Synthesis of DNA oligomers possessing a covalently cross-linked Watson - Crick base pair model

Article abstract of DOI:10.1002/1521-3773(20010417)40:8<1471::AID-ANIE1471>3.0.CO;2-I

Building bridges: A feasible solidphase synthesis of antiparallel n-, h-, and H-type DNA oligomers (see picture) has been demonstrated. The oligomers possess a CH₂-bridged base-pair model that should be conformationally flexible, even after base pairing.

Full text of DOI:10.1002/1521-3773(20010417)40:8<1471::AID-ANIE1471>3.0.CO;2-I

COMMUNICATIONS
[3] L. E. Overmann, R. L. Freenks, C. B. Petty, L. A. Clizbe, R. K. Ono,
G. F. Taylor, P. J. Jessup, J. Am. Chem. Soc. 1981, 103, 2816; Y. Becker,
A. Eisenstadt, J. K. Stille, J. Org. Chem. 1980, 45, 2145; J. W. F. K.
Barnick, J. L. van der Baan, F. Bickelhaupt, Synthesis 1979, 787; J. H.
Rigby, Synlett 2000, 1.
[4] For reviews, see: Pd: B. M. Trost, Acc. Chem. Res. 1990, 23, 34; B. M.
Trost, Chem. Eur. J. 1998, 4, 2405; Ru: B. M. Trost, Chem. Ber. 1996,
129, 1313; general: B. M. Trost, M. J. Krische, Synlett 1998, 1.
[5] T. P. Gill, K. R. Mann, Organometallics 1982, 1, 485.
[6] B. M. Trost, F. D. Toste, Tetrahedron Lett. 1999, 40, 7739; B. M. Trost,
F. D. Toste, J. Am. Chem. Soc. 2000, 122, 714.
and lack of conformational flexibility between the base
pairs.[1]
We have recognized the possibility that CH2-bridged base
pair models may be uniquely suited to the chemical explora-
tion of covalently cross-linked nucleosides/nucleotides. In
addition to their increased chemical stability, these base pair
models are expected to adopt only Watson ± Crick or reverse
Watson ± Crick base pairings while maintaining conforma-
tional flexibility along the CH2 bridge. We have focused on
two specific types (types I and II) of base pair models
(Scheme 1), with the anticipation that they might exhibit
[7] T. Kondo, Y. Mousaki, S.-Y. Uenoyama, K. Wada, T. A. Mitsudo, J.
Am. Chem. Soc. 1999, 121, 8657; Y. Morisaki, T. Kondo, T. A. Mitsudo,
Organomet. 1999, 18, 4742; S.K. Kang, D.-Y. Kim, R.-K. Hong, P.-S.
Ho, Synth. Commun. 1996, 26, 3225.
[8] New compounds have been satisfactorily characterized (see Support-
ing Information).
[9] B. M. Trost, A. F. Indolese, T. J. J. Müller, B. Treptow, J. Am. Chem.
Soc. 1995, 117, 615.
[10] B. M. Trost, J. A. Martinez, R. J. Kulawiec, A. F. Indolese, J. Am.
Chem. Soc. 1993, 115, 10402.
[11] V. Cadierno, M. P. Gamasa, J. Gimeno, J. Borge, S. Garcia-Granda,
Organometallics 1997, 16, 3178; D. Touchard, N. Pirio, P. H. Dixneuf,
Organometallics 1995, 14, 4920; J. R. Selegue, B. A. Young, S. L.
Logan, Organometallics 1991, 10, 1972, and references therein. See
also B. M. Trost, J. A. Flygare, J. Am. Chem. Soc. 1992, 114, 5476.
[12] B. M. Trost, M. Machacek, M. J. Schnaderbeck, Org. Lett. 2000, 1761.
[13] B. M. Trost, T. J. J. Müller, J. Martinez, J. Am. Chem. Soc. 1995, 117,
1888.
major
Type I
groove
Base Pairs
N
H
N
R2
N
N
R2
H
N
O
O
N
Me
CH2
Me
CH2
O
N
N
H
O
H
N
R1
N
R1
minor
groove
Corresponds to
Watson-Crick T-A or C-G
Corresponds to
Reverse Watson-Crick T-A or C-G
major
groove
[14] B. M. Trost, Science 1991, 254, 1471; B. M. Trost, Angew. Chem. 1995,
107, 285; Angew. Chem. Int. Ed. Engl. 1995, 34, 259.
Type II
Base Pairs
[15] As it does in the case of Pd, see B. M. Trost, J. Y. L. Chung, J. Am.
Chem. Soc. 1985, 107, 4586; B. M. Trost, D. C. Lee, J. Org. Chem. 1989,
54, 2271; B. M. Trost, G. J. Tanoury, M. Lautens, C. Chan, D. T.
MacPherson, J. Am. Chem. Soc. 1994, 116, 4255; B. M. Trost, D. L.
Romero, F. Rise, J. Am. Chem. Soc. 1994, 116, 4268.
H
N
N
R2
N
R2
N
H
N
O
O
N
Me
CH2
O
Me
CH2
N
N
H
O
H
minor
groove
N
R1
N
R1
Corresponds to
Watson-Crick T-A or C-G
Corresponds to
Reverse Watson-Crick T-A or C-G
Scheme 1. Covalently cross-linked Watson ± Crick base pair models. A ˆ
2'-deoxyadenosine, C ˆ 2'-deoxycytidine, G ˆ 2'-deoxyguanosine, Tˆ 2'-
deoxythymidine.
Synthesis of DNA Oligomers Possessing a
Covalently Cross-Linked Watson ± Crick Base
Pair Model**
differing structural characteristics. Although both the type I
and type II models adopt only Watson ± Crick or reverse
Watson ± Crick base pairings in the sense of primary hydro-
gen-bonded base pairing, only the type II model provides a
structural motif for formation of Hoogsteen triplets such as
T± AT and C ± GC: see the two bold arrows in Scheme 1
which indicate possible hydrogen-bonding sites in a major
groove. We recently reported the synthesis and structural
properties of type I and II base pairs.[4] In this paper, we
present a method for incorporating these base pair models
into DNA oligomers.
A priori, we considered that phosphoramidite-based solid-
phase synthesis[3, 5] would best meet with our future needs, and
we studied its applicability to three types of oligomers
incorporating a covalently cross-linked base pair, the n, h,
and H types. Our synthetic plan is schematically depicted in
Scheme 2, with the order of chain elongation being indicated
by a circled number. First, all of the proposed chain
elongations take place in the 3' !5' direction for n- and h-
type oligomers, but the last chain elongation takes place in the
5' !3' direction for H-type oligomers. Second, the order of the
Hong-Yu Li, Yao-Ling Qiu, Elisabeth Moyroud, and
Yoshito Kishi*
The concept of covalently cross-linked sections with
molecular architecture similar to Watson ± Crick hydrogen-
bonded base pairs was introduced by Devadas and Leonard in
the mid-1980ꢁs.[1] Since then, several types of covalently linked
systems have been developed. However, these systems[2, 3]
were generated from preformed double helices, as seen in
the seminal work of Ferentz and Verdine. The Leonard system
may offer unique opportunities to address questions regarding
the chemistry of DNA and RNA, but this system has several
drawbacks, including difficulty in attempted duplex formation
[*] Prof. Dr. Y. Kishi, Dr. H.-Y. Li, Dr. Y.-L. Qiu, Dr. E. Moyroud
Department of Chemistry and Chemical Biology
Harvard University, Cambridge, MA 02138 (USA)
Fax : (‡1)617-495-5150
E-mail: kishi@chemistry.harvard.edu
[**] Financial support from the National Institutes of Health (Grant: NS
12108) is gratefully acknowledged.
Angew. Chem. Int. Ed. 2001, 40, No. 8
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
1433-7851/01/4008-1471 $ 17.50+.50/0
1471

COMMUNICATIONS
n Type
H Type
h Type
5'
5'
3'
5'
3'
CH2
1
2
3'
5'
2
1
2
1
4
3
CH2
CH2
3
3'
5'
3'
5'
3'
Y
O
Y
O
N
Y
O
N
CH2
CH2
CH2
Me
N
N
Me
Me
N
N
N
O
N
N
N
N
O
X
DMTO
2
O
N
X
O
X
Alloc-O
3
O
BzO
DMTO
2
DMTO
2
LevO
3
O
O
O
O
O
P
AcO
O
P
O
P
Alloc-O
AcO
NC
NC
NC
4
O
N(iPr)2
O
N(iPr)2
O
N(iPr)2
1
1
1
1a: X=NH-Fmoc; Y=H
2a: X=H; Y=NH-Fmoc
1b: X=NH-Fmoc; Y=H
2b: X=H; Y=NH-Fmoc
1c: X=NH-Fmoc; Y=H
2c: X=H; Y=NH-Fmoc
Scheme 2. Synthetic plan for n-, h-, and H-type oligomers. Ac ˆ acetyl, Alloc ˆ allyloxycarbonyl, Bz ˆ benzoyl, DMTˆ 4,4'-
dimethoxytrityl, Fmoc ˆ 9-fluorenylmethoxycarbonyl, Lev ˆ levulinoyl.
second and third chain elongations for the h and H types can
be inverted simply by switching the DMT group with the
Alloc or the Lev group in 1b and 2b or 1c and 2c,
O
N
CH2
Me
N
N
a)
b)
N
1a
1b
1c
NH2
O
respectively. Third, an Fmoc group is chosen for protection
of the amino group; this allows for simultaneous removal,
along with deprotection of the protecting groups on the A, T,
G, and C nucleotides, upon detachment of the oligomers from
the solid support. Fourth, we plan to employ solid-phase
synthesis with controlled-pore glass (CPG) supports.[6] Thus,
alcoholic protecting groups compatible with this solid support
are required; Alloc and Lev protecting groups are known to
meet with this requirement.[7, 8]
To put this plan in practice, we first required base pairs 1a ±
c[9] and 2a ± c, which were efficiently synthesized as summar-
ized in Scheme 3. With the arbitrarily chosen base sequences
shown in Scheme 4, the feasibility of this plan was then tested.
Clearly, the crucial step in these syntheses entails adding a
CH2-bridged base pair model onto an elongating oligonucleo-
tide chain, that is, step 8 for n-type oligomer syntheses, step 3
for h-type, and step 5 for H-type. The base pairs 1a ± c and
2a ± c could be incorporated into these oligomers approx-
imately as efficiently as the nonmodified nucleotides (see
Experimental Section for more details).[10] However, to
achieve this efficiency, longer coupling times and higher
concentrations of the modified bases were required.[11] Under
these conditions, approximately 10 ± 15 equivalents of a
modified base were delivered to a DNA synthesizer. How-
ever, the unused modified base could be recovered almost
quantitatively by employing the method recently reported.[12]
In the case of h-type oligomer syntheses, the solid support was
removed from the synthesizer after completion of the second
chain elongation, subjected to Alloc deprotection on the solid
support,[7] and reintroduced into the synthesizer to complete
the third chain elongation, step 11. In the case of H-type
TBSO
DMTO
O
O
c)
TBSO
AcO
I[4a, 9]
NH2
O
N
CH2
Me
N
N
N
O
d)
d)
O
2a
2b
2c
TBSO
DMTO
O
d)
TBSO
II[4b]
AcO
Scheme 3. Reagents and conditions. a) 1. FmocCl, DMAP, Pyr, RT, 92%;
2. nBu4NF, AcOH, THF, RT, 93%; 3. BzCl, Pyr, CH2Cl2, 08C, 63% of the
product, along with 33% of starting material; 4. (iPr)2NP(Cl)O(CH2)2CN,
Et(iPr)2N, CH2Cl2, RT, 93%; b) 1. FmocCl, DMAP, Pyr, RT, 92%; 2. TCA,
CH2Cl2, RT, 96%; 3. Alloc-OBt, DMAP, Pyr, THF, RT, 95%; 4. nBu4NF,
AcOH, THF, RT, 89%; 5. DMTCl, DMAP, CH2Cl2, RT, 92%; 6. (iPr)2NP-
(Cl)O(CH2)2CN, Et(iPr)2N, CH2Cl2, RT, 95%; c) 1. NH3, MeOH, RT,
100%; 2. FmocCl, DMAP, Pyr, RT, 95%; 3. Alloc-OBt, Pyr, DMAP, THF,
RT 96%; 4. TCA, CH2Cl2, RT, 90%; 5. (Lev)2O, Pyr, RT, 92%; 6. nBu4NF,
AcOH, THF, RT, 87%; 7. DMTCl, Pyr, CH2Cl2, RT 93%; 8. (iPr)2NP-
(Cl)O(CH2)2CN, Et(iPr)2N, CH2Cl2, RT, 100%; d) steps (h)1 ± 3, (h)1 ± 5,
or (h)1 ± 7 of Scheme 1 in ref. [4b] as appropriate, followed by (iPr)2NP-
(Cl)O(CH2)2CN, Et(iPr)2N, CH2Cl2, RT, 94% for 2a, 95% for 2b, 97% for
2c. Abbreviations: Bt ˆ benzotriazol-1-yl, nBu ˆ normal butyl, DMAP ˆ
4-dimethylaminopyridine, Et ˆ ethyl, iPrˆ isopropyl, Pyrˆ pyridine,
TBS ˆ tert-butyldimethylsilyl, TCA ˆ trichloroacetic acid, THF ˆ tetrahy-
drofuran.
oligomer syntheses, the column containing the solid support
was removed from the synthesizer twice (after completion of
the second and third chain elongations), subjected to Lev and
1472
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
1433-7851/01/4008-1472 $ 17.50+.50/0
Angew. Chem. Int. Ed. 2001, 40, No. 8

COMMUNICATIONS
H Type
n Type
h Type
5'-AcO
3'-DMTO
C
C
A
T
C
A
O
P O
G
G
T
G
T
O
P O
10
9
5'-AcO
11
10
9
8
7
21
20
19
18
C
C
A
C
A
O
P O
8
7
6
5
(Fmoc NH)
O
N
CH2
Me
N
N
N
O
Fmoc
NH
O
NC
O
O
NC
NC
O
BzO
O
17
O
O
(Fmoc NH)
O
N
(Fmoc NH)
N
CH2
CH2
Me
Me
N
N
4
N
O
N
O
6
O
OAc
9
Fmoc
NH
O
Fmoc
NH
O
NC
O P
O
N
O
O
O
N
8
7
O
O P O
O
O
O
O
CN
A
T
T
A
T
10
11
12
13
14
15
16
6
5
4
3
2
1
O
OAc
11
O
O
A
G
G
T
G
T
NC
3
NC
O P
O
O P
O
12
5
4
O P O
O
O P O
O
C
C
A
C
A
O
O
CN
CN
A
C
G
C
T
2
1
T
G
A
C
G
A
C
A
12
13
13
3
2
1
T
14
15
16
G
5'-DMTO
5'-DMTO
T
5'-AcO
Scheme 4. Synthesis of n-, h-, and H-type oligomers. The order in which the chains were elongated is indicated. For synthetic details and yields, see the
Experimental Section.
Alloc deprotections on the solid support,[7, 8] and reintroduced
into the synthesizer to complete the third and fourth chain
elongations, steps 12 and 17. The overall efficiencies of these
syntheses are summarized in the Experimental Section and
demonstrate that the CH2-bridged base pair models 1a ± c and
2a ± c could be efficiently incorporated into n-, h-, and H-type
oligomers.[13]
After completion of the chain elongations, the oligomers
were detached from the solid support, isolated, and purified
under the conditions specified in the Experimental Section.
The syntheses were carried out on 1-mmol scale and furnished
approximately 0.45 mmol, 0.35 mmol, and 0.15 mmol of HPLC-
purified n-, h-, and H-type oligomers, respectively (Figure 1).[14]
The molecular weight of each synthetic oligomer was
established by mass spectrometry. In the early phase of these
studies, we relied on electrospray ionization (ESI) mass
spectrometry, but in later phases, we used the matrix-assisted
laser desorption/ionization (MALDI) mass spectrometry for
technical reasons.[15] By knowing both the order of nucleotides
introduced on an elongating chain and the molecular weight
of the products, we could secure the structure of the synthetic
oligomers.
In conclusion, we have demonstrated the synthetic feasi-
bility of n-, h-, and H-type DNA oligomers possessing a
Figure 1. HPLC chromatograms of crude and purified DNA oligomers.
Traces a) ± c) indicate the purity of the crude products obtained after the
detachment/deprotection step: a) type II, n type; b) type II, h type; c) type
II, H type. Traces A) ± C) indicate the purity of the isolated products after
covalently cross-linked Watson ± Crick base pair model. It
should be noted that the current study is focused on the
synthesis of anti-parallel DNA duplexes. However, with
adjustments of the protecting groups in 1a ± c and 2a ± c, the
plan depicted in Scheme 2 can be extended to the synthesis of
parallel DNA duplexes. Circular dichroism and NMR spec-
troscopic studies, which demonstrate that the structural
properties of DNA oligomers containing a covalently cross-
linked Watson ± Crick base pair model resemble those of the
corresponding native duplex DNA, will be reported else-
where.[16]
the final purification step: A) type II, n type; B) type II, h type; C) type II,
H type. For further details, see the Experimental Section.
Exerimental Section
Synthesis of n-, h-, or H-type DNA oligomers was carried out on 1-mmol
scale with an Applied Biosystems 392 DNA/RNA synthesizer (coupling
time: 25 seconds for 3'-CE phosphoramidites; 10 minutes for the modified
Angew. Chem. Int. Ed. 2001, 40, No. 8
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
1433-7851/01/4008-1473 $ 17.50+.50/0
1473

COMMUNICATIONS
dinucleoside phosphoramidites and for 5'-CE phosphoramidites) or with a
PerSeptive Expedite 8909 (coupling time: 90 seconds for 3'-CE phosphor-
amidites; 15 minutes for the modified dinucleoside phosphoramidites and
for 5'-CE phosphoramidites). CE ˆ b-cyanoethyl.
[6] The following phosphoramidites and dT-lcaa-CPG (500 Š; loading
capacity:40 mmolg 1) from Glen Research, Virginia, were used for the
current studies. iBu ˆ isobutyryl, lcaa ˆ long chain alkylamino.
O
O
NHBz
NHBz
The following reagents were used: deblocking mix: TCA/CH2Cl2; activa-
tor: 1H-tetrazole/MeCN; cap mix A: Ac2O/THF; cap mix B: 1-methyl-
imidazole/Pyr/THF; oxidizing solution: I2/H2O/Pyr/THF. All the reagents
and the dA-, dT-, dC-, or dG-3'/5'-CE phosphoramidites were purchased
from Glen Research, Virginia. A 0.1m solution of dA-, dT-, dC-, or dG-3'/5'-
CE phosphoramidites was used with the Applied Biosystems 392 DNA/
RNA synthesizer, whereas a 0.06m solution was used with PerSeptive
Biosystems Expedite 8909 DNA/RNA synthesizer. A 0.08 ± 0.1m solution
was used for coupling 1a ± c or 2a ± c onto a chain. Removal of the Alloc
protecting group was performed with [Pd(PPh3)4] and HCO2NH4 in THF at
RT over 24 h and removal of the Lev protecting group was required
NH2NH2 ´ H2O (0.5m) in Pyr/HOAc at RT for 10 min.
Me
N
N
N
NH
HN
iBuHN
DMTO
N
N
O
N
N
O
N
N
O
N
DMTO
DMTO
DMTO
O
O
O
O
P
O
O
O
CN
P
CN
P
CN
P
CN
(iPr)2N
O
(iPr)2N
O
(iPr)2N
O
(iPr)2N
O
dGiBu
dABz
Me
dCBz
dT
O
N
NH
O
Based on the amount of DMT released in a coupling cycle over that in the
preceding coupling cycle, the coupling yield was estimated to be 97 ± 100%
for step 8 of the n-type, steps 3 and 11 of the h-type, and steps 5, 12, and 17
of the H-type oligomer syntheses, and 98 ± 100% for all the other steps. The
overall yields, estimated from the amount of DMT released in the second-
to-last coupling cycle over that in the first coupling cycle, were 70 ± 80%,
60 ± 70%, and 50 ± 60% for the n-, h-, or H-type oligomers, respectively.
DMTO
O
O
OAc
MeO OMe
H
N
O
O
Si
O
N
H
Controlled Pore Glass
O
O
[7] Y. Hayakawa, S. Wakabayashi, H. Kato, R. Noyori, J. Am. Chem. Soc.
1990, 112, 1691.
[8] a) J. A. J. dan Hartog, G. Wille, J. H. van Boom, Recl. Trav. Chim.
Pays-Bas 1981, 100, 320; b) S. Iwai, E. Ohtsuka, Nucleic Acids Res.
1988, 16, 9443.
Purification/isolation of an oligomer was achieved through a five-step
procedure: 1) 28% NH4OH, 558C, 24 h (detachment/deprotection of an
oligomer from solid support); 2) HPLC purification under the conditions
indicated below; 3) 10% AcOH (deprotection of DMT); 4) C18 cartridge
filtration; 5) lyophilization. The HPLC chromatograms (Figure 1) were
recorded on a VydacC4 column (10 Â 250 mm); solvent system: MeCN/
100 mm AcONH4, 3% ± 50% over 20 min for a) ± c) and 2% ± 15% over
20 min for A) ± C); flow rate: 3 mLmin 1. Traces a ± c indicate the purity of
the crude products obtained after step 1 of the purification. Retention
times: a) type II, n type: 15.0 min; b) type II, h type: 16.1 min; c) type II, H
type: 17.2 min. Traces A ± C indicate the purity of the products after step 5
of the purification. Retention times: A) type II, n type: 15.4 min); B) type
II, h type: 15.9 min; C) type II, H type: 16.1 min.
[9] Two modifications have been made on the original synthesis of I.[4a]
These are: 1) NaBH4 reduction of the mixed anhydride, that is,
step (b)5. in Scheme 1 of ref. [4a], is better performed at
78 ! 308C, and 2) separation of N1- and N3-glycosides can most
practically be achieved after the NaBH4 reduction. With these
improvements, a 20-gram scale synthesis of I can be carried out, with
one silica filtration. The overall yield of 5 from 1 in Scheme 1 in
ref. [4a] is approximately 26%.
[10] At an earlier stage of this study, the solid-phase synthesis was carried
out on an Applied Biosystems 392 DNA/RNA Synthesizer in the
Verdine group. We thank Professor G. Verdine for his generosity and
help. In later stages, the synthesis was performed on a PerSeptive
Biosystems Expedite 8909. No obvious difference in synthetic
efficiency was noticed between the two instruments.
Received: November 20, 2000
Revised: January 18, 2001 [Z16143]
[1] a) B. Devadas, N. J. Leonard, J. Am. Chem. Soc. 1986, 108, 5012; b) for
the latest publication on this subject from the Leonard group, see: B.
Bhat, N. J. Leonard, H. Robinson, A. H.-J. Wang, J. Am. Chem. Soc.
1996, 118, 10744.
[11] Using a PerSeptive Biosystems Expedite 8909, optimal conditions
were briefly studied for the case of 1a and 2a. To achieve the efficient
attachment of 1a and 2a onto the elongating chain, their concen-
tration should be at least 0.08m and the coupling time for attaching
them on the elongating chain should be longer than the coupling time
(90 seconds) for the nonmodified nucleotides.
[12] W. Wang, Q. Song, R. A. Jones, Tetrahedron Lett. 1999, 40, 8971.
[13] We tested the feasibility of H-type oligomer syntheses with polystyr-
ene-based solid support. According to the procedure reported by
Strazewski (S. Gunzenhauser, E. Biala, P. Strazewski, Tetrahedron
Lett. 1998, 39, 6277), the polystyrene-based solid support loaded with
T (loading capacity: 25 mmolg 1) was prepared and used for the solid-
phase synthesis. Two modified base pairs i and ii were successfully
incorporated into H-type oligomers with an efficiency similar to that
outlined in the Experimental Section. We thank Dr. Peter Strazewski,
University of Basel, for a generous gift of polystyrene polymer.
[2] For examples of covalently linked base pairs, see the contributions
from Verdine (A. E. Ferentz, G. L. Verdine, J. Am. Chem. Soc. 1991,
113, 4000), Benkovic (M. Cowart, S. J. Benkovic, Biochemistry 1991,
30, 788), Mitchell and Kelly (M. A. Mitchell, R. C. Kelly, N. A.
Wicnienski, N. T. Hatzenbuhler, M. G. Williams, G. L. Petzold, J. L.
Slighton, D. R. Siemieniak, J. Am. Chem. Soc. 1991, 113, 8994),
Lippert (I. Dieter-Wurm, M. Sabat, B. Lippert, J. Am. Chem. Soc.
1992, 114, 357), Armstrong (R. W. Armstrong, M. E. Salvati, M.
Nguyen, J. Am. Chem. Soc. 1992, 114, 3144), Hopkins (J. J. Kirchner,
S. T. Sigurdsson, P. B. Hopkins, J. Am. Chem. Soc. 1992, 114, 4021 and
H. Huang, M. S. Solomon, P. B. Hopkins, J. Am. Chem. Soc. 1992, 114,
9240), Hurley (F. C. Seaman, L. Hurley, Biochemistry 1993, 32,
12577), and Glick (S. E. Osborne, R. J. Cain, G. D. Glick, J. Am.
Chem. Soc. 1997, 119, 1172).
Y
O
i: X=NH-Fmoc, Y=H
CH2
Me
ii: X=H, Y=NH-Fmoc
N
N
[3] Sigurdsson, Hopkins, and co-workers have recently reported a solid-
phase synthesis of a nitrous acid interstrand cross-linked duplex DNA:
E. A. Harwood, S. T. Sigurdsson, N. B. F. Edfeldt, B. R. Reid, P. B.
Hopkins, J. Am. Chem. Soc. 1999, 121, 5081.
N
O
TBDPS=tert-butyldiphenylsilyl
X
O
N
TBDPSO
DMTO
O
[4] a) X. Qiao, Y. Kishi, Angew. Chem. 1999, 111, 977; Angew. Chem. Int.
Ed. 1999, 38, 928; b) Y.-L. Qiu, H.-Y. Li, G. Topalov, Y. Kishi,
Tetrahedron Lett. 2000, 41, 9425.
O
P
Alloc-O
NC(CH2)2O N(iPr)2
[5] M. D. Matteucci, M. H. Caruthers, J. Am. Chem. Soc. 1981, 103, 3185;
for examples of reviews on this subject, see: a) M. H. Caruthers, Acc.
Chem. Res. 1991, 24, 278; b) J. W. Engels, E. Uhlmann, Angew. Chem.
1989, 101, 733; Angew. Chem. Int. Ed. Engl. 1989, 28, 716; c) S. L.
Seaucage, R. P. Iyer, Tetrahedron 1993, 49, 6123.
[14] The low yields of H-type oligomers, relative to the estimated overall
yields (see the Experimental Section), were largely due to the
difficulties encountered during HPLC purification.
1474
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
1433-7851/01/4008-1474 $ 17.50+.50/0
Angew. Chem. Int. Ed. 2001, 40, No. 8

COMMUNICATIONS
[15] n Type: m/z calcd: 5445; type I found: 5444 (negative ion ESI-MS);
type II found: 5440 (MALDI-MS, matrix: 3-hydroxypicolinic acid/
ammonium citrate). h Type: m/z calcd: 4154; type I found: 4151
(MALDI-MS, matrix: 2,6-dihydroxyacetophenone/ammonium cit-
rate); type II found: 4127 (MALDI-MS, matrix: 2,6-dihydroxyaceto-
phenone/ammonium citrate). H Type: m/z calcd: 6683; type I found:
6669 (MALDI-MS, matrix: 2,6-dihydroxyacetophenone/ammonium
citrate); type II found: 6671 (MALDI-MS, matrix: 3-hydroxypicolinic
acid/ammonium citrate).
OH
HO
HO
O
CO2H
HO
HO
HO
Kdo
OH
Kdo
O
O
CO2H
O
O
O
[16] H.-Y. Li, Y.-L. Qiu, Y. Kishi, ChemBioChem, in press.
O
O
(HO)2P
HO
O
O
O
O
HN
O
O
O
HN
O P(OH)2
O
O
O
OH
O
Re LPS (1)
OH
O
First Total Synthesis of the Re-Type
Lipopolysaccharide**
Hiroaki Yoshizaki, Naohiro Fukuda, Kenjiro Sato,
Masato Oikawa,* Koichi Fukase, Yasuo Suda, and
Shoichi Kusumoto*
lipid A
Lipopolysaccharides (LPS) are ubiquitous glycoconjugates
located on the surface of Gram-negative bacteria such as
Escherichia coli, and exhibit multiple and potent biological
activity, both toxic and beneficial, towards higher animals.[1]
LPS consist of a phosphorylated acyl glucosamine disacchar-
ide covalently bonded to a polysaccharide. The former
glycolipid, lipid A, is the entity responsible for most of the
biological activity of LPS.[2, 3] Lipid A is therefore never
present on living bacterial cells in the free form; it is an
artificial product obtained only by the acid hydrolysis of LPS.
The simplest LPS molecule that is known to be present on
bacterial cells and that comes into contact with animal cells is
the so-called Re-type LPS produced by the Escherichia coli
Re mutant (Re LPS, 1).[4] This LPS derivative is composed of
lipid A and two molar equivalents of 3-deoxy-d-manno-2-
octurosonic acid (formerly called 2-keto-3-deoxy-d-manno-
octonic acid (Kdo)). The chemical synthesis of Re LPS is not
only a highly challenging goal, but would also contribute
towards the elucidation of the biological and physicochemical
role of the sugar moieties linked to lipid A. It has often been
implied that the polysaccharide portion enhances or modifies
the activity of Re LPS.[5] In fact, Re LPS that only has two
units of Kdo was reported to exhibit more potent antitumor
and cytokine-inducing activity than lipid A.[6, 7] However,
these observations have never been confirmed owing to the
inherent heterogeneity and the difficult purification of natural
specimens. An efficient synthesis of Re LPS would also
provide easy access to various structural congeners required
for the study of structure ± activity relationships. Such a
strategy based on chemical synthesis has been effective in
our studies of the role of acyl and phosphoryl moieties in the
bioactivity of lipid A.[8]
Until now, only partial syntheses of Re LPS have been
reported, for example, our synthesis of the 1-O-dephospho
analogue.[9] The total synthesis of Re LPS has not yet been
accomplished because of the difficulty in synthesizing the
complex structure, which contains highly acid-labile glycosyl
phosphate, as well as base-labile ester functional groups in the
amphiphilic structure. A high-yielding and stereoselective
formation of a-Kdo glycosides[10±13] is a key step towards the
synthesis of Re LPS. We have recently completed an efficient
synthesis of lipid A[8c, 14] and its labeled analogues[15] by means
of an optimized synthetic pathway coupled with a high-
yielding efficient purification method for the final products
based on liquid ± liquid partition.[16] These achievements
prompted us to undertake the synthesis of Re LPS
(Scheme 1). Herein we describe an efficient stereoselective
glycosylation method with Kdo donors, and the first total
synthesis of Re LPS. A direct comparison of the biological
activities of synthetic and natural specimens is also reported.
Based on our experience with the synthesis of lipid A and
its analogues,[8, 9, 15, 16] we employed benzyl-type groups that
are removable by neutral hydrogenolysis, as they are resilient
protecting groups for the hydroxy, carboxy, and phosphate
groups in 2. The 4-(trifluoromethyl)benzyl group, which was
found to be resistant to oxidation but readily removable by
hydrogenolysis,[17] was used especially for the protection of the
hydroxy groups on the 3-hydroxyacyl residues. Unsubstituted
benzyl groups at these positions are prone to air-oxidation and
gradually are transformed into the corresponding benzoyl
groups during storage, even at 58C.[17, 18] The tetrasaccharide
backbone of 2 was formed by the successive coupling of two
[*] Dr. M. Oikawa, Prof. Dr. S. Kusumoto, H. Yoshizaki, N. Fukuda,
K. Sato, Dr. K. Fukase, Dr. Y. Suda
Department of Chemistry, Graduate School of Science
Osaka University
Toyonaka, Osaka 560-0043 (Japan)
Fax : (‡81)6-6850-5419
E-mail: moik@chem.sci.osaka-u.ac.jp
skus@chem.sci.osaka-u.ac.jp
[**] This work was supported by the Research for the Future Program
(No. 97L00502) from the Japan Society for the Promotion of Science.
H.Y. is grateful for a JSPS Research Fellowship for Young Scientists
(No. 1241) from the Japan Society for the Promotion of Science. The
authors are grateful to Mr. Seiji Adachi for his skillful measurement of
NMR spectra.
Angew. Chem. Int. Ed. 2001, 40, No. 8
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
1433-7851/01/4008-1475 $ 17.50+.50/0
1475