Proofing of Photolithographic DNA Synthesis with 3′,5′-Dimethoxybenzoinyloxycarbonyl-Protected Deoxynucleoside Phosphoramidites

Article abstract of DOI:10.1021/jo970872s

We have evaluated in a microchip format the photochemical solid-phase phosphoramidite DNA synthesis method we previously developed. A set of nucleoside building blocks with "easy-off" base protecting groups was prepared bearing photolabile 5′-O-dimethoxybenzoincarbonate (DMBOC) groups. Photolysis rates and cycle yields for these DMBOC-protected nucleotides covalently attached to planar, derivatized glass surfaces were determined by fluorescence imaging-based methods earlier developed by McGall et al. and described in detail elsewhere. Data were obtained for both 280/310 and 365/400 nm irradiation in a range of solvents. Deprotection of the DMBOC occurs fastest in a nonpolar medium or without solvent. The coupling efficiency of these amidites in the synthesis of homopolymers was determined to be in the range 80-97%, with purines generally showing lower efficiency than pyrimidines. These DMBOC-protected monomers were used to prepare a 4 × 4 array of 16 decanucleotides of the sequence 5′-AAXTAXCTAC-chip, where X = A, C, G, or T. The array was hybridized with a target deoxyeicosanucleotide of the sequence fluorescein-5′-CTGAACGGTAGCATCTTGAC. Surface fluorescence imaging demonstrated sequence-specific hybridization to this probe.

Full text of DOI:10.1021/jo970872s

J . Org. Chem. 1998, 63, 241-246
241
P r oofin g of P h otolith ogr a p h ic DNA Syn th esis w ith
3′,5′-Dim eth oxyben zoin yloxyca r bon yl-P r otected Deoxyn u cleosid e
P h osp h or a m id ites
Michael C. Pirrung* and Lara Fallon
Department of Chemistry, P. M. Gross Chemical Laboratory, Duke University,
Durham, North Carolina 27708-0346
Glenn McGall
Affymetrix, 3380 Central Expressway, Santa Clara, California 95051
Received May 16, 1997 (Revised Manuscript Received November 3, 1997X)
We have evaluated in a microchip format the photochemical solid-phase phosphoramidite DNA
synthesis method we previously developed. A set of nucleoside building blocks with “easy-off” base
protecting groups was prepared bearing photolabile 5′-O-dimethoxybenzoincarbonate (DMBOC)
groups. Photolysis rates and cycle yields for these DMBOC-protected nucleotides covalently attached
to planar, derivatized glass surfaces were determined by fluorescence imaging-based methods earlier
developed by McGall et al. and described in detail elsewhere. Data were obtained for both 280/310
and 365/400 nm irradiation in a range of solvents. Deprotection of the DMBOC occurs fastest in
a nonpolar medium or without solvent. The coupling efficiency of these amidites in the synthesis
of homopolymers was determined to be in the range 80-97%, with purines generally showing lower
efficiency than pyrimidines. These DMBOC-protected monomers were used to prepare a 4 × 4
array of 16 decanucleotides of the sequence 5′-AAXTAXCTAC-chip, where X ) A, C, G, or T. The
array was hybridized with a target deoxyeicosanucleotide of the sequence fluorescein-5′-CTGAACG-
GTAGCATCTTGAC. Surface fluorescence imaging demonstrated sequence-specific hybridization
to this probe.
In tr od u ction
chemistry to be applied so that the number of sequences
synthesized far exceeds the number of chemical steps
required. One combinatorial chemistry method that has
been applied to the preparation of DNA on chip-like
surfaces is light-directed synthesis.5 We have developed
a novel photoremovable protecting group, dimethoxyben-
zoincarbonate (DMBOC),6 for light-directed synthesis and
evaluated it for 5′-hydroxyl protection in phosphoramid-
ite-based DNA synthesis.7 We wished to evaluate this
chemistry in a chip format using methods that have been
developed earlier and described elsewhere.8 A minor
modification from earlier practice was required, involving
rapidly removable base protecting groups (phenoxyacetyl
(PAC) for A and G, isobutyryl (i-Bu) for C)9 so that the
final deprotection of the array could be carried out under
mild conditions that would not cleave the oligonucleotides
The determination and utilization of DNA sequence
information is crucial to modern biology. Avant garde
methods for analysis of nucleic acid sequences include
“chip” methodologies1 in which DNA is attached at
microscopic sites on a surface and permitted to hybridize
to complementary sequences in an analyte DNA. The
detection of such hybrid formation has been based on a
variety of technologies, including fluorescence microscopy
and primer extension reactions.2 The preparation of
microarrays can be accomplished by delivery of presyn-
thesized oligonucleotides3 or in situ synthesis.4 The latter
affords the opportunity for the principles of combinatorial
X Abstract published in Advance ACS Abstracts, December 15, 1997.
(1) (a) Mirzabekov, A. D. Trends Biotechnol. 1994, 12, 27-32. (b)
Lysov, Y. P.; Chernyi, A. A.; Balaeff, A. A.; Beattie, K. L.; Mirzabekov,
A. D. J . Biomol. Struct. Dyn. 1994, 11, 797-812. (c) Livshits, M. A.;
Florentiev, V. L.; Mirzabekov, A. D. J . Biomol. Struct. Dyn. 1994, 11,
783-95. (d) Mirzabekov, A. D.; Drmanac, S.; Drmanac, R. BioTech-
niques 1994, 17, 328-9, 332-6. (e) Drmanac, R.; Drmanac, S.;
Strezoska, Z.; Paunesku, T.; Labat, I.; Zeremski, M.; Snoddy, J .;
Funkhouser, W. K. Science (Washington, D.C., 1883-) 1993, 260, 1649-
52. (f) Southern, E. M. Trends Genet. 1996, 12, 110-15. (g) Cronin,
M. T.; Fucini, R. V.; Kim, S. M.; Masino, R. S.; Wespi, R. M.; Miyada,
C. G. Human Mutation 1996, 7, 244-255. (h) Sapolsky, R. J .; Lipshutz,
R. J . Genomics 1996, 33, 445-456. (i) Kozal, M. J .; Shah, N.; Shen,
N.; Yang, R.; Fucini, R.; Merigan, T. C.; Richman, D. D.; Morris, M.
S.; Hubbell, E.; Chee, M. S.; Gingeras, T. R. Nature Med. 1996, 2, 753-
759. (j) Chee, M. S.; Huang, X.; Yang, R.; Hubbell, E.; Berno, A.; Stern,
D.; Winkler, J .; Lockhart, D. J .; Morris, M. S.; Fodor, S. P. A. Science
1996, 274, 610-614. (k) Lockhart, D. J .; Dong., H.; Byrne, M. C.;
Follettie, M. T.; Gallo, M. V.; Chee, M. S.; Mittmann, M.; Wang, C.;
Kobayashi, M.; Horton, H.; Brown, E. L. Nature Biotechnology 1996,
14, 1675-1680.
(3) Khrapko, K. R.; Lysov, Y. P.; Khorlyn, A. A.; Ivanov, I. B.; Yerhov,
G.; Vasilenko, S. K.; FLorentiev, V. L.; Mirzabekov, A. D. DNA
Sequencing 1991, 1, 375-388.
(4) (a) Maskos, U.; Southern, E. M. Nucleic Acids Res. 1992, 20,
1675-8. (b) Maskos, U.; Southern, E. M. Nucleic Acids Res. 1992, 20,
1679-84. (c) Matson, R. S.; Rampal, J . B.; Coassin, P. J . Anal. Biochem.
1994, 217, 306-10. d. Blanchard, A. P.; Kaisern R. J .; Hood, L. E.
Biosensors Bioelectronics 1996, 11, 687-90.
(5) (a) Pease, A. C. Solas, D.; Sullivan, E. J .; Cronin, M. T.; Holmes,
C. P.; Fodor, S. P. A. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 5022-6.
(b) Lipshutz, R. J .; Morris, D.; Chee, M.; Hubbell, E.; Kozal, M. J .;
Shah, N.; Shen, N.; Yang, R.; Fodor, S. P. A. BioTechniques 1995, 19,
442-7. (c) Lipshutz, R. J .; Fodor, S. P. A. Curr. Opin. Struct. Biol.
1994, 4, 376-80. (d) McGall, G.; Labadie, J .; Brock, P.; Wallraff, G.;
Nguyen, T.; Hinsberg, W. Proc. Natl. Acad. Sci. U.S.A. 1996, 93,
13555-13560.
(6) Pirrung, M. C.; Bradley, J .-C. J . Org. Chem. 1995, 60, 1116.
(7) Pirrung, M. C.; Bradley, J .-C. J . Org. Chem. 1995, 60, 6270.
(8) McGall, G. H.; Barone, A. D.; Diggelmann, M.; Fodor, S. P. A.;
Gentalen, E.; Ngo, N. J . Am. Chem. Soc. 1997, 119, 5081-90.
(2) Shumaker, J . M.; Metspalu, A.; Caskey, C. T. Hum. Mutat. 1996,
7, 346-354.
S0022-3263(97)00872-4 CCC: $15.00 © 1998 American Chemical Society
Published on Web 01/01/1998

242 J . Org. Chem., Vol. 63, No. 2, 1998
Pirrung et al.
Ch a r t 1
F igu r e 1. Surface fluorescence image of surface-bound DM-
BOC-protected isobutyryldeoxycytidine 7 after deprotection at
365 nm for up to 200 s followed by staining with fluorescein
phosphoramidite.
from the surface. We also earlier showed that benzoyl
C was photoreactive,7 so the benzoyl protecting group had
to be avoided.
Sch em e 1
Resu lts
Synthesis of three novel nucleoside phosphoramidite
monomers 2-4 (Chart 1) needed for this study was
accomplished along lines parallel to those used in our
earlier work. That is, the azeotropically dried, base-
protected A, G, and C nucleosides were treated with
(dimethoxybenzoin)carbonylimidazolium triflate to se-
lectively derivatize the 5′-hydroxyl. The 3′-hydroxyl was
converted to the 2-cyanoethyl phosphoramidite using
commercial reagents according to literature protocols.10
These compounds were purified by repetitive precipita-
tion from 6:1 hexanes/methylene chloride. Compound 1
was prepared in our earlier work.
Hydroxyalkylsilanated glass microscope slides5a were
derivatized with a poly(ethylene glycol) linker terminated
with a MeNPOC photoremovable nitrobenzyl group by
coupling with MeNPOC-PEG-2-cyanoethyl phosphor-
amidite.8 The resulting substrate (5) was clamped in a
sealed flowcell through which a modified automated DNA
synthesizer could deliver reagents.
The rate of photolysis at two different wavelengths of
these monomer units when bound to the glass surface
was first determined by the protocol described in Scheme
1, wherein the black regions indicate the area of the
surface irradiated in a given step.8 Rear irradiation of
this surface (contacted with dioxane) through a 0.8 × 12.8
mm horizontal striped photolithographic mask was used
to generate (in quadruplicate) parallel regions of free
hydroxyl groups. Successive coupling of phosphoramid-
ites 1-4 to the hydroxylated regions was performed. This
resulted in surface 6. It was then irradiated through a
single striped mask with a 0.4 × 12.8 mm vertical feature
for increasing time intervals (0-200 s at 280/310 nm,
0-800 s at 365/400 nm), translating the mask by 0.4 mm
between irradiation cycles. The resulting surface 7 was
then coupled to a fluorescein phosphoramidite, the
protecting groups were removed by brief base treatment,
and the slide was subjected to quantitative fluorescence
scanning using a confocal microscope. Examples of a
typical image and kinetic plot are shown in Figures 1
and 2. The data show a first-order exponential increase
(9) (a) Schulhof, J . C.; Molko, D.; Teoule, R. Bioact. Mol. 1987, 3,
143-8. (b) Schulhof, J . C.; Molko, D.; Teoule, R. Nucleic Acids Res.
1987, 15, 397-416. (c) Vu Huynh; McCollum, C.; J acobson, K.; Theisen,
P.; Vinayak, R.; Spiess, E.; Andrus, A. Tetrahedron Lett. 1990, 31,
7269-72.
(10) Sinha, N. D.; Biernat, J .; McManus, J .; Koster, H. Nucleic Acids
Res. 1987, 12, 4539-57.

Proofing of Photolithographic DNA Synthesis
J . Org. Chem., Vol. 63, No. 2, 1998 243
Sch em e 2
F igu r e 2. Plots of the data derived from Figure 1 of the
photodeprotection at 365 nm of surface-bound DMBOC-
protected (2) deoxyadenosine, (9) deoxyguanosine, ([) deoxy-
cytidine, and (]) thymidine in dioxane.
Ta ble 1. Ha lf-Lives for P h otoch em ica l Dep r otection of
DMB-ca r bon a te-P r otected Nu cleosid es on a Gla ss
Su r fa ce in th e P r esen ce of Dioxa n e
t1/2, s
DMB-carbonate nucleoside
310 nm
365 nm
dA
dG
dC
T
11
13
5.5
12
14
17
5.6
16
in fluorescence from which the half-lives were derived
by curve fitting. Data on wavelength dependence were
obtained with a mercury arc source filtered selectively
in the “near-UV” (primarily Hg lines at 365 and 405 nm)
and the “mid-UV” (lines at 280 and 310 nm). Sum-
marized in Table 1 are the data from these experiments.
The effect of solvent on the rate of deprotection was
examined for DMBOC-T and DMBOC-dCi-Bu. Surface
8 (Scheme 2) was prepared by uniform irradiation of 5
and coupling with 1. The kinetics of deprotection at 310
nm were determined using an experimental protocol
similar to that described in Scheme 1, in the presence of
dioxane, toluene, and methanol and without solvent.
Data are given in Table 2.
The coupling efficiency of amidites 1-4 in the synthesis
of homopolymers up to dodecamers was determined using
the protocol described in Scheme 2.8 Surface 8 and
analogues bearing the other three bases were prepared
by uniform irradiation of 5 and coupling with 1-4. A
photolithographic mask with a 6.4 × 12.8 mm vertical
stripe was used in a first irradiation and coupling cycle.
The irradiation was conducted for 8-10 half-lives based
on the values earlier determined. Coupling of the phos-
phoramidite was performed, the mask was translated by
492 µm, and the cycle was repeated. Due to overlap of
the areas of irradiation at each of the steps, this protocol
leads to the synthesis of dodecanucleotides at the center
of the slide, reflecting 12 coupling steps. The flanking
regions reflect monotonically decreasing numbers of
coupling cycles. After the full surface was coupled, half
of the array was subject to uniform irradiation to free
the 5′-hydroxyl groups, which were coupled with a
fluorescein phosphoramidite. Fluorescence imaging with
correction for background permitted the yield to be
determined for each cycle on the basis of the relative
fluorescence intensity of adjacent regions, whose oligo-
nucleotides differ in length by one base. In Figure 3 is
the fluorescence image for a sample (dC)n array. The
Ta ble 2. Solven t Dep en d en ce of Ha lf-Lives for Mid -UV
P h otoch em ica l Dep r otection of
DMB-ca r bon a te-P r otected Nu cleosid es on a Gla ss
Su r fa ce
t1/2
solvent
T
dC
dioxane
toluene
methanol
acetonitrile
dry
12
9.3
12
10
5.8
5.5
6.5
8.4
7.4
5.6
surface fluorescence decreases somewhat toward the
center of the pattern, as the yield of the full-length
oligomer decreases with increasing length. Quantitative
evaluation of the fluorescence intensity (I) data using eq
1 gives step yields.
% yield (step n) ) 100 × (In/In-1
)
(1)
Data for cycle yields and calculated aggregate yields
for homopolymer synthesis from the experiment in
Scheme 2 are tabulated in the Supporting Material.
Representative data are shown in Figure 4. The ranges
are as follows: T, 91 f 98%; C, 82 f 95%; G, 79 f 92%;
A, 74 f 84%.
The amidites 1-4 were used in the preparation of a
small oligonucleotide array for a hybridization study.
Sixteen decanucleotides of the sequence 5′-AAXTAXC-
TAC-chip were prepared, where the positions marked
by X comprise all combinations of nucleotides. These
oligonucleotides were prepared in the layout shown in
Figure 5, using striped masks with 3.2 × 12.8 mm
features, resulting in a 4 × 4 array of square hybridiza-

244 J . Org. Chem., Vol. 63, No. 2, 1998
Pirrung et al.
F igu r e 6. Fluorescence image of hybridization array surface
10 after staining at 23 °C with the probe fluorescein-5′-
CTGAACGGTAGCATCTTGAC at 10 nM concentration. The
region of complementarity with the target site in the array is
emboldened.
F igu r e 3. Fluorescence image of surface 9 after the synthesis
of an array of (dC)n homooligomers followed by staining with
fluorescein phosphoramidite.
Discu ssion a n d Con clu sion
Similar (only ∼2-fold different) DMBOC deprotection
half-lives were observed with 365 nm compared to 310
nm irradiation, though the intensity produced by the
mercury source is much higher at the longer wavelength.
This is not due simply to the lower absorptivity of the
chromophore in this region, since it is almost negligible
at λ > 350 nm. The quantum efficiency for removal of
DMBOC groups is increased at longer wavelength.11
Irradiation with wavelengths <340 nm should be avoided
in any event for oligonucleotide synthesis, based on the
potential photochemical damage to the DNA.12 Solvent
(including its complete omission) has only a modest
influence on the deprotection rate of surface-bound
DMBOC groups, with fastest rates observed without
solvent. This is not to say that the surface was in contact
with a vacuum during irradiation; it was rinsed with
anhydrous acetonitrile and dried under a stream of
argon, and indeed it is possible that some residual solvent
adheres to the surface. The data do show that solvent
is not a major influence with this protecting group. The
significant variation in the deprotection half-lives for the
four different nucleotides is somewhat surprising. It is
interesting to compare the half-lives determined in this
surface format with those earlier found13 for DMB-
phosphates (using a Rayonet reactor) in solution: T, 109
s; C, 105 s; A, 95 s; G, 81 s; 5′-Ac-T, 82 s; 5′-Ac-C, 71 s.
There is not a consistent pattern within this latter data
set, and trends between the sets are also not apparent.
It was also earlier observed on larger scale6,7 that the
rate slows with greater extent of deprotection, but the
concentration of photolabile groups in the chip format is
sufficiently low that this is not observed.
F igu r e 4. Plot of the calculated coupling efficiencies versus
cycle number for (dN)n homooligomer synthesis: (9) deoxy-
adenosine, (]) deoxyguanosine, ([) deoxycytidine, and (2)
thymidine.
F igu r e 5. Format of a 16-element hybridization array 10
based on the decanucleotide sequence 5′-AAXTAXCTAC-chip.
tion areas of dimensions 3.2 × 3.2 mm. The array was
hybridized at 23 °C with a target deoxyeicosanucleotide
of the sequence fluorescein-5′-CTGAACGGTAGCATCT-
TGAC, where the emboldened residues are complemen-
tary to one of the sequences (the G, G hybridization area)
of the surface 10. The fluorescence image of the hybrid-
ized probe is shown in Figure 6. The average signal-to-
background ratio of the hybridized region compared to
the nonmatching cells is >6.
(11) Cameron, J . F.; Willson, C. G.; Fre´chet, J . M. J . J . Chem. Soc.,
Chem. Commun. 1995, 923.
(12) Cadet, J .; Vigny, P. In Bioorganic Photochemistry, Vol. 1;
Morrison, H.; Ed.; Wiley: New York, 1990; p 1-272.
(13) Pirrung, M. C.; Shuey, S. W. J . Org. Chem. 1994, 59, 3890.

Proofing of Photolithographic DNA Synthesis
J . Org. Chem., Vol. 63, No. 2, 1998 245
anhydrous pyridine and cooled in an ice bath. After the
addition of 4.3 mL (33.9 mmol) of trimethylsilyl chloride, added
dropwise over 20 min, the reaction mixture was allowed to
stir for 1 h at 25 °C. In a separate round-bottom flask, 2.19 g
(16.21 mmol) of 1-hydroxybenzotriazole was azeotroped from
benzene and dissolved in 10 mL of dry CH3CN. Phenoxyacetyl
chloride (2.33 mL, 16 87 mmol) was slowly syringed into the
solution of hydroxybenzotriazole and the reaction was allowed
to stir for 5 min, when the mixture solidified. Pyridine and
CH3CN were added until the reaction mixture was homoge-
neous. The nucleoside was then transferred via cannula to
the acylating solution and the reaction was allowed to stir
overnight at 25 °C. The solution was cooled in an ice bath
and quenched by the addition of 10 mL of H2O followed by 5
mL of NH4OH. The solution was stirred for 20 min and
concentrated in vacuo. The residual oil was taken up in 100
mL of H2O and washed with 50 mL of CHCl3 (3×) and 50 mL
EtOAc (1×), upon which a lavender solid began to precipitate
out of the water. The solution was cooled in an ice bath and
the solid collected and dried in vacuo to give 1.32 g of
N-phenoxyacetyl-2′-deoxyguanosine (59%). 1H NMR (DMSO):
δ 2.24-2.33 (1H, m), 2.54-2.64 (1H, m), 3.50-3.60 (2H, m),
3.85 (1H, m), 4.39 (1H, m), 4.87 (2H, s), 4.98 (1H, m), 5.34
(1H, m), 6.23 (1H, t, J ) 6.9 Hz), 6.99 (3H, m), 7.32 (2H, m),
8.26 (1H, s).
To a solution of N-phenoxyacetyl-2′-deoxyguanosine (0.93
g, 2.32 mmol, azeotroped twice from pyridine) was added 4.6
mL of 0.5 M DMIT followed by 4 mL of dry pyridine. The
reaction mixture was stirred overnight at 25 °C and then
concentrated in vacuo. The residual oil was taken up in
dichloromethane and washed with saturated NaHCO3 and
saturated NaCl, dried over MgSO4, and concentrated. The
resulting oil was purified by flash chromatography on a silica
gel column using a step gradient of 2:98-5:95-10:90 EtOH/
CH2Cl2 to give 0.91 g of the desired compound (56%). 1H NMR
(CDCl3): δ 2.50 (1H, m), 2.60-2.80 (1H, m), 3.69 (6H, m),
4.25-4.80 (6H, m), 6.30 (1H, m), 6.36 (1H, m), 6.49 (2H, m),
6.54 (1H, m), 6.93 (2H, m), 7.05 (1H, m), 7.26-7.40 (4H, m),
7.46 (1H, m), 7.81 (2H, m), 7.95 (1H, m). UV (CH3CN): λmax
251 (ꢀ 18 800), 276 (ꢀ 13 000) nm (ꢀ280 14 200, ꢀ310 5800, ꢀ365
51, ꢀ405 16). Anal. Calcd for C35H33N5O11: C, 59.38; H, 4.84;
N, 10.19; O, 25.59. Found: C, 59.51; H, 4.86; N, 10.08.
N-P h en oxyacetyl-2′-deoxyaden osin e-5′-(3′′,5′′-dim eth ox-
yben zoin )ca r bon a te. To a solution of 2.00 g (7.96 mmol) of
2′-deoxyadenosine (azeotroped twice from pyridine) in 80 mL
of anhydrous pyridine, cooled in an ice bath, was added 5.10
mL (40.2 mmol) of trimethylsilyl chloride. The solution was
allowed to stir for 30 min at 25 °C, followed by the addition of
2.20 mL (15.9 mmol) of phenoxyacetyl chloride. The reaction
was stirred for 2 h at 25 °C and then quenched by the addition
of 100 mL of saturated NaHCO3. The pyridine was removed
in vacuo, followed by the addition of 300 mL of CH2Cl2. The
organic layers were combined and washed with saturated
NaCl, dried over MgSO4, and concentrated to give a yellow
foam. The foam was purified by flash chromatography on
silica gel using a step gradient of 5:95-10:90-15:85 EtOH/
CH2Cl2 to give 1.38 g of N-phenoxyacetyl-2′-deoxyadenosine
(46%). 1H NMR (DMSO): δ 2.32-2.39 (1H, m), 2.74-2.83 (1H,
m), 3.52-3.68 (2H, m), 3.91 (1H, m), 4.46 (1H, m), 5.04 (2H,
s), 5.38 (1H, m), 6.47 (1H, t, J ) 6.6 Hz), 6.95-7.00 (3H, m),
7.29-7.34 (2H, m), 8.69 (1H, s), 8.72 (1H, s).
While 98% cycle yields have been obtained in a best
case (T), none of the other heterocyclic bases is as
efficient. Cytidine can also approach this efficiency, but
cycle yields with purines can be lower. These coupling
efficiencies agree with those determined in our earlier
work in which the oligonucleotides were prepared on
controlled-pore glass. The cause of the base dependence
is unknown: we believe the base may influence the
course of the product-forming photoreactions and thereby
the final absolute yield of deprotected product (5′-OH).
While these results demand improvement in order to
maximize the fidelity of the in situ synthesized oligo-
nucleotides and therefore of nucleic acid hybridization,
it is interesting that the results of an actual hybridization
experiment show little difficulty with discriminating a
perfect hybrid from internal one-base mismatches despite
the imperfect sequence fidelity that is a consequence of
the less-than-quantitative cycle yields. The interest is
increased by the fact that the purines, which are less
efficient in the synthesis, are in the majority in the
perfectly matched hybridization probe. An earlier ex-
periment with a 16 × 16 array of octanucleotides gave
comparable results in terms of the fidelity of hybridi-
zation.5a It is clear that light-directed in situ synthesis
of short oligonucleotides will be an effective method to
prepare hybridization surfaces (“DNA chips”) for detec-
tion of complementary sequences. This work further
highlights the issues of coupling efficiency in DNA array
synthesis.
Exp er im en ta l Section
Gen er a l. The 2′-deoxynucleosides A, G, and T were ob-
tained from Cruachem. Protected 2′-deoxycytidine was ob-
tained from Sigma. DNA synthesis reagents were obtained
from Applied Biosystems. Fluorescein phosphoramidite (Fluo-
reprime) and DMTr-thymidine-CE-amidite were obtained
from Pharmacia. Other reagents were from Aldrich.
P r ep a r a tion of a 0.50 M Solu tion of (3′′,5′′-Dim eth oxy-
ben zoin )ca r bon ylim id a zoliu m Tr ifla te (DMIT) in Ni-
tr om eth a n e. To a solution of 1.96 g (11.96 mmol) of carbo-
nyldiimidazole (azeotroped twice from benzene) in 19.0 mL of
nitromethane, cooled in an ice bath, was added 2.71 mL (23.95
mmol) of methyltrifluoromethanesulfonate dropwise over 10
min. The solution was stirred for 30 min at 25 °C and then
transferred via cannula to a flask containing 3.26 g (11.97
mmol) of dimethoxybenzoin (azeotroped twice from benzene).
The reaction mixture was allowed to stir for 3 h at 25 °C.
Complete formation of the DMIT reagent was confirmed by
1H NMR.
N-Isobu tyr yl-2′-d eoxycytid in e-5′-(3′′,5′′-d im eth oxyben -
zoin )ca r bon a te. To a solution of N-isobutyryl-2′-deoxycyti-
dine (0.83 g, 2.80 mmol, azeotroped twice from pyridine) was
added 3.32 mL of 0.5 M DMIT followed by 2 mL of dry
pyridine. The reaction mixture was stirred at 25 °C for 8 h
and then concentrated in vacuo. The residual oil was taken
up in dichloromethane and washed with saturated NaHCO3
and saturated NaCl, dried over MgSO4, and concentrated. The
resulting oil was purified by flash chromatography on a silica
gel column with EtOH/CH2Cl2 (4:96 v/v) to give 0.83 g of the
desired compound (50%). 1H NMR (CDCl3): δ 1.21 (6H, m),
2.10-2.23 (1H, m), 2.58-2.85 (2H, m), 3.76 (6H, m), 4.20-
4.60 (4H, m), 6.35 (1H, m), 6.42 (1H, m), 6.57 (2H, m), 6.64
(1H, m), 7.38-7.60 (3H, m), 7.94 (2H, m), 8.08 (1H, d, J ) 7.5
Hz), 8.17 (1H, d, J ) 7.5 Hz), 8.5 (1H, br s). HRMS (FAB,
MH+) calcd for C30H34N3O10: 596.2244. Found: 596.2232. UV
(CH3CN): λmax 245 (ꢀ 10 350), 295 (ꢀ 3740) nm (ꢀ280 3400, ꢀ310
3200, ꢀ365 30, ꢀ405 5).
To a solution of N-phenoxyacetyl-2′-deoxyadenosine (1.20
g, 3.11 mmol, azeotroped twice from pyridine) was added 6.23
mL of 0.5 M DMIT followed by 10 mL of dry pyridine. The
reaction mixture was stirred overnight at 25 °C and then
concentrated in vacuo. The residual oil was taken up in
dichloromethane and washed with saturated NaHCO3 and
saturated NaCl, dried over MgSO4, and concentrated. The
resulting oil was purified by flash chromatography on a silica
gel column using a step gradient of 2:98-6:94 EtOH/CH2Cl2
to give 0.58 g of the desired compound (30%). 1H NMR
(CDCl3): δ 2.55-2.66 (1H, m), 2.86-2.95 (1H, m), 3.68-3.75
(6H, m), 4.27-4.56 (3H, m), 4.73-4.89 (3H, m), 6.41 (1H, m),
6.56 (1H, m), 6.59 (2H, m), 6.64 (1H, s), 7.05 (3H, m), 7.30-
7.60 (5H, m), 7.91 (2H, m), 8.32 (1H, m), 8.77 (1H, s), 9.40
N-P h en oxyacetyl-2′-deoxygu an osin e-5′-(3′′,5′′-dim eth ox-
yben zoin )ca r bon a te. 2′-Deoxyguanosine (1.5 g, 5.61 mmol,
azeotroped twice from pyridine) was dissolved in 30 mL of

246 J . Org. Chem., Vol. 63, No. 2, 1998
Pirrung et al.
(1H, br s). HRMS (FAB, MH+) calcd for C35H34N5O10: 684.2305.
Found: 684.2308. UV (CH3CN) λmax 271 (ꢀ 28300), 256 (ꢀ
27500) nm (ꢀ280 21200, ꢀ310 1900, ꢀ365 68, ꢀ405 19).
UV Sp ectr a . To enable the relative photoefficiencies of the
different compounds under different experimental conditions
to be calculated, detailed UV data are provided for the
foregoing compounds and the known thymidine-5′-(3′′,5′′-
dimethoxybenzoin)carbonate): UV (CH3CN) λmax 250 (ꢀ 21 400)
nm (ꢀ280 8900, ꢀ310 1140, ꢀ365 22).
droxyethyl)(aminopropyl)triethoxysilane as described previ-
ously8
and then by adding to the surface a protected linker,
MeNPOC-hexaethylene glycol-(2-cyanoethyl) N,N-diisopro-
pylphosphoramidite. A 50 mM solution of each nucleoside
phosphoramidite monomer was prepared in anhydrous CH3-
CN. All phosphoramidite coupling reactions were performed
on a custom-built automatic exposure tool and flowcell inter-
faced to a modified Applied Biosystems 392 DNA synthesizer
for reagent delivery. The desired synthesis order containing
coupling and photolysis cycle data was programmed into a
computer interfaced with both the synthesizer and lithographic
equipment. The synthesis program consisted of the base
sequence to be coupled to the surface, the mask position during
each coupling step, and the irradiation time. Minor adjust-
ments were made to the standard instrument coupling pro-
gram to accommodate the particular volume and mixing
requirements of the flowcell, to eliminate the detritylation step,
and to pause for automated mask positioning and exposure.
Exposures were made through a chrome-on-quartz mask in
contact with the back of the substrate. Light was from a 500
W collimated mercury arc light source (model 87330, Oriel
Instruments, Stratford, CT) with interchangeable dichroic
reflectors providing output in either the mid-UV (280 nm/0.55
mW/cm2, 310 nm/5.0 mW/cm2) or near-UV (365 nm/44.5 mW/
cm2, 405 nm/98 mW/cm2) range. The glass substrate was held
in a flow cell under argon allowing DNA coupling reagents to
be delivered to the surface under anhydrous conditions. The
output of the UV light source was measured through the
photolithographic mask and glass substrate to be 5.0 mW/cm2
at 310 nm and 44.5 mW/cm2 at 365 nm. After the synthesis
was complete, all substrates were deprotected in a 50%
solution of 1,2-diaminoethane or ethanolamine in ethanol for
1 h at RT to remove all base, phosphate, and fluorescein
protecting groups, then rinsed with deionized water, and dried
under a stream of nitrogen. Deprotected surface hydroxyl
groups were visualized and quantitated by fluorescence stain-
ing, wherein the chip is treated with a solution of 5 mM
fluorescein CE-amidite/50 mM DMTr-thymidine-CE-amidite
in acetonitrile, with tetrazole, and subjected to conventional
oxidation.
N-Isobu tyr yl-2′-d eoxycytid in e-5′-(3′′,5′′-d im eth oxyben -
zoin )ca r b on a t e -3′-cya n oe t h yl(d iisop r op yl)p h osp h or -
a m id ite (2). To a solution of N-i-Bu-5′-(DMBOC)-2′-deoxy-
cytidine (0.70 g, 1.18 mmol, azeotroped twice from benzene)
in 20 mL of anhydrous CH2Cl2, cooled in an ice bath, was
added 410 µL (2.35 mmol) of diisopropylethylamine and 315
µL (1.41 mmol) of (2-cyanoethoxy)(diisopropylamino)chloro-
phosphine. The reaction was allowed to stir overnight at 25
°C. The reaction mixture was taken up in EtOAc and washed
with 10% Na2CO3 and saturated NaCl, dried over Na2SO4, and
concentrated in vacuo. Purification of the phosphoramidite
was accomplished by repetitive trituration of the residual oil
from 6:1 hexane/CH2Cl2 to give 0.65 g of the desired compound
as a white foam (70%). 1H NMR (CDCl3): δ 1.14-1.29 (20H,
m), 2.20 (1H, m), 2.60-2.80 (4H, m), 3.59 (2H, m), 3.74 (6H,
s), 4.25-4.75 (4H, m), 6.32 (1H, m), 6.42 (1H, m), 6.57 (2H,
m), 6.65 (1H, m), 7.42 (2H, m), 7.53 (1H, m), 7.93 (2H, m),
8.04 (1H, m), 8.15 (1H, m), 8.39 (1H, br s). 31P NMR (CDCl3):
δ 149.94, 149.57, 149.38, 150.09. HRMS (FAB, MH+) calcd
for C39H51N5O11P: 796.3322. Found: 796.3295.
N-P h en oxyacetyl-2′-deoxygu an osin e-5′-(3′′,5′′-dim eth ox-
yb e n zoin )ca r b on a t e -3′-cya n oe t h yl(d iisop r op yl)p h os-
p h or a m id ite (3). To a solution of N-phenoxyacetyl-5′-(DM-
BOC)-2′-deoxyguanosine (0.75 g, 1.07 mmol, azeotroped twice
from benzene) in 20 mL of anhydrous CH2Cl2, cooled in an ice
bath, was added 374 µL (2.14 mmol) of diisopropylethylamine
and 287 µL (1.29 mmol) of (2-cyanoethoxy)(diisopropylamino)-
chlorophosphine. The reaction was allowed to stir overnight
at 25 °C. The reaction mixture was taken up in EtOAc and
washed with 10% Na2CO3 and saturated NaCl, dried over Na2-
SO4, and concentrated in vacuo. Purification of the phos-
phoramidite was accomplished by repetitive trituration of the
residual oil from 6:1 hexane/CH2Cl2 to give 0.64 g of the desired
compound as a white foam (66%). 1H NMR (CDCl3): δ 1.15-
1.30 (14H, m), 2.55 (1H, m), 2.65 (2H, m), 2.90 (1H, m), 3.66
(2H, m), 3.74 (6H, m), 4.30-4.80 (6H, m), 6.28 (1H, m), 6.40
(1H, m), 6.53 (2H, s), 6.59 (1H, m), 6.95-7.12 (3H, m), 7.30-
7.55 (5H, m), 7.80-7.98 (3H, m). 31P NMR (CDCl3): δ 149.71,
149.61, 149.57. HRMS (FAB, MH+) calcd for C44H51N7O12P:
900.3333. Found: 900.3312.
N-P h en oxyacetyl-2′-deoxyaden osin e-5′-(3′′,5′′-dim eth ox-
yb e n zoin )ca r b on a t e -3′-cya n oe t h yl(d iisop r op yl)p h os-
p h or a m id ite (4). To a solution of N-phenoxyacetyl-5′-(DM-
BOC)-2′-deoxyadenosine (0.54 g, 0.79 mmol, azeotroped twice
from benzene) in 12 mL of anhydrous CH2Cl2, cooled in an ice
bath, was added 276 µL (1.58 mmol) of diisopropylethylamine
and 212 µL (0.949 mmol) of (2-cyanoethoxy)(diisopropylamino)-
chlorophosphine. The reaction was allowed to stir overnight
at 25 °C. The reaction mixture was taken up in EtOAc and
washed with 10% Na2CO3, saturated NaCl, dried over Na2-
SO4 and concentrated in vacuo. Purification of the phosphor-
amidite was accomplished by repetitive trituration of the
residual oil from 6:1 hexane/CH2Cl2 to give 0.50 g of the desired
compound as a white foam (71%). 1H NMR (CDCl3): δ 1.19-
1.29 (14H, m), 2.60-3.0 (4H, m), 3.63 (2H, m), 3.75 (6H, m),
4.35-4.60 (3H, m), 4.83 (1H, m), 4.88 (2H, s), 6.42 (1H, m),
6.52 (1H, m), 6.60 (2H, m), 6.65 (1H, m), 7.30-7.60 (5H, m),
7.93 (2H, m), 8.33 (1H, m), 8.79 (1H, m), 9.39 (1H, m). 31P
NMR (CDCl3): δ 149.92, 149.89, 149.59, 149.47. HRMS (FAB,
MH+) calcd for C44H51N7O11P: 884.3384. Found: 884.3366.
DNA Syn th esis. Glass substrates were prepared for oli-
gonucleotide synthesis by derivatization with N,N-bis(hy-
Im a gin g. The pattern and intensity of the surface fluo-
rescence was imaged with a specially constructed scanning
laser fluorescence microscope, with 488 nm argon ion laser
excitation. Emitted light was collected through confocal optics
with a 530 ( 15 nm band-pass filter and detected with a
photomultiplier tube equipped for photon counting. Relative
quantitation of surface-bound fluorescein molecules in various
regions of the substrate was taken directly from output
intensity values in photon counts/s. All values were corrected
for nonspecific background fluorescence, measured as the
surface fluorescence in nonilluminated regions of the substrate.
Hybr id iza tion . Array hybridization was carried out in a
flowcell fixed to the stage of the microscope with the labeled
oligonucleotide at 10 nm concentration in 6× SSPE buffer (0.9
M NaCl, 60 mM NaH2PO4, 6 mM EDTA, pH 7.5) for 60 min.
After removal of the oligonucleotide solution, the array was
washed briefly with 6× SSPE, and the image was obtained.
Ack n ow led gm en t. Financial support was provided
by NIH GM-46720. We thank S. Manyem for obtaining
UV spectra.
Su p p or tin g In for m a tion Ava ila ble: Table of coupling
yields (1 page). This material is contained in libraries on
microfiche, immediately follows this article in the microfilm
version of the journal, and can be ordered from the ACS; see
any current masthead page for ordering information.
J O970872S