WO2012027561A1

WO2012027561A1 - High-stability derivatized glass surfaces for dna sequencing applications

Info

Publication number: WO2012027561A1
Application number: PCT/US2011/049133
Authority: WO
Inventors: Peter Williams; Teresa Gerrity
Original assignee: University of Arizona; Arizona State University ASU
Current assignee: University of Arizona; Arizona State University ASU
Priority date: 2010-08-25
Filing date: 2011-08-25
Publication date: 2012-03-01
Anticipated expiration: 2013-02-25

Abstract

The invention relates to the use of a high temperature treatment of silane-derivatized glass surfaces that may be used to immobilize DNA molecules for sequencing and other applications.

Description

TITLE

HIGH -STABILITY DERIVATIZED GLASS SURFACES FOR DNA SEQUENCING

APPLICATIONS

BACKGROUND OF THE INVENTION

[0001 ] Next-generation (next-gen) DNA sequencing refers to sequencing approaches that have been developed to improve upon and supplant the classical Sanger sequencing process. All are characterized by orders of magnitude improvements in data acquisition, and reductions in reagent amounts and cost. From the estimated cost and time in 2004 of $10,000,000 and ~ 1 -2 years to sequence a second human genome using the industrialized Sanger sequencing technologies developed in the Human Genome Project, next-gen technologies have already shrunk the cost by almost 1 ,000 and the time requirement (for sequence acquisition alone) by a factor close to 100. The dominant commercial technologies (lllumina, Lifetech, 454, Helicos) all use some variant of a sequencing-by-synthesis approach in which surface-attached DNA is exposed to reagents such as deoxyribonucleotide triphosphates (dNTPs) and DNA polymerase enzyme, or short DNA strands and DNA ligase enzyme and extension or ligation is sensed, and a local sequence of one or several DNA bases is typically identified by visualizing a fluorescent probe covalently attached to the dNTP incorporated by the polymerase or the DNA strand attached by the ligase. (454 senses a pyrophosphate molecule released by dNTP incorporation, via a series of enzyme-catalyzed reactions that culminate in a chemiluminescent product.) High data rates and low reagent cost are achieved by carrying out these reactions simultaneously on an array of spots, each containing a set of identical DNA strands with a unique sequence (for Helicos the spots are individual molecules). The arrays contain millions (billions in the Helicos approach) of unique sequences, all of which are read out in parallel giving high throughput and low cost.

[0002] Common to all the next-gen sequencing approaches is a problem of limited read length. The length of contiguous sequence that can be read in a single pass has long been felt to be an important figure of merit in DNA sequencing because the human genome in particular contains numerous sections of repetitive sequence and ideally one would like to be able to read through such a sequence completely and overlap into non-repetitive, unique sequence at either end in order to locate the sequence segment correctly in the overall genome. Read lengths using Sanger technology were improved to ~ 400 bases in the high-throughput sequencing machines used in the final phase of the Human Genome Project, while more specialized systems achieved read lengths ~ 800 - 1 ,000 bases. The current next- gen technologies suffer from significantly shorter read length: 454, ~ 300 bases, lllumina, ~ 120 bases, Lifetech, ~ 50 bases, Helicos, ~ 35 bases. In the case of lllumina, Helicos and probably Lifetech, the read length appears to be limited by steady loss of DNA from the array surface. For Helicos, loss of a single molecule terminates the sequence read for that molecule. For the others, loss of DNA from an array spot reduces the readout signal strength ultimately to a point where data becomes noisy and unreliable. Another phenomenon that limits read length in the non-single molecule techniques is the growth of out-of-phase signals due to molecules that for various reasons failed to be extended at some cycle in the process and subsequently extend at a later cycle producing signals that are erroneous and lead to a growing noise background that eventually makes the sequence readout unusable.

[0003] Although the next-gen companies have shown great ingenuity in working to assemble data of such short read lengths into contiguous lengths of sequence, and although the existence of the gold-standard genome from the Human Genome Project provides a useful scaffold on which to assemble these short sequence snippets, stabilization of the DNA attachment to the surface, and minimization of out-of-phase signal growth, that resulted in extending these read lengths would clearly be of significant value.

BRIEF SUMMARY OF THE INVENTION

[0004] The present invention describes a method to improve the robustness of glass slides with bound DNA, that are used for automated sequencing machines. By using heat treatment along with a linker molecule, the process removes loosely bound DNA that could be dislodged during the sequencing or otherwise adversely affect the accuracy of sequencing, leading to errors that limit the read length of commercial sequencing machines.

[0005] The heat treatment which is carried out immediately after the DNA attachment to the aminosilylated glass slides removes "extractable oligomers" (DNA bound to weakly attached polysiloxane) before the sequencing begins. These siloxane oligomers are reasonably stable under static conditions, but tend to be washed away by repeated rinse steps (leading to signal loss and out-of-phase errors). This removal of the blocking overlayer of extractable siloxane oligomers exposes DNA which is bound to a strongly-attached siloxane layer that tends not to be washed away during sequencing and hence increases the efficiency with which the glass slide can be used in the sequencing reactions.

[0006] It is contemplated that the elevated temperature treatment in an aqueous environment will tend to partially hydrolyze the siloxanes where they bind to the glass surface, making them subject to hydrolysis and removal during later sequencing reaction cycles. Thus, if needed, an additional dry heat treatment can be specified to reform these bonds.

[0007] Thus in specific embodiments, the methods of the invention relate to methods for immobilizing nucleic acid molecules on a surface for enhanced DNA sequencing comprising:

(a) silylating a surface with an aminosilane thereby forming a modified surface that is aminosilanized;

(b) reacting the aminosilanized surface with a heterobifunctional crosslinker, thereby forming a further modified substrate comprising a cross-linker surface;

(c) attaching an amine-terminated nucleic acid strand to the cross- linker to obtain a silylated surface with immobilized nucleic acid;

(d) subjecting the silylated surface containing the immobilized nucleic acid to a high temperature treatment in a buffered aqueous environment to remove from the surface loosely-attached DNA from the surface of the slide; and

(e) optionally, further subjecting the slide to a dry high-temperature treatment to reverse any partial hydrolysis of the siloxane bonds and ensure that all aminopropylsilicons are attached to the surface by a maximum number of siloxane bonds.

[0008] Notably, the heat treatment of step (d) may in fact be performed immediately after the silylation of the surface to form an aminosilanized surface. This would equally well remove the extractable oligomers and then the cross-linker and DNA could be attached to the now-stable surface for immobilization of the DNA. [0009] The surface may be any surface on which DNA mobilization can occur. In exemplary embodiments, the surface is a glass slide or other glass surface. Alternatively, the surface is a siliceous substrate that comprises amorphous silica, fumed amorphous silica, fused silica, or a combination thereof.

[0010] In exemplary embodiments, the high temperature treatment in a buffered aqueous environment to remove from the surface loosely-attached oligomers or DNA attached oligomers from the surface comprises subjecting the surface to a high temperature of about 90 °C - 95° for about 10 to 15 minutes.

[001 1 ] In more particular embodiments, the aminosilane used for derivatizing the surface may be for example, aminopropyl)dimethylethoxysilane (APDMES); (3- aminopropyl)triethoxysilane (APTES); alkyl trichlorosilane, tetramethoxysilane, tetraethoxysilane, tetrapropoxysilane, methyltrimethoxysilane, methyltriethoxysilane, methyltris(methylethylketoxime)silane, methyltris(acetoxime)silane, methyltris(methylisobutylketoxime)silane, dimethyldi(methylethylketoxime)silane, trimethyl(methylethylketoxime)silane, vinyltris(methylethylketoxime)silane, methylvinyldi(methylethylketoxime)silane, methylvinyldi(cyclohexanoneoxime)silane, vinyltris(methylisobutylketoxime)silane, phenyltris(methylethylketoxime)silane, methyltriacetoxysilane, or tetracetoxysilane substituted with an amine group; or a combination thereof. Preferably, the aminosilane comprises APDMES, APTES, or a combination thereof. More preferably, the aminosilane comprises APTES.

[0012] In more particular embodiments, the heterobifunctional crosslinker comprises glutaraldehyde, N-(p-maleimidophenyl)isocyanate (PMPI), m- maleimidobenzoyl-N-hydroxysuccinimide ester (MBS) or N-(. gamma. - maleimidobutyryloxy)sulfosuccinimide ester (sulfo-GMBS).

[0013] In specific embodiments, the method is carried out wherein the aminosilane comprises APTES, wherein the heterobifunctional crosslinker comprises glutaraldehyde, and wherein the amino-terminated nucleic acid molecule comprise amino-terminated DNA.

BRIEF DESCRIPTION OF SEVERAL VIEWS OF THE DRAWINGS

[0014] Fig. 1 illustrates the idealized desireable surface, with a monolayer of multiply-anchored aminosiloxanes to which DNA can be attached. Such a surface should allow free access by bulky enzymes to surface-attached DNA (assuming that the DNA itself is not attached so densely that strands start to impede access to other strands).

[0015] Fig. 2 illustrates the true nature of the APTES surface as deduced from surface spectroscopic studies.

[0016] Fig. 3 illustrates the effects of intermediate primers blocking primer extension and effects of DNA strands buried in pores being exposed by removal of extractable oligomer overlayers.

[0017] Fig. 4 shows the signal loss effect in successive extension reactions on a surface that had NOT been treated at high temperature to remove extractable oligomers. The extension signal has dropped to 50% of the initial signal after only 6- 7 reactions.

[0018] In Fig. 5, the slide was heat-treated at 90 °C for 10 minutes following attachment of the primers, and extractable oligomers and any DNA attached to them were thereby removed, while primers buried in the pores of the siloxane layer were exposed for reaction. It can be seen that the signal loss is very significantly reduced by this process. In this study the reactions were carried out using unlabeled dNTPs and the signals measured were those for fluorescein-labeled primer (with no template attached) and fluorescein-labeled template (attached to unlabeled primer on different spots). Note that the signal loss is very significantly reduced.

[0019] In Fig. 6 the primer signal loss (as a proxy for extension signal losses) is extrapolated out to a 50% signal drop. It is seen that read length is extended out to ~ 90 reaction cycles by this heat treatment and stabilization of the slide surface. Note that although the template signal losses appear larger, such losses can be eliminated by covalently crosslinking the template and primer strands as disclosed in M5-072L.

[0020] Fig. 7 Schematic for the production of high stability derivatized glass surfaces for DNA sequencing operations.

DETAILED DESCRIPTION OF THE INVENTION

[0021 ] The invention describes a method for fabricating silane-derivatized glass surfaces of very high stability for use in DNA sequencing applications that perform chemical manipulations in solution on surface-attached DNA strands. The surface-modifying compound is aminopropyltriethoxysilane (APTES) or related compounds capable of forming one or more siloxane bonds to a glass surface and also carrying a reactive group (examples are amine, sulfhydryl) by which a suitably modified DNA strand can be covalently coupled. Reaction of APTES with a glass surface produces a complex, nanoporous three-dimensional surface network due to self-polymerization of the APTES competing with or adding to attachment to the surface. DNA subsequently attached via the reactive group on the APTES has a hierarchy of binding strengths to the surface and a significant fraction of DNA strands and their siloxane anchors are steadily lost on exposure to a series of reagent and rinse flows during a typical sequencing reaction series. Additionally, the nanoporous surface has varying accessibility for DNA sequencing reagents, in particular the relatively bulky DNA polymerase or ligase enzymes, and the loss of a blocking siloxane may open access to a DNA strand at some time after sequencing has begun, producing an out-of-phase, undesired sequencing signal. The invention describes a high-temperature treatment in water that is used to remove from the surface loosely-attached siloxanes prior to DNA attachment. Subsequently the DNA to be sequenced is attached to a surface that is significantly stabilized to degradation in reagent and rinse flows. Further stabilization of the surface may be achieved by a high-temperature treatment in an inert atmosphere or in air following the high temperature water treatment, to reverse any partial hydrolysis of the siloxane bonds and ensure that all aminopropylsilicons are attached to the surface by a maximum number of siloxane bonds.

[0022] The following is a description of the method used to prepare an APTES-derivatized glass surface for spotting DNA arrays. There are various methods described in the literature, and commercial entities typically have not described their own approaches, but the approach described below should be taken as fairly typical and results in a moderately stable surface, as indicated by the survival of the signal from fluorescently-labeled amino-modified DNA, linked to the amino-silane surface groups by a glutaraldehyde linker, for several weeks under a static buffer solution.

[0023] To begin microarray preparation, borosilicate glass slides (25x75x1 mm, VWR) were placed in a Piranha solution (1 :3, H₂0₂:H₂S0 ) for thirty minutes. The slides were then rinsed with de-ionized water and placed in 10% NH₄OH solution for thirty minutes. A 3-aminopropyltriethoxysilane (APTES) solution (10 mM in acetate buffer, pH 5.2) was freshly prepared, and the slides were rinsed again with de-ionized water before being placed into the APTES solution at 95 °C for three hours. After this time, the slides were rinsed with de-ionized water, blown dry with N₂ and placed in a desiccator at 120 °C for at least thirty minutes, to allow curing of the surface (further polymerization enhanced by dehydration of the surface). The cooled slides were then placed in a freshly made solution of 5% glutaraldehyde (in PBS, pH 8.0) for two hours at room temperature. After this time, the slides were rinsed with de-ionized water, blown dry with N₂ and were then ready for spotting with the amino-modified primers. In our approach we first link bifunctional glutaraldehyde molecules to the surface. The aldehyde group at one end of the glutaraldehyde links to the free amine group to form a Schiff base. The same reaction can subsequently be used to attach an amine-terminated DNA strand to the free aldehyde group at the other end of the glutaraldehyde. For the sequencing experiments described below, the surface-attached DNA was attached via an amine group at its 5' terminus and had a free 3' terminus at which DNA polymerase-catalyzed extension could occur. For extension, lengths of template DNA with complementary sequence to the primer DNA were introduced in an appropriate buffer and allowed to hybridize to the surface attached primers.

[0024] The experimental observations described below can be understood by reference to sketches illustrating the nanoscopic nature of the APTES-derivatized surface (adapted from F.R. Jones).

[0025] Jones defines a layer of "extractable oligomers", consisting of single and partially polymerized APTES molecules held to the surface by non-covalent interactions between the polar -OH and -NH₂ or NH₃ ⁺ groups. These extractable oligomers are reasonably stable under static solution conditions, and so amino-DNA linked to them will also be reasonably stable; however, repeated rinsing tends to wash away an increasing fraction of these oligomers and attached DNA. Signal losses due to detachment of extractable oligomers carrying DNA is one process that limits read length.

[0026] A second problem with the nanoporous polysiloxane surface shown in Fig. 2 is that DNA molecules can frequently penetrate into the pores and be attached in subsurface sites that subsequently are inaccessible to bulky polymerase or ligase enzymes. As the extractable oligomers are slowly washed away, these subsurface DNA molceules become enzyme-accessible and start to participate in the extension or ligation reactions for sequencing, but obviously do so out of phase with the molecules that were exposed from the start and thus produce erroneous signals. The growth of erroneous, out-of-phase sequencing signals is another process that limits read lengths.

Demonstrations

[0027] The phenomenon of DNA strands buried in pores being exposed by removal of extractable oligomer overlayers is illustrated in Fig. 3.

[0028] In this experiment a surface that had not been treated to remove the extractable oligomers was subjected to a DNA polymerase-catalyzed extension reaction with a single species of Cy5-labeled dNTP (the sequencing approach used here is described in U.S. Patent No. 7645596 and U.S. Patent No. 6780591 , and is similar for the purposes of the present disclosure to the approaches of lllumina and Helicos). For spots with two different template sequences the typical extension signals are shown as the white bars. The differences in signal are due to effects of the local DNA sequence on the response of the Cy5 fluorescent signal. The Cy5 fluorescent label was then inactivated by treatment with sodium borohydride (see disclosure M7-078L) and a second, identical extension reaction was carried out. The tiny dark color bars show the signal due to reaction of previously unreacted DNA in this second reaction. Around 5% of the initial signal is seen, and in view of the step to be described next, this is interpreted as extension of unreacted DNA strands exposed by rinsing away of some fraction of extractable oligomers in the steps that followed the first extension reaction (these consist of rinsing away the dNTP solution, flushing in fresh buffer for the fluorescent imaging step, rinsing away the imaging buffer and flowing in fresh dNTP solution). The slide was then heated in buffer to 90 °C for 10 minutes, rehybridized with fresh template and subjected to a further extension reaction, again with the same dNTP. The hatched bars show that an amount of DNA amounting to ~ 40 - 60% of the initially exposed template was now accessible for reaction as a result of the removal of the blocking overlayer of extractable oligomers. This illustrates a process of exposure of unreacted material that can create out-of-phase signals.

Further stabilization

[0029] It may be noted in Fig 6 that both template and primer signals show an initial drop before stabilizing to a slower loss rate. (It is presumed that the initial rise in the primer signal is due to losses of primer that reduce the self-quenching effect of fluorophores that are too densely packed on the surface.) It appears probable that the heat treatment that removes the extractable oligomers may also cause partial hydrolysis of siloxanes that were initially multiply linked to the surface and therefore stable enough to survive the 10 minute high temperature treatment, but at the end of the treatment are linked only by a single siloxane bond that is fairly easily broken - i.e. while the heat treatment removes most extractable oligomers, it may also create a small population of new extractable oligomers. It may be possible to re-attach those broken bonds by a high temperature treatment in a dry atmosphere, reversing the hydrolysis, thus further stabilizing the surface.

[0030] This invention describes techniques and processes to produce glass surfaces with covalently attached DNA, for sequencing by synthesis and sequencing by ligation, that have increased resistance to signal loss by DNA strand detachment and increased immunity to production of erroneous or out-of-phase signals.

[0031 ] This invention uses an understanding of the nanometer-scale structure of siloxane-derivatized gass surfaces to prescribe methods to produce surfaces that have high stability for increased read length in DNA sequencing applications and also result in greater accuracy in the sequencing process.

[0032] It should be understood that it is not possible to know in any detail what surface preparation methods the existing next-generation DNA sequencing companies currently use. It is presumed that they will have taken measures to prepare surfaces that exhibit a high level of stability. Nevertheless it is highly suggestive that the read lengths reported by the various companies seem to scale rather well with the number of process steps involving reagent or rinse flows. Thus Helicos reports read lengths ~ 35 bases. They use a 1 -color dNTP approach, which means they need 4 complete cycles of dNTP addition, rinses, imaging etc in order to complete 1 round of extension with all 4 bases. (Note that the DNA strands that are not extended are still subject to the reagent and rinse flows, and associated losses.) On average one four-reaction cycle will yield ~2 bases of sequence per strand, i.e. to extend by 35 bases requires a total of ~ 60 reactions. If the Helicos surface is less stable than that of Fig 6 their read length might be improvable by ~ 50%, which would make them comparable with Lifetech. Illumina uses a 4-color approach, which means that they can extend with all 4 dNTPs in a single cycle. A four-reaction cycle yields 4 bases of sequence. If lllumina's 120 base read length (requiring 120 reactions) corresponds to 50% signal loss (there is no standard definition in the field) then lllumina's surface stability is better than that described here. The impact of out- of-phase signals in the Illumina approach is not known.

Claims

1 . A method for immobilizing nucleic acid molecules on a surface for enhanced DNA sequencing comprising:

(c) attaching an amine-terminated nucleic acid strand to the cross-linker to obtain a silylated surface with immobilized nucleic acid;

2. A method for immobilizing nucleic acid molecules on a surface for enhanced DNA sequencing comprising:

(b) subjecting the silylated surface to a high temperature treatment in a buffered aqueous environment to remove from the surface loosely-attached aminosilated oligomers from the surface of the slide;

(c) optionally, further subjecting the slide to a dry high-temperature treatment to reverse any partial hydrolysis of the siloxane bonds and ensure that all aminopropylsilicons are attached to the surface by a maximum number of siloxane bonds

(d) reacting the aminosilanized surface with a heterobifunctional crosslinker, thereby forming a further modified substrate comprising a cross-linker surface; and

(e) attaching an amine-terminated nucleic acid strand to the cross-linker to obtain a silylated surface with immobilized nucleic acid.

3. The method of claim 1 or claim 2, wherein the surface is a glass surface.

4. The method of claim 1 or claim 2 wherein the surface is a siliceous substrate that comprises amorphous silica, fumed amorphous silica, fused silica, or a combination thereof.

5. The method of claim 1 , wherein step (d) comprises subjecting the surface to a high temperature of about 90 °C - 95° for about 10 to 15 minutes.

6. The method of claim 2, wherein step (b) comprises subjecting the surface to a high temperature of about 90 °C - 95° for about 10 to 15 minutes.

7. The method of claim 1 or claim 2, wherein the aminosilane comprises aminopropyl)dimethylethoxysilane (APDMES); (3-aminopropyl)triethoxysilane (APTES); alkyl trichlorosilane, tetramethoxysilane, tetraethoxysilane, tetrapropoxysilane, methyltrimethoxysilane, methyltriethoxysilane, methyltris(methylethylketoxime)silane, methyltris(acetoxime)silane, methyltris(methylisobutylketoxime)silane, dimethyldi(methylethylketoxime)silane, trimethyl(methylethylketoxime)silane, vinyltris(methylethylketoxime)silane, methylvinyldi(methylethylketoxime)silane, methylvinyldi(cyclohexanoneoxime)silane, vinyltris(methylisobutylketoxime)silane, phenyltris(methylethylketoxime)silane, methyltriacetoxysilane, or tetracetoxysilane substituted with an amine group; or a combination thereof.

8. The method of claim 4, wherein the aminosilane comprises APDMES, APTES, or a combination thereof.

9. The method of claim 8, wherein the aminosilane comprises APTES.

10. The method of claim 1 or claim 2, wherein the heterobifunctional crosslinker comprises glutaraldehyde, N-(p-maleimidophenyl)isocyanate (PMPI), m- maleimidobenzoyl-N-hydroxysuccinimide ester (MBS) or N-(. gamma. - maleimidobutyryloxy)sulfosuccinimide ester (sulfo-GMBS).

1 1 . The method of claim 1 or claim 2, wherein the aminosilane comprises APTES, wherein the heterobifunctional crosslinker comprises glutaraldehyde, and wherein the amino-terminated nucleic acid molecule comprise amino-terminated DNA.