaflavin H-phosphonate derivative 11 (synthesized
from aniline and 6-chlorouracil) after activation with ada-
mantanecarboxylic acid chloride (Scheme 2).[4] Oxidation of
the conjugate on the solid support with iodine and then
cleavage and deprotection with sodium hydroxide afforded
the conjugates 1 and 4 in high yields (60 and 53%,
respectively). The conjugates 1 and 4 were characterized by
1H and 31P NMR spectrometry, mass spectrometry (matrix-
assisted laser desorption/ionization time-of-flight (MALDI-
TOF) and electrospray ionization (ESI)), and absorption
spectrophotometry.
Angew. Chem. Int. Ed. 2002, 41, No. 3
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
1433-7851/02/4103-0487 $ 17.50+.50/0
487

COMMUNICATIONS
probe, to discriminate targets of variable complementarity.
Significantly decreased quenching occurs because of de-
creased base pairing, which occurs as a result of mutations
at the 3' part of the sequence facing the 5' part of conjugate 4
to which the deazaflavin moiety is attached (compare
ICAM20 to ICAM17 and ICAM14) or because of mutations
in the interior of the DNA targets (compare ICAM20 to
ICAM18). This effect probably results from a combination of
an increased distance between the deazaflavin and the target
and decreased amounts of the hybridized probe at equili-
brium, and hence might have applications for the detection of
mutations in solution.
Since deazaflavin oligonucleotide conjugates might serve
as fluorescent probes in the new concept of DNA microarrays
proposed here, we finally checked whether the same duplex-
specific quenching effect occurred on a solid surface. In this
experiment 20 mL of a 10 mm solution of conjugate 4 were
deposited at several places on a glass slide and then dried. By
shining a mercury 100 W lamp of an epi-fluorescence micro-
scope through an appropriate dichroic mirror for excitation at
396 nm, fluorescence emission could be detected with a high
sensitivity spectrophotometer through an optical fiber mount-
ed on the microscope in place of the camera.[6] This arrange-
ment allowed the detection of a well-resolved signal at 500 nm
(intensity: 1100 arbitrary units) corresponding to the fluo-
rescence of the deazaflavin moiety of conjugate 4 on the glass
surface. Duplex formation was allowed for 10 min on several
spots after the addition of five equivalents (20 mL in 0.1m
NaCl) of the complementary oligonucleotide ICAM20. An
assay of the fluorescence of each spot after drying the surface
revealed a reproducible large quenching (intensity: 100
arbitrary units). No quenching of fluorescence could be
detected when oligonucleotides with no complementary
sequence were instead added on fluorescent spots under the
same conditions and then dried. This result demonstrates that
quenching was specific for the hybridized probe and not the
result of some unspecific effect of the added oligonucleotide.
In conclusion, we propose that flavins or deazaflavins might
be suitable fluorescent markers for oligonucleotide probes to
be used in DNA microarrays. First, these markers do not
affect the affinity for the complementary strand. Second, their
fluorescence is also detectable on a solid surface by fluo-
rescence scanners, which thus provides a means of controlling
the quality of the array. Third, their fluorescence is specifically
quenched upon hybridization to a complementary sequence,
thus allowing DNA detection. Fourth, the hybridization/
fluorescence quenching is reversible. Even though, in general,
a generation of a signal is preferred to a decay, the concept of
labeled DNA arrays proposed here is also likely to be
applicable and to provide significant advantages.
Figure 1. Effect of the addition of increasing amounts of ICAM20 on the
fluorescence intensity (I) of conjugate 4 (2 mm in 50 mm tris(hydroxyme-
thyl)aminomethane-HCl (Tris-HCl), pH 7.5, 10 mm NaCl, 258C) at 450 nm
(excitation at 396 nm). x represents the target/4 ratio.
melting, as monitored by fluorescence spectroscopy, were
successfully performed. All these highly reproducible results
showed that the fluorescence quenching observed here was a
specific property of a deazaflavin moiety attached to an
oligonucleotide on hybridization to the complementary
strand.
Similar results were obtained with the flavin oligonu-
cleotide conjugates 2, 3, and 5 (synthesized as described
previously[2]) and using the oligonucleotides 6, 7, and ICAM20
as the complementary strands, respectively. In the case of
probe 3, sequence variations in the proximity of the flavin
moiety (GC, CG, or CC in place of the GG doublet adjacent
to the polyA strand) in the target 7 had no effect on the
quenching efficiency. The results taken together indicate that
the extent of fluorescence quenching, with both deazaflavin
and flavin as markers, does not depend on particular target
lengths and sequences. This observation seems to rule out
specific interactions, such as hydrogen bonding between the
probe fluorophore and bases of the target, as the origin of the
quenching.
An interesting observation is shown in the Stern Volmer
representation Io/I(x) plotted as a function of increasing
amounts x of various targets (Figure 2). The results indicate
that fluorescence quenching makes it possible, with a given
Received: July 26, 2001
Revised: November 19, 2001 [Z17607]
Figure 2. Effect of the addition of increasing amounts of different targets
[1] C. M. Niemeyer, D. Blohm, Angew. Chem. 1999, 111, 3039 3043;
Angew. Chem. Int. Ed. 1999, 38, 2865 2869.
~
&
*
*
(
ICAM2O, ICAM18, ICAM17, ICAM14) on the fluorescence of
¬
[2] C. Frier, J. L. Decout, M. Fontecave, J. Org. Chem. 1997, 62, 3520 3528.
conjugate 4 under the same conditions as for Figure 1. Io represents the
initial fluorescence intensity of 4 and I(x) its intensity after the addition of
x equivalents of the target oligonucleotide.
¬
[3] C. Frier, J. F. Mouscadet, J. L. Decout, C. Auclair, M. Fontecave, Chem.
Commun. 1998, 2457 2458.
488
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
1433-7851/02/4103-0488 $ 17.50+.50/0
Angew. Chem. Int. Ed. 2002, 41, No. 3

COMMUNICATIONS
[4] A low-yielding synthesis of modified thymidine decamers covalently
linked to the N3 atom of 5-deazaflavin derivatives has been reported,
see Y. Eikyu, Y. Nakamura, T. Akiyama, F. Yoneda, K. Tanaka, K. Fuji,
Chem. Pharm. Bull. 1992, 40, 291 293; Y. Nakamura, T. Akiyama, Y.
Yoneda, K. Tanaka, F. Yoneda, Chem. Pharm. Bull. 1993, 41, 778 780.
[5] Measurements of Tm values were carried out as described in: K. J.
Breslauer, R. Frank, H. Blˆcker, L. A. Marky, Proc. Natl. Acad. Sci.
USA 1986, 83, 3746 3750.
growth process to occur at the aluminum centers or, in other
words, a transition metal catalyzed ™Aufbaureaktion∫. Later it
was realized that in these polymerization systems, the polymer
chain actually grows on the transition metal and not on the
aluminum center. But, to complicate matters, in recent years it
has been observed that, for certain early as well as late
transition metal polymerization catalysts, under certain con-
ditions, the polymer chain is transferred to the aluminum
centers during the polymerization reaction, a process that is
referred to as chain transfer to aluminum. This reaction is
sometimes the only transfer process,[5] but more often it occurs
alongside the more common b-H transfer process.[6 9] If chain
transfer to aluminum constitutes the sole transfer mechanism,
and the exchange of the growing polymer chains between the
transition metal and the aluminum centers is very fast and
reversible, the polymer chains will appear to be growing on
the aluminum centers. This can then reasonably be described
as a transition metal catalyzed chain growth reaction on
aluminum or, using Ziegler×s terminology, a transition metal
catalyzed ™Aufbaureaktion∫. An attractive manifestation of
this type of chain-growth reaction (cata-
[6] The high-sensitivity spectrophotometer with a grating optimized for
visible light (360 850 nm) was purchased from Ocean Optics Europe
Company.
Iron-Catalyzed Polyethylene Chain Growth on
Zinc: Linear a-Olefins with a Poisson
Distribution**
George J. P. Britovsek, Steven A. Cohen,
Vernon C. Gibson,* Peter J. Maddox, and
Martin van Meurs
lyzed or uncatalyzed) is a Poisson distri-
bution of main group alkyl products, as
opposed to the Schulz Flory distribution
of olefinic products that arises when b-H
transfer accompanies propagation.[10]
Systems where these criteria are fulfil-
led, or are close to being fulfilled, have
been described for transition metal and
rare-earth metallocenes (on Al and
Fifty years ago Ziegler and co-workers reported the
Aufbaureaktion for alkylaluminum compounds.[1] This reac-
À
tion involves a stepwise insertion of ethylene into the Al C
bonds of, for example, triethylaluminum to give long-chain
trialkylaluminum compounds [Eq. (1)]. Today, this chain-
growth reaction is commercially exploited for the synthesis of
linear a-olefins and primary alcohols from ethylene.[2]
14]
Mg)[11
and recently for half-sandwich complexes of chro-
mium (on Al).[16] Here we report the first highly active chain-
growth process on zinc, catalyzed by bis(imino)pyridineiron
catalyst 1 [Eq. (2)]. We also describe a nickel-catalyzed
Subsequently, Ziegler and co-workers discovered the effect
of transition metals such as colloidal nickel on this reaction,
which suppressed chain growth to give only butene.[3] The
elucidation of this ™nickel effect∫, and the subsequent screen-
ing of the periodic table for the effects of other metals led to
the discovery of the first transition metal catalyzed ethylene
polymerization.[4]
In their early reports,[4] Ziegler and co-workers generally
described the aluminum compound as the polymerization
catalyst, with the transition metal compound functioning as
the cocatalyst, and therefore considered the whole chain-
displacement of the grown alkyl chain, in the presence of
ethylene, to regenerate diethylzinc and to give a Poisson
distribution of a-olefins [Eq. (3); acac ˆ acetylacetonate].
[*] Prof. V. C. Gibson, Dr. G. J. P. Britovsek, M. van Meurs
Department of Chemistry
In the absence of ZnEt2, a polymerization system contain-
ing 1 (5 mmol) and 100 equivalents of methylalumoxane
(MAO) generates high molecular weight polyethylene with
a broad molecular weight distribution (Table 1, Figure 1). A
small amount of a low molecular weight fraction is also
obtained, due to chain transfer to the AlMe3 present in MAO.
An increased proportion of low molecular weight material is
obtained upon addition of 100 equivalents of ZnEt2 (run 2,
Table 1). At 500 equivalents of ZnEt2 (2.5 mmol) a very
narrow molecular weight distribution product results (Mn ˆ
700, Mw ˆ 800, Mw/Mn ˆ 1.1). The polymer yield after hydro-
Imperial College of Science, Technology and Medicine
Exhibition Road, London SW72AY (UK)
Fax : (‡44)20-7594-5810
E-mail: V.Gibson@ic.ac.uk
Dr. S. A. Cohen
BP America Inc., Naperville Complex
150 W. Warrenville Road, Naperville, IL, 60563 (USA)
Dr. P. J. Maddox
BP plc, Sunbury Research Center
Chertsey Road, Sunbury on Thames, Middlesex TW16 7LN (UK)
[**] This work was supported by BP plc. Dr. Gary Audley and Dr. Jane
Boyle are thanked for GPC and NMR measurements, respectively.
Angew. Chem. Int. Ed. 2002, 41, No. 3
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
1433-7851/02/4103-0489 $ 17.50+.50/0
489