ABSTRACT
Four types of polyacrylamide or polydimethylacrylamide gels for regioselective
(by immobilization at the 3
'
end) of short oligonucleotides have been designed for use in manufacturing
oligonucleotide microchips. Two of these supports contain amino or aldehyde
groups in the gel, allowing coupling with oligonucleotides bearing aldehyde or
amino groups, respectively, in the presence of a reducing agent. The aldehyde
gel support showed a higher immobilization efficiency relative to the amino
gel. Of all reducing agents tested, the best results were obtained with a
pyridine-borane complex. The other supports are based on an acrylamide gel
activated with glutaraldehyde or a hydroxyalkyl-functionalized gel treated with mesyl chloride. The use of
dimethylacrylamide instead of acrylamide allows subsequent gel modifications in organic solvents. All the immobilization methods are easy and
simple to perform, give high and reproducible yields, allow long durations of
storage of the activated support, and provide high stability of attachment and
low non-specific binding. Although these gel supports have been developed for
preparing oligonucleotide microchips, they may be used for other purposes as
well.
Sequencing by hybridization (SBH) on oligonucleotide matrix or microchip
containing a complete set of octanucleotides (or oligonucleotides of other
reasonable length) is a new analytical method which was originally developed
for fast sequencing of genomic DNA. Despite certain limitations, biological
microchips show considerable promise as diagnostics for genetic diseases (for
review see
1
).
There are two alternative approaches to microchip preparation. In one approach,
oligonucleotide probes are synthesized directly on a flat surface using either
conventional chemical procedures (
2
,
3
) or photolithographic VLSIPS technology (
4
). Over the last few years we have developed another approach based on the
immobilization of presynthesized, purified oligonucleotides in microcells of a
functionalized polyacrylamide gel (
5
,
6
). The advantages of our approach are the higher capacity of the support and the
possibility of immobilizing different amounts of oligonucleotides or
oligonucleotides of different length in different cells; in particular, these
features enable us to design a `normalized' microchip that yields equal hybridization signals for
sequences of different G+C content by adjusting the concentrations of
immobilized oligonucleotides (
7
).
Oligonucleotides are immobilized upon their 3'-dialdehyde termini by reaction with the gel hydrazide groups
(
Scheme 1
). Unfortunately, the hydrazide chemistry does not provide sufficient stability of attachment in repeated hybridization experiments. A variety of
methods have been described for immobilization of biomolecules (
8
,
9
); and for attachment of oligonucleotides to polyacrylamide gels (
10
), some of them are well developed and corresponding gel supports are
commercially available. However, functionalized gels for oligonucleotide
microchips should meet some specific requirements, such as easy preparation and
modification of the gel; high and reproducible immobilization yield; long
storage life of the activated support; and high stability of attachment, and
low nonspecific binding. Here we describe new functionalized acrylamide gels for
oligonucleotide immobilization suitable for SBH applications.
All solvents were analytical or HPLC grade. General reagents were purchased from
Fluka and Sigma Chemical Co. Glass slides for matrix preparation were from
Corning Glass Works.
1
H NMR spectra were recordered on a Varian XL100, 100 MHz spectrometer; UV spectra were recorded on a Shimadzu UV-160A spectrophotometer.
Deoxyoligonucleotides were synthesized on an Applied Biosystems ABI 392-05 synthesizer using standard phosphoramidite chemistry.
N
-Me-uridine CPG (
11
) and amine CPG (Clontech) were used for the introduction of the methyluridine
unit and the amino group into the 3' end of the oligonucleotides, respectively. Purification of
oligonucleotides was performed by reverse phase HPLC (Gilson) using the
following conditions: C
18
Nucleosil column, 7 [mu]m, 10-50%, CH
3
CN in 0.1 M triethylammonium acetate (TEAA), 30 min for oligonucleotides containing DMT protection groups, 0-25%, 30 min for unprotected oligonucleotides.
Radioactive labeling of oligonucleotides at their 5'
ends with [[gamma]-
32
P]ATP was performed following standard protocol (
12
). The labeled oligonucleotides were isolated by electrophoresis on a denaturing
polyacrylamide gel. 3'-methyluridine oligonucleotide was oxidized with 0.1 M NaIO
4
for 15 min at room temperature before immobilization to generate the 3'-dialdehyde unit.
Thin layer chromatography was carried out on Merck Kieselgel 60 TLC plates.
Solvent systems for TLC: chloroform-methanol (9:1, system A); chloroform-methanol-hexane (9:1:10, system B); isopropyl alcohol-NH
4
OH-H
2
O (7:2:1, system C).
To a solution of acryloyl chloride (8.5 ml, 105 mmol) in dry dichloromethane
(250 ml) cooled to 0oC, a mixture of 3-aminopropyltriethoxysilane (23.5 ml, 100 mmol) and triethylamine
(13.9 ml, 100 mmol) in dichloromethane (50 ml) was added dropwise. After
completion of the addition, the reaction mixture was stirred for 30 min and
then filtered. The filtrate was evaporated to oil, diluted with hexane (100 ml)
and filtered. The filtrate was concentrated to oil and distilled
in vacuo
. The product was obtained as a viscous oil (21.2 g, 77%), b.p. 142-145oC at 1 mm Hg.
1
H NMR (CDCl
3
): 0.72 (t, 2H, SiCH
2
), 1.27 (t, 9H, CH
3
), 1.73 (m, 2H, CH
2
), 3.39 (m, 2H, CH
2
NH), 3.88 (q, 6H, OCH
2
), 5.67-6.35 (m, 3H, CH=CH
2
), 6.35 p.p.m. (brs, 1H, NH).
A solution of ethylenediamine (4.2 ml, 63 mmol) in dry ether (25 ml) was added
slowly to a cooled (0oC) solution of acryloyl chloride (2.25 ml, 65 mmol) in dry ethyl ether (200 ml) under intensive stirring.
After completion of the addition, (2-aminoethyl)acrylamide hydrochloride was filtered, washed with ether on the filter, and dried. The yield was 9.4 g (100%).
1
H NMR (DMSO-d
6
): 3.02 (s, 2H, CH
2
NH
2
HCl), 3.12 (m, 2H, CH
2
NH), 5.29-6.28 (m, 3H, CH=CH2), 7.97 p.p.m. (brs, 1H, NH).
Compound
Ib
was synthesized following the procedure described for
Ia
, with a 99% yield.
1
H NMR (DMSO-d
6
): 1.36 (brs, 8H, 4CH
2
), 2.78 (t, 2H, CH
2
NH
2
HCl), 3.10 (m, 2H, CH
2
NH), 5.48-6.40 (m, 3H, CH=CH
2
), 8.10 p.p.m. (brs, 1H, NH).
1,2,6-Trihydroxyhexane (11.7 g, 87 mmol) was refluxed in the chloroform-acetone mixture (3:1 v/v, 300 ml) in the presence of catalytic
amounts of
p
-toluenesulfonic acid for 6 h with azeotropic removal of water. The
resulting solution was neutralized with 10 ml of triethylamine and washed with 10% aqueous sodium bicarbonate. The
organic solution was dried over sodium sulfate and concentrated. The residual
oil was distilled, yielding 11.9 g (79%) of 5,6-isopropylidenedioxahexane-1-ol (
IIa
). R
f
0.45, solvent A.
1
H NMR (CDCl
3
): 1.64 (s, 3H, CH
3
), 1.69 (s, 3H, CH
3
), 1.64-2.00 (m, 6H, 3CH
2
), 2.34 (brs, 1H, OH), 3.56-4.28 p.p.m. (m, 5H, CH, CH
2
, CH
2
OH). Compound
IIa
(10 g, 57 mmol) in Py (100 ml) was reacted with CH
3
SO
2
Cl (6.6 g, 57 mmol, 4.5 ml) at room temperature for 12 h with stirring. The
reaction was quenched with water (10 ml) and the reaction mixture was
concentrated. The residue was diluted with chloroform, washed with water, dried
over Na
2
SO
4
, and concentrated. The product, 5,6-isopropylidene-dioxahexane-1-mesylsulfonate (
IIb
) (11.92 g, 83%), was used without further purification. R
f
0.7, solvent A.
1
H NMR (CDCl
3
): 1.64 (s, 3H, CH
3
), 1.69 (s, 3H, CH
3
), 1.64-2.35 (m, 6H, 3CH
2
), 3.18 (s, 3H, SO
2
CH
3
), 3.56-4.28 (m, 3H, CH, CH
2
), 4.34 p.p.m. (t, 2H, CH
2
OMs). Mesilate
IIb
(11.22 g, 44 mmol) was dissolved in dry DMF (10 ml), and LiN
3
(4.4 g, 88 mmol) was added. The reaction mixture was heated at 150oC for 30 min and then cooled and partitioned between water and ether
(100/100 ml). The organic layer was dried and evaporated, yielding 5,6-isopropylidenedioxahexane-1-azide (
IIc
) (7.7 g, 88%), R
f
0.7, solvent B.
1
H NMR (CDCl
3
): 1.64 (s, 3H, CH
3
), 1.69 (s, 3H, CH
3
), 1.64-2.35 (m, 6H, 3CH
2
), 3.18 (s, 3H, SO
2
CH
3
), 3.47 (t, 2H, CH
2
N
3
), 3.56-4.28 p.p.m. (m, 3H, CH, CH
2
). To the solution of compound
IIc
(7.5 g, 37 mmol) in pyridine (100 ml), Ph
3
P (10.1 g, 37 mmol) was added under stirring. After 3 h, nitrogen liberation was
completed. Aqueous ammonia (100 ml) was then added and the solution was kept at
room temperature for 12 h. The reaction mixture was concentrated and diluted
with water (100 ml). Ph
3
PO was filtered and the aqueous solution was evaporated. The residue was
coevaporated several times with acetonitrile to remove water and distilled
in vacuo
. 5,6-Isopropylidene-dioxahexane-1-amine (
IId
) was obtained as colourless liquid (3.8 g, 60%), b.p. 110-113oC at 15 mm. Hg. R
f
0.8, solvent C.
1
H NMR (CDCl
3
): 1.64 (s, 3H, CH
3
), 1.69 (s, 3H, CH
3
), 1.60-2.25 (m, 6H, 3CH
2
), 2.05 (s, 2H, NH
2
), 2.92 (t, 2H, CH
2
N), 3.56-4.28 p.p.m. (m, 3H, CH, CH
2
). A mixture of amine
IId
(1.03g, 5.9 mmol) and triethylamine (0.83 ml, 5.9 mmol) in dichloromethane (5
ml) was added dropwise to a vigorously stirred solution of acryloyl chloride
(0.53 ml, 6.5 mmol) in dichloromethane (5 ml) at 0oC. After 30 min, hexan (15 ml) was added to the reaction mixture.
Triethylamine hydrochloride was filtered off and the filtrate was evaporated.
N
-(5,6-di-
O
-isopropylidene)hexylacrylamide (
II
) was obtained as a viscous liquid (1.31 g, 97%), R
f
0.5, solvent A.
1
H NMR (CDCl
3
): 1.66 (s, 3H, CH
3
), 1.71 (s, 3H, CH
3
), 1.60-2.1 (m, 6H, 3CH
2
), 3.40-4.32 (m, 5H, CH, CH
2
, CH
2
N), 5.62-6.42 p.p.m. (m, 3H, CH=CH
2
).
A
chemically cleaned (washed in concentrated H
2
SO
4
, then with distilled water, and dried) glass slide was treated with Repel-Silane (Pharmacia) for 1-2 min and then washed with distilled water. After drying in a
nitrogen stream, it was used as a hydrophobic cover glass. A second glass plate
of the same surface quality was treated for 15-20 min with 5-10% solution of 3-(triethoxysilyl)propylacrylamide in 95% alcohol. After this,
it was washed intensively with alcohol and dried in a nitrogen stream. This
glass was used further as the gel binding glass. Two 20 [mu]m (or other size) teflon spacer strips were put on the sides of glass and
then covered with the hydrophobic glass slide. The glass `sandwich' was fixed
with two climbs. The monomer mixture was filtered before use. TEMED (0.1% v/v)
was added to the monomer solution. The polymerizing mixture was then poured
between the glass plates and was allowed to stand for 15-20 min. After completion of polymerization, the glass sandwich was rinsed
with distilled water and carefully taken apart.
A monomer mixture of the following content was used: 0.9 M
N
,
N
-dimethylacryamide, 0.1 M
N
-(2-aminoethyl)acrylamide hydrochloride, 0.02 M
N
,
N
-methylene-bis-acrylamide and 0.5-1 mg/ml ammonium persulfate. Before immobilization,
amino matrices were washed in 0.1 M KOH for 5 min (to generate free amines), in
distilled water for 30 min, and were then dried.
The monomer mixture contained 0.9 M
N
,
N
-dimethylacrylamide, 0.1 M
N
-(5,6-di-
O
-isopropylidene)hexylacrylamide, 0.02 M
N
,
N
-methylene-bis-acrylamide and 0.5-1 mg/ml ammonium persulfate. Aldehydes were generated
by treatment with 80% acetic acid at 60oC for 1 h and then with 0.1 M NaIO
4
at room temperature for 30 min. The aldehyde matrix was then washed at 30oC for 1 h in distilled water and dried.
The monomer mixture contained 1 M acryamide, 0.02 M
N
,
N
-methylene-bis-acrylamide and 0.5-1 mg/ml ammonium persulfate. Gel chips were treated
with hydrazine hydrate at room temperature for 12 h, then washed with 5% acetic
acid for 1 h and with distilled water for 1 h, and dried.
The monomer mixture contained 0.9 M
N
,
N
-dimethylacryamide, 0.1 M 2-hydroxyethylacrylate, 0.02 M
N
,
N
-methylene-bis-acrylamide and 0.5-1 mg/ml ammonium persulfate. The polymerized gel matrix
was treated with hydrazine hydrate at room temperature for 12 h, then washed
with 5% acetic acid for 1 h and with distilled water for 1 h, and dried.
The monomer mixture contained 0.9 M
N
,
N
-dimethylacryamide, 0.1 M 2-hydroxyethylacrylate, 0.02 M
N
,
N
-methylene-bis-acrylamide and 0.5-1 mg/ml ammonium persulfate. Polymerized gel matrix was
treated with 50% 3-aminopropanol in EtOH at 50oC for 6 h. After a 1 h wash in distilled water at 30oC, the matrix was dried and treated with 1 M CH
3
SO
2
Cl in dry pyridine for 8 h at room temperature. Then matrix was washed for 5 min
with ethanol, for 1 h with distilled water, and dried.
The monomer mixture contained 1 M acryamide, 0.02 M
N
,
N
-methylene-bis-acrylamide and 0.5-1 mg/ml ammonium persulfate. Gel chips were treated
with a mixture of glutaraldehyde (25% aqeous solution) and phosphate buffer pH 7.5 (1:1) at 40oC for 12 h, were washed with distilled water for 1 h, and then were dried.
The following conditions were used for immobilization. Hydrazide matrices: 1 [mu]l of oxidized oligonucleotide in water, overnight; amino matrix (or aldehyde
matrix): 1 [mu]l of oxidized oligonucleotide (or amino-oligonucleotide for aldehyde support) in water plus 1 [mu]l of reducing agent (0.1 M solution in water for NaCNBH
3
and Me
3
NBH
3
, 0.1 M solution in 10% aqueous methanol for PyBH
3
), 1.5 h; methanesulfonate-activated matrix: 1 [mu]l of amino-oligonucleotide in 0.1 M K
2
CO
3
, overnight; glutaraldehyde activated matrix: 1 [mu]l of amino-oligonucleotide in water, overnight. Immobilization was carried out at
room temperature in a chamber at 100% humidity. Unmodified oligonucleotides in
the same solutions as the immobilized oligonucleotides were used as a control.
Each spot (~1 mm in diameter) containing immobilized or control oligonucleotide had 15
000-25 000 c.p.m. of activity and contained 200-500 pmol of oligonucleotide. After immobilization, each matrix was
washed for 20 min with 0.1 M TEAA at 50oC.
Unlike some other applications in which gel beads can be used without any
additional support, gel-based SBH on microchips requires solid support for the gel cells.
Conventionally in gel microchip preparation, acrylamide or another acrylic
monomer polymerizes in the form of a thin film on a flat glass surface
derivatized with acrylic groups. The polymer film attached to glass is then
used for chemical treatment (hydrazinolysis or other), cell preparation and immobilization. Insufficient resistance of the linkage between gel and glass towards modifying reagents may result in exfoliation of cells. The commonly used methacryloxypropyltriethoxysilane (Bind-Silane) modification of the surface gives an attachment of the gel that is
unstable to hydrazine treatment. We found that 3-(triethoxysilylpropyl)acrylamide provides a much better resistance to
drastic conditions of gel treatment. This compound was synthesized starting
from 3-aminopropyltriethoxysilane and acryloyl chloride. Several hours of exposure to hydrazine hydrate at
room temperature or to 50% 3-aminopropanol in ethyl alcohol at 50oC did not cause any damage or exfoliation of a 20 [mu]m polydimethylacrylamide film.
The common methods for polyacrylamide (and polymorpholino- acrylamide) gel functionalization are based on treatment of the
polymerized support with reagents such as hydrazine or ethylenediamine (
13
-
15
). In routine microchip production, we also use hydrazinolysis of polymerized
acrylamide films (Scheme 1
). In this work, we introduced the functional groups by copolymerization
reactions and by subsequent modification of the active groups in the copolymer.
Some examples of functionalized polyacrylamide copolymers for coupling of
glycoprotein and enzymes have been described in which
N
-hydroxysuccinimide acrylic derivatives were used (
11
,
16
). Commercially available oxirane acrylic support (Sigma) is another example of
a derivatized copolymer. The copolymerization approach offers easy, controlled
modification of acrylic polymers. The use of substituted acrylamides
(dimethylacrylamide or morpholinoacrylamide) provides an additional possibility
for subsequent modification both in aqueous and organic media. Thus, a variety
of functional groups can be generated in the gel, or the linker length can be
increased by the usual synthetic methods.
Immobilization was carried out with 5'-
32
P-labeled oligonucleotides: ATGCTACT-X, where X is an oxidized ribose unit from the methyluridine
fragment or an NH
2
group (oligoR or oligoA, respectively). An amino group was attached to the 3'
end of the oligonucleotide through a C
6
linker. The unmodified octanucleotide (oligoP) was used as a control. The
chemistry of immobilization with simultaneous reduction is shown in Schemes
2
and
5 . The efficiency of binding was estimated by measuring radioactivity before and
after washing for 20 min in 0.1 M TEAA buffer at 50oC. As expected, both the amino and aldehyde matrices provided fast
attachment in high yield (Table
1
). The slightly lower efficiency of the amino matrix is probably caused by
partial elimination [E1cB mechanism (
18
)] of the 3'-dialdehyde moiety as pH rises in the course of reduction. The highly efficient aldehyde support exhibited, however, an increased
non-specific binding of the control octamer. We used sodium cyanoborohydride,
pyridine-borane and trimethylamine-borane complexes as reducing agents. The best results were obtained with
pyridine-borane complex for both the amino and the aldehyde supports (yields were
74 and 97%, respectively). A probable explanation for this fact could be the
higher reactivity of the pyridine-borane complex under the conditions used.
The stability of oligonucleotide matrices was monitored by measuring the
decrease in remaining radioactivity during 9 h of washing in 0.1 M TEAA buffer
at 60oC and pH 7.0. Regardless of the type of attachment, all the matrices showed
a rapid loss of the bound material (~20%) in the first 2 h. Further decay was 10-100 times slower. The relatively fast initial breakdown of the
matrix seems to be caused by the elimination of short, oligonucleotide-containing polyacrylamide (or polydimethylacrylamide) chains that are not
linked to the entire polymer lattice.
For estimating the relative stability for different matrices, we fitted
experimental points by the two-exponential decay function:
{{B = A} sub 1} cdot e x p ( - {{italic k} sub 1} {{{italic t} ) + A} sub 2} cdot e x p ( - {{italic k} sub 2} {italic t} {roman )}
where B is amount of bound labeled and A
1
, A
2
,
k
1
and
k
2
are constants. The first term of the equation represents the fast initial
decrease of radioactivity (rate constant
k
1
). The rate constant of the slow stage (
k
2
) reflects the overall sum of reactions resulting in loss of label: cleavage of
the oligonucleotide-to-gel bonds, hydrolysis of internucleotide and terminal phosphates,
destruction of polyacrylamide gel, etc. Only the first of all these processes
depends on the type of attachment. The experimental data and fitting curves are
presented in Figure
1
; the calculated parameters A
1
, A
2
,
k
1
and
k
2
are given in Table
2
.
Table 2
Kinetics of matrix degradation
It is important to note here that none of the three borane complexes reduce
N
-(acylamino)dihydroxymorpholine cycle (hydrazide matrix, Scheme
1
), as evidenced by the absence of changes in hydrazide matrix stability after
treatment with reductants.
Thus, from all tested methods, high-yield immobilization of amino-oligonucleotides on aldehyde gel supports providing stable
attachment may be considered as the most effective procedure for
oligonucleotide gel microchip preparation. In addition, it does not require
additional steps for activation of oligonucleotides (oxidation of 3'-ribo unit) and makes possible immobilization on the 3'
or 5' end.
We thank Dr A. Mozoleva, EIMB, for recording NMR spectra, and D. Prudnikov,
EIMB, for labeling of oligonucleotides. This work was supported by Grants 558
and 562 of the Russian Human Genome Program, by Grant DE-FG02-93ER61538 of the US Department of Energy, and by a collaboration
agreement with Affymetrix. We thank L. Novikova and D. E. Nadziejka for
editorial assistance.
The fast initial decrease depends only slightly on the gel preparation mode,
while the rates of the second slow stage differ by a factor of >10. The results
obtained are clearly divisible into two main groups. The first group includes
oligonucleotides attached to the gel through relatively easily hydrolyzable
bonds such as hydrazone or imine. In these cases, the half-life of matrix degradation is ~40 h (Table
2
). The same half-life for matrices treated by trimethylamine-borane complex indicates that this agent does not reduce imino
groups under the conditions used. The matrices in which oligonucleotides are
coupled to the gel by a very stable aminoalkyl bond (half-life of destruction is ~450 h) form the second group. The glutaraldehyde-activated supports, as well as the amine and aldehyde matrices
reduced by pyridine-borane complex, fall into this category. In this case, the imino groups
are reduced completely. In contrast with the pyridine-borane complex, sodium cyanborohydride reduces imines only partly, as
judged from the half-life of matrix degradation (100-130 h).
REFERENCES
Return