abstracts from the
Mugasimangalam Raja, Vincent Molloy, Lev
Lvovsky, James Akowski, Jonathan Schisler,
Yakov Kogan#, Michael Fonstein#, Robert
Haselkorn# and Levy Ulanovsky
We are using DENS (Differential Extension with Nucleotide Subsets, see Ref.1 and the accompanying abstract by Raja et al.) for sequencing the genome of Rhodobacter capsulatus by primer walking without custom primer synthesis. A cosmid library of R. capsulatus was constructed and mapped with high resolution (2). Of the total 3.5 Mb genome, 1.4 Mb has already been sequenced by limited shotgun sequencing followed by conventional primer walking. Forty-eight plasmid subclones containing 5 kb inserts were isolated from each cosmid and both ends sequenced. The resulting sequences were then assembled into contigs. In conventional primer walking, approximately 100 custom synthesized primers per cosmid are required to close the sequencing gaps and generate the second strand. Around 10 additional primers are needed to fill the physical gaps between subclones by direct primer walking on the cosmid.
In primer walking by DENS, the conventional custom synthesized walking primers are replaced with DENS octamers (containing two degenerate positions each) from our presynthesized library of 2,048 octamers (50% of all possible sequences). At the current DENS success rate of 62% on dsDNA templates, the use of DENS is cheaper than primer walking using conventional custom-made primers, even though the latter yields an 85% success rate. Future closed end automation of DENS primer walking, made possible by the instant availability of primers, should reduce sequencing cost by more than an order of magnitude. The results of the R. capsulatus sequencing show that DENS is a viable option. Updated results on this pilot project and implications of the DENS technique will be discussed.
1. Raja et al. (1997) Nucleic Acids Res. 25: 800-805.
2. Fonstein et al. (1995) EMBO J. 14: 1827-1841.
Mugasimangalam Raja*, Vincent Molloy*, Lev
Lvovsky# and Levy Ulanovsky*#
DENS (Differential Extension with Nucleotide Subsets) is a technique used for template directed enzymatic synthesis of unique primers, avoiding the chemical synthesis step in primer walking (1). DENS works by selectively extending a short primer and making it a long one at the intended site only. The procedure starts with an initial extension of the primer (at 20-30C) in the presence of only two out of the four possible dNTPs. The primer is extended by 5 bases or longer at the intended priming site, which is deliberately selected, (as is the 2 dNTP set), to maximize the extension length. The subsequent termination reaction at 60-65C then accepts the primer extended at the intended site but not at alternative sites, where the initial extension (if any) is generally much shorter.
Until recently we were getting about 70% success rate on ssDNA plasmid and asymmetric PCR products and 62% on ds plasmids. Now, we have found that the usable template size for DENS is not limited to several kb (typically plasmids), as we thought previously. It turns out that DENS sequencing can be performed on cosmid-sized templates on the condition that the sequence of most of the template is known (as in filling physical gaps). Then the computer excludes candidate priming sites for whom the chosen octamer primers have alternative priming sites in the known part of the template if at any alternative site the extension crosses the failure threshold (about 4-5 bases). Of 17 DENS sequencing reactions performed using Lambda phage as the template, 10 worked well.
In addition, we have developed a novel version of DENS that uses only one out of the 4 possible dNTPs in the differential extension (either dCTP or dGTP extending the octamer by 3 or more bases). One-dNTP DENS allows primer walking on such a template as a whole cosmid, even if most of its sequence is completely unknown. What makes it possible is that the probability of the occurrence of alternative sites with long differential extensions using a single dNTP is dramatically lower than that of extensions using a two-dNTP subset. Of five octamer primers tested on Lambda using one-dNTP DENS, four produced good sequences at the intended unique sites. The occurrence of suitable priming sites for the one-dNTP DENS is lower than for the two-dNTP DENS but is still high enough to perform primer walking if all 4,096 octamers are used.
1. Raja et al (1997) Nucleic Acids Res. 25, 800-805.
Dina Zevin-Sonkin#, Anahit Ghochikyan#, Arthur
Liberzon#, Lev Lvovsky# and Levy Ulanovsky*#
DENS allows primer walking without custom primer synthesis and each walk involves a two step procedure. It starts with a limited initial extension of an 8-mer primer (degenerate in 2 positions) at 20-30C in the presence of only 2 out of the 4 possible dNTPs. The primer is extended by 5 bases or longer at the intended priming site, which is deliberately selected, (as is the two-dNTP set), to maximize the extension length. The subsequent termination reaction at 60C then accepts the primer extended at the intended site, but not at alternative sites, where the initial extension (if any) is generally much shorter (see Ref.1). Both steps involve thermocycling.
Here we show an example of DENS primer walking on three human genomic DNA subclones of the 3-4 kbp length each from telomeric region of human chromosome 7, kindly provided by Robert Moysis group, LANL, Los Alamos, NM. The full length sequences (3024, 3473 and 3707 bp) were obtained from both strands of each subclone using dye- terminators and the ABI-373 sequencer. For these clones we tested an easy method for ss template preparation avoiding plasmid prep of any kind. The entire insert of a plasmid clone was amplified by PCR in which one of the two primers was 5'-phosphorylated. One of the two strands was digested with lambda exonuclease (Boehringher Mannheim, cat#1666908) which selectively digests 5'-phosphorylated strand only (Ref.2). Using this method ssDNA template is produced by PCR using vector-specific primers directly from overnight bacterial culture.
For the DENS reaction, we have optimized parameters related to the stability of the extended primers and used them for the selection of DENS priming sites. Upon this optimization, the success rate of DENS primer walking using the ss template preparation by PCR was 70%.
1. Raja et al. (1997) Nucleic Acids Res. 25, 800-805.
2. Little et al. (1967) J. Biol. Chem. 242, 672-678.
Dina Zevin-Sonkin#, Anahit Ghochikyan#, Arthur
Liberzon#, Lev Lvovsky#, and Levy Ulanovsky*#
Primer walking using conventional primers is believed to be problematic when tandem Alu repeats are present in the template. In contrast to conventional 18-20mer primers, DENS (Differential Extension with Nucleotide Subsets, see Ref. 1 and accompanying abstracts by Raja et al.) uses octamers degenerate in 2 positions. We have found that DENS has an advantage over conventional primer walking in sequencing through tandem Alu repeats. A single mismatch in the octamer/template complex prevents priming, enabling discrimination between nearly identical repeats. It is possible to walk through repeat-rich regions by selecting hypervariable sites for DENS priming within the Alu consensus sequence. Additional mismatch discrimination is provided during the differential extension stage, since the extension will be shorter if a template base is non-complementary to either of the two dNTPs provided (e.g. if dATP and dGTP were the only nucleotides present, and a "G" were encountered in the template). Finally, the extended primer is also destabilized by a single mismatch at the higher annealing temperature of the cycle sequencing stage. All of these factors cause DENS to be very effective at distinguishing similar, but not identical, priming sites within a DNA template, whereas a conventional long primer is less able to provide such discrimination.
1. Raja et al. (1997) Nucleic Acids Res. 25, 800-805.
Stanley Tabor, Sylvie Doublié, Tom Ellenberger and Charles Richardson
Current methods for DNA sequencing require a DNA polymerase to extend a primer with each of the four natural nucleotides, as well as a variety of analogs such as fluorescent and chain-terminating dideoxynucleotides. We have been characterizing the structure and function of DNA polymerases to better understand and to modify those properties important for DNA sequencing, in particular the incorporation of nucleotide analogs and the processivity of DNA synthesis. Our work has focused on the DNA polymerases belonging to the Pol I family, that includes T7 DNA polymerase, Taq DNA polymerase, and E. coli DNA polymerase I. An important advance in our understanding of these enzymes has resulted from our recent determination of the 2.2 crystal structure of T7 DNA polymerase locked in a replicating complex with a dideoxy-terminated primer-template, an incoming dNTP, and the processivity factor thioredoxin . The incoming dNTP fits snugly into a pocket formed by the fingers of the polymerase closing on its palm and the 3'-terminus of the primer. Numerous interactions of the bound nucleotide with the template base, the polymerase, and two metal ions specify the correct base-pair in the active site, and provide insight into the mechanism of discrimination against analogs with modifications in the sugar moiety (e.g. dideoxynucleotides, ribonucleotides and 3' fluoro derivatives) as well as bases containing bulky fluorescent substituents. We are making use of this structure to construct mutant polymerases that incorporate various nucleotide analogs more efficiently that are modified in either their sugar, base or triphosphate groups.
The processivity factor for T7 DNA polymerase, E. coli thioredoxin, is located at the tip of the thumb of the polymerase in a position poised to prevent dissociation of the DNA from the polymerase. It binds to a 74 residue domain in T7 DNA polymerase that is attached to the thumb by a flexible tether. While this domain is unique to T7 DNA polymerase, it is a modular in that it can be transferred to other homologous polymerases by gene fusion to generate hybrid enzymes that have dramatically increased processivity for DNA synthesis . The crystal structure of the T7 DNA polymerase complex suggests that thioredoxin is acting to stabilize the region it binds to, allowing a number of critical basic residues in T7 DNA polymerase to interact electrostatically with the DNA backbone and thus prevent its dissociation. We are carrying out extensive mutagenesis of this region in order to further define the critical structural features and to engineer new DNA polymerases that have increased processivity.
This work is funded in part by DOE grant DE-FG02-96ER62251 (Stanley Tabor, P. I.)
 Crystal Structure of Bacteriophage T7 DNA Polymerase Complexed to a Primer-Template, a Nucleoside Triphosphate, and its Processivity Factor Thioredoxin. Sylvie Doublié, Stanley Tabor, Alexander Long, Charles C. Richardson and Tom Ellenberger, Nature, in press.
 The Thioredoxin Binding Domain of Bacteriophage T7 DNA Polymerase Confers Processivity on Escherichia coli DNA Polymerase I. Ella Bedford, Stanley Tabor and Charles C. Richardson. Proc. Natl. Acad. Sci. USA 94, 479-484 (1997).
Mark W. Knuth, Scott A. Lesley, and Heath E. Klock
The objective of this grant is to develop an engineered RNA/DNA polymerase for primerless DNA sequencing, among other applications. In order to develop this enzyme, systems have been developed for facile semiautomated mutagenesis, expression, micropurification, and screening of mutants of T7 RNA polymerase.
Two publications have reported T7 RNAP mutations that confer some ability to incorporate dNTPs; we have further characterized their performance in parameters important to DNA sequencing and other applications, and have created screens to test for improved performance.
To date, clones representing all possible amino acid substitutions at 96 sites have been created, using a novel mutagenesis kit developed at Promega. These mutant enzymes are being expressed, purified and tested for increased dNTP incorporation. We anticipate that at full capacity, around 12 positions will be screened per week, using techniques and equipment that can be generalized to other enzymes requiring some purification before screening.
John J. Dunn and Matthew Randesi
An ordered set of nested deletions whose ends are separated by ~400 bp allows rapid sequencing across one strand of a cloned fragment, using a universal primer. Any gaps remaining after this process can be closed by primer walking on the original clone. Even highly repeated DNA can easily be assembled correctly, knowing the relative locations of the sequences obtained. We have developed vectors and protocols that allow simple and reliable production of nested deletions suitable for such a sequencing strategy, from cloned fragments at least as large as 17 kbp and potentially 40 kbp.
We have made two single-copy, amplifiable vectors suitable for this strategy, and have validated and made reliable a previously described method for generating nested deletions enzymatically. In both vectors, clones are stably maintained at low copy number by the F replication and partitioning functions and can be amplified from an IPTG-inducible P1 lytic replicon to prepare DNA. A synthetic version of the phage f1 origin of replication is located a short distance upstream of the multiple cloning site. Vector pND-1 is used primarily for obtaining clones by transformation or electroporation; pND-2 has phage lambda cos sites that allow efficient cloning of 30-40 kbp fragments in a lambda packaging system.
Reaction conditions have been defined where purified f1 gene 2 protein efficiently introduces a strand-specific single nick in the f1 origin sequence with very little rejoining. Large amounts of stable gene 2 protein are easily obtained from a clone by a rapid purification procedure we developed. To create the nested deletions, the nick is expanded unidirectionally into the cloned fragment by 3' to 5' digestion with E. coli Exo III, the resulting single-stranded regions are digested with S1 nuclease, and the ends are repaired and ligated with T4 DNA polymerase and ligase. The Exo III digestion is highly synchronous and processive, and the deletion lengths are proportional to incubation time. To prevent undeleted DNA from giving rise to clones, the treated DNA is digested with one of several restriction enzymes whose 8-base recognition sequences lie between the f1 origin and the cloning site. Nested deletion clones are then obtained by electroporation.
Pooling samples from several different times of Exo III digestion before subsequent treatment generates a good distribution of deletion clones. Growth and amplification of randomly selected clones in 1 ml of medium in 96-well format followed by a simple DNA preparation protocol provides ample DNA for analyzing deletion length by gel electrophoresis and for DNA sequencing reactions. Imaging and sizing software is now being tested for automated selection of an appropriate set of deletions for sequencing.
Exploratory work has demonstrated the effectiveness of this nested deletion strategy for sequencing fragments at least as large as 17 kbp cloned from a human BAC.
Charles L. Asbury, Ger van den Engh
We present a method, analogous to optical trapping, which allows manipulation of DNA molecules in aqueous solution. Due to the induction of an electric dipole, DNA molecules are pulled by a gradient force to regions of high electric field strength. With the use of very thin gold films and oscillating fields, molecules can be trapped individually or locally concentrated. The molecules do not become permanently attached to the gold. Spatial control over the trapped molecules is achieved because they are confined to a width of ~5 mm perpendicular to the edges of the gold films. It is possible to mix static and oscillating electric fields in order to move trapped molecules from one edge to another, or to make them follow very precise trajectories along the edges. This phenomenon may be useful in microdevices for manipulation of small quantities or single molecules of DNA.
Su-Chun Hung*, Yiwen Wang**, Richard A.
Mathies** and Alexander N. Glazer
Energy transfer (ET) primers are markedly superior to single dye-labeled primers in DNA sequencing, and in multicomponent genetic analyses, such as forensic identification, and genetic typing of short tandem repeats.1,2 We describe here improvements in the properties of ET fluorescent primers that enhance their value in all of these applications. In designing improved ET primers, we focused on the spectroscopic characteristics most important for DNA sequencing and PCR product analysis: the relative acceptor fluorescence emission intensity and the amount of residual donor fluorescence emission, and on improving the match in electrophoretic mobilities of the DNA fragments extended from ET primers. Hung et al.3 showed that 3-(e-carboxypentyl)-3'-ethyl- 5,5'-dimethyloxa-carbocyanine (CYA or C), a dye with a high absorption cross-section but a low fluorescence quantum yield, is superior to FAM (F) as a donor in ET primers and that 6-carboxyrhodamine-6G (R6G or G) and 5&6-carboxyrhodamine-11) (R110) can serve as alternative acceptors to JOE and FAM, respectively, in ET primers. CYA-labeled ET primers have stronger acceptor emission and higher spectral purity than ET primers with a FAM donor. The ET primer set C10R110, C10G, C10T, and C10R with only rhodamine derivatives (R110, R6G, TAMRA, and ROX) as the four acceptors has superior spectroscopic properties and gives results in DNA sequencing with near-perfect match in the electrophoretic mobility of single-base extension DNA fragments both in capillary and slab gel electrophoresis.3-5 ET primers with CYA as donor and a donor-acceptor spacing of 4-6 nucleotides offer excellent acceptor emission intensities coupled with negligible donor emissions. In multiplex separations, this allows precise quantitation of the ratio of the signals from DNA fragments labeled with one or another of two different primers. We have applied the two-color ET primer sets C4G/C4R and C6G/C6R to bladder cancer diagnosis based on electrophoretic analyses of polymerase chain reaction-amplified short tandem repeats (STRs) where the diagnosis depends on the detection of loss of heterozygosity at particular loci. The success of the analysis depends on the accurate multiplex quantitation of the amplified DNA fragments from two different samples, normal and tumor cell DNA. Multiplex analyses with the two-color primer sets C4G/C4R and C6G/C6R allowed quantitative determination of allelic ratios with a precision of 10%.6 We performed a quantitative comparison of sets of primers differing in the nature of the donor-acceptor combinations, CYA-ROX, FAM-ROX, and BODIPY503/512- BODIPY581/591. Variables examined included the length of the 5'-amino linker arm, the number of base pairs between the donor and acceptor, and the excitation wavelength (488 or 514 nm). Of the primers examined, CYA-ROX primers offer the best combination of acceptor fluorescence emission intensity and spectral purity.7
Supported by the Director, Office of Energy Research, Office of Health and Environmental Research of the U.S. Department of Energy under Contract DE-FG-91ER61125. Financial support from the Amersham Life Science Inc. is also gratefully acknowledged.
1. J. Ju, A.N. Glazer, and R.A. Mathies Nature Medicine 2, 246-249 (1996).
2. A.N. Glazer and R.A. Mathies Curr. Opinion Biotechnol. 8, 94-102 (1997).
3. S-C. Hung, J. Ju, R.A. Mathies, and A.N. Glazer Anal. Biochem. 243, 15-27 (1997).
4. S-C. Hung, J. Ju, R.A. Mathies, and A.N. Glazer Anal. Biochem. 238, 165-170 (1996)
5. S-C. Hung, R.A. Mathies, and A.N. Glazer Anal. Biochem. 251, (1997) - in press.
6. Y. Wang, S-C. Hung, J.F. Linn, G. Steiner, A.N. Glazer, D. Sidransky, and R.A. Mathies Electrophoresis 18, (1997) - in press.
7. S-C. Hung, R.A. Mathies, and A.N. Glazer Anal. Biochem. - in press.
Kenneth W. Porter, Dima Sergueev, Ahmad
Hasan, J. David Briley and Barbara Ramsay
DNA can be simultaneously amplified and sequenced using a new class of nucleotides containing boron. During the polymerase chain reaction, boron-modified nucleotides, i.e. 2'-deoxynucleoside 5'-alpha-[P-borano]- triphosphates, are incor-porated into the product DNA. The boranophosphate linkages are resistant to nucleases and thus the positions of the boranophosphates can be revealed by exonuclease digestion, thereby generating a set of fragments that terminate in a boranophosphate linkage and define the DNA sequence. The boranophosphate method offers an alternative to current PCR sequencing methods. Single-sided primer extension with dideoxynucleotide chain terminators is avoided, with the consequence that the sequencing fragments are derived directly from the original PCR products. Boranophosphate sequencing has been demonstrated with the Pharmacia and the Applied Biosystems 373A automatic sequencers, producing data that is comparable to cycle sequencing.
The method has been improved recently by streamlining sample preparation and by employing modified boranophosphate nucleotides. Sample preparation is streamlined by implementing direct digestion of the PCR products immediately after amplification. The base-specific PCR products are combined and digested by direct addition of Exonuclease III and Phosphodiesterase I. Subsequently, the digestion products are purified by spin column chromatography, concentrated by isopropanol precipitation, and separated by gel electrophoresis. The uniformity of the digestion products is increased by the addition of modified boranophosphates to the PCR amplification and by the addition of Phosphodiesterase I to the digestion mixture.
Additionally, we have synthesized a 2'-deoxyadenosine alpha-borano-triphosphate with a fluorescent tag attached at the C-8 position of adenine via an alkylamine. Experiments are in progress to determine its ability to be incorporated during amplification and its nuclease resistance, and to develop a screen that will allow for rapid evaluation of other naturally occurring or mutant exonucleases.
The boron method may find use in applications where high resolution of longer fragments requires stronger signals at longer read lengths, because the distribution of fragments produced by nuclease digestion should be skewed to long fragments. Direct sequencing of PCR products simplifies bidirectional sequencing and provides a simple, direct, and complementary method to cycle sequencing.
(Supported by DOE grant 94-ER61882 and continuing grant 97ER-62376 to B.R.S.)
Indu Kheterpal, James R. Scherer, William W. Ja,
Yiwen Wang and Richard A. Mathies
Recent advancements in DNA analysis with capillary array electrophoresis (CAE) have included (i) the optimization of sample preparation, sample loading and separation matrices for DNA sequencing, (ii) the development of a scanner for detecting separations in large numbers of capillaries (~1000) in parallel, (iii) the development and evaluation of new 3-color extended binary coding strategies, and (iv) the development of rapid and sensitive methods for high-throughput cancer screening:
(1) Four-color confocal fluorescence CAE scanners using our standard flat bed design1 are being used along with energy transfer (ET) primers for sequencing of mitochondrial (mt) DNA, Chlamydia, and Anabaena. Twelve motifs of the hypervariable region I of human mt DNA from a Sierra Leone population have been sequenced with an average accuracy of 99.7%.2 Over 40 kilobases of Chlamydia clones have been resequenced and a 8 kilobase fragment has been assembled with 99.63 % accuracy. Several genes from Anabaena have also been sequenced as a part of an undergraduate research project and fragments of up to 8 kilobases have been assembled.
(2) A capillary array scanner (CAE) capable of acquiring four-color data at the rate of 4 Hz from over 1000 capillaries has been constructed. The scanner has a rotating objective which excites and collects fluorescence from one to 1088 capillaries which are positioned in precisely machined grooves (spaced at 260 microns) in a cylindrical objective housing. Four-color data from four photomultipliers is obtained simultaneously with four independent ADCs. The acquired data is stripped of non-sample gaps, averaged across each capillary and also averaged across a variable number of successive rotations as it is acquired. The instrument has been designed to use replaceable matrices introduced by pressure filling. This scanner will be evaluated through sequencing of Chlamydia in collaboration with Ron Davis' group at Stanford.
(3) New three-color extended binary coding strategies for multiplex DNA sequencing have been developed using ET primers.3 These three-color methods are found to be nearly as good as traditional four-color coding methods with sequencing accuracy rates of 99.6% and 99.9%, respectively. Methods have been developed to deconvolve the three-color data into the four base concentrations. This three-color approach is the first step towards the development of higher order multiplex coding schemes for DNA sequencing and other analyses.
(4) Finally, short tandem repeat (STR)-based bladder cancer diagnosis methods have been developed using two-color labeling with ET primers and CAE electrophoresis.4 Rapid (< 35 min.) separations are achieved on capillary arrays using replaceable separation matrices and the allelic ratios are quantitatively determined with a precision of 10%. These methods provide a significant improvement in the speed, ease and precision of STR analyses.
Supported by the Director, Office of Energy Research, Office of Health and Environmental Research of the U. S. Department of Energy under Contract DE-FG-91ER61125.
1. R. A. Mathies, X. C. Huang, Nature (London), 359, 167-169, 1992.
2. I. Kheterpal, J. R. Scherer, S. M. Clark, A. Radhakrishnan, J. Ju, C. L. Ginther, G. F. Sensabaugh, R. A. Mathies, Electrophoresis, 17, 1852-1859, 1996.
3. I. Kheterpal, L. Li, T. P. Speed, R. A. Mathies, Anal. Chem., submitted
4. Y. Wang, S. -H. Hung, J. F. Linn, G. Steiner, A. N. Glazer, D. Sidransky and R. A. Mathies, Electrophoresis 18, in press, 1997.
Michael S. Westphall, David R. Rank and Lloyd M. Smith
It has been recognized from the beginning of the Human Genome project that in order to successfully complete this tremendous undertaking the development of new and improved technologies for DNA sequencing would be required. We have focused on several aspects of this technology development from automated Front End sample preparation to base calling of collected data. Central to the development of this and any other automated sequencing system is the electrophoresis platform. To meet the electrophoresis throughput requirements of our automated Front End DNA sample preparation system, we have developed a 4 color fluorescent sequencing instrument emphasizing long read lengths, dense sample loading, and low construction cost.
The system employs a 250 micron thick cross-linked polyacrylamide gel, 10 inches wide x 24 inches long which is temperature regulated on both sides. A scanning four color detection system is employed to collect the emitted fluorescence. Data is collected bi-directionally with one of four bandpass filters in position per scan. Image processing and base calling is performed using GelImager and BaseFinder software packages developed in our lab.
The system provides sufficient resolution to base-call DNA fragments beyond 1000 bases in length (albeit requiring high quality template preparations and sequencing chemistry) and to routinely deliver runs with 750 bases of useable sequence (98% accuracy). The system currently process 88 samples in parallel and can be built at a cost of $23,000. Instrument design details will be presented along with performance characteristics (in combination with our analysis software) as obtained through in-house sequencing projects.
Edward S. Yeung* and Hongdong Tan
DNA sequencing as practiced today involves a series of steps starting from isolating DNA from the biological sample, cutting these into fragments with convenient sizes, amplifying the fragments biochemically, introducing a label for detection while generating a nested set of ordered fragments, separating the ordered set of fragments and identifying the nucleotide sequence, and reassembling the short sequence data into a continuous sequence. Recent developments of capillary electrophoresis, especially in multiplexed arrays, show great promise for substantially increasing the speed and throughput of the separation and identification steps. The issue of cost when combined with the other steps in the whole sequencing process still remains. Our project is aimed towards the development of novel front-end strategies, whereby the speed and throughput of sample preparation can be significantly increased while the amount of manual operation and total cost can be significantly reduced. We will present our latest results on multiplexed sample preparation in small volumes without the use of robotics. This promises to reduce the cost of reagents and at the same time provide high-speed high- throughput operation.
An integrated on-line prototype for coupling a microreactor to capillary electrophoresis for DNA sequencing has been demonstrated. A dye-labeled terminator cycle-sequencing reaction is performed in a fused-silica capillary. Subsequently, the sequencing ladder is directly injected into a size-exclusion chromatographic column operated at ~95C for purification. On-line injection to a capillary for electrophoresis is accomplished at a junction set at ~70C. High temperature at the purification column and injection junction prevents the renaturation of DNA fragments during on-line transfer without affecting the separation. The high solubility of DNA in and the relatively low ionic strength of 1x TE buffer permit both effective purification and electrokinetic injection of the DNA sample. The system is compatible with highly efficient separations by a replaceable poly(ethylene oxide) polymer solution in uncoated capillary tubes.
We will present data from an 8-capillary system, where individual templates are simultaneously injected from standard microtiter wells. After injection, a completely computer controlled system takes the samples through terminator-labeled cycle sequencing, purification, introduction to 8 parallel electrophoresis capillaries for separation and detection, and base calling. Scaling up to 100 simultaneous channels is straightforward. Future research should allow starting from single bacterial colonies injected into each capillary and the nucleotide sequence identified in a fully on-line and automated system, perhaps multiplexed to 1000 at a time.
Jian Jin, William F. Kolbe, Yunian Lou and Earl W. Cornell
As the Human Genome Project moves towards a scale-up of its sequencing phase, it is apparent that gel electrophoresis will remain the main technology for DNA sequencing. Although slab gel electrophoresis is currently the dominant technology used in production-level sequencing, recent rapid progress in capillary electrophoresis technology suggests that it soon will have the capability of replacing slab gel sequencer in production work. At LBNL we have developed a beta-test version of a 96-capillary system capable of production level sequencing at increased rates and more importantly, improved automation. The system is based on an adaptation of the best available technology currently being developed by several laboratories. In particular, we employ a sheath-flow excitation/detection geometry derived from earlier work by Norman Dovichi at the University of Alberta. Our system, fully integrated from the assembling of the capillary array to automated DNA sequencing, to base-calling, has demonstrated the effectiveness of the sheath-flow approach using 96-channel array, with a separation speed of 700 bases/hour/channel. The instrument's operation has been fully computerized with automatic control of the laser, high voltage power supply, data acquisition and processing. The gel replacement and sample loading were nearly automated, and the sequencing cycle time was less than 2 hours. A standardized baseline protocol for capillary coating and filling and sample preparation has also been developed. Currently, this system has undergone a series of beta-testing runs in the production-sequencing environment at LBNL's sequencing Center. By loading our system with the sample remainders left by production sequencing at LBNL and directly comparing our sequencing results with those generated by the ABI-377 machines, we have found that a comparable sequencing quality has been achieved by our instrument. Details of results will be represented.
This work was supported by the Director, Office of Energy Research, Office of Health and Environmental Research, Human Genome Program, of the U.S. Department of Energy under Contract No. DE-AC03-76SF00098.
Salas-Solano, O., Carrilho, E., Ruiz-Martinez1,
M.C., Goetzinger2, W., Kotler, L., Sosic, Z., and
Capillary electrophoresis (CE) is being developed in our lab and elsewhere for high speed automated DNA sequencing. However, while several multichannel sequencing instruments are currently under development by different groups, the significant issues of robustness and high overall throughput of the DNA sequencing analysis have not been adequately addressed. We are developing a fully automated multicapillary DNA sequencing system, starting from sample preparation to DNA sequence generation. To reduce the failure rate to a minimum and, hence, increase the throughput of such a system, proper care is necessary for every step of the process. We have developed a robust protocol for thorough purification of DNA samples that includes both desalting and template removal. This procedure also resulted in 10-fold increase in the amount of DNA sequencing fragments compared to conventional desalting by ethanol precipitation or gel filtration columns. The purification procedure is economical, very reproducible and compatible with different sequencing chemistries. Conditions for electrokinetic injection of the purified samples have also been optimized. With respect to the column, we previously found that long read length sequencing runs may be achieved using 2% w/w linear polyacrylamide (LPA) with a molecular weight close to 9 MDa. A new, convenient method for high molecular weight LPA preparation using inverse emulsion polymerization has been developed. With this procedure, LPA forms a white powder of unlimited shelf life and allows fast and reproducible preparation of working polymer solutions. Additionally, base calling software has been improved to increase the accuracy of extended read length (see a separate abstract of A.W. Miller and B.L. Karger).
We are currently working on a fully automated multicapillary array DNA sequencing system, which will include a robotic sample preparation and purification system, separation matrix replacement and sample injection automated devices. We will present our latest results in this area. The described sequencing technology advances will greatly increase overall throughput of the automated multi-channel capillary DNA electrophoresis system.
This work is being supported by DOE grant #DE-FG02-90ER60895.
(1) Present address: Curagen Corp., Branford, CT 06504
(2) Present address: Arqule Inc., Medford, MA 02155
Qingbo Li, Thomas E. Kane, Changsheng Liu,
John Kernan, and James R. Hoyland
A high throughput DNA sequencer is constructed, where the instrument operation (e.g., sample introduction, DNA separation, instrument reconditioning, data processing) is carried out and tightly controlled by the instrument computer. The complete instrument consists of two units. The main unit houses the optical detection system, the 96-capillary cartridge, the mechanical system for sample introduction, and the associated electronic controllers. The other unit is the liquid handling module dedicated to gel delivering and capillary reconditioning. The two units interface through a liquid conducting tubing. In the detection system, an air-cooled argon ion laser is efficiently coupled with the optics to excite fluorescence from all 96 capillaries. The instrument is portable. In the sample introduction system, a carrousel assembly allows automatic processing of seven 96-well sample trays without human intervention.
The high throughput arises primarily from four advantages: (1) the instrument design that allows complete automatic operation; (2) the use of 96 discrete capillaries, with which 96 separate DNA samples can be analyzed simultaneously; (3) the use of a high speed CCD camera that allows simultaneous monitoring of fast separation in all 96 capillaries; (4) the use of dilute replaceable gel matrix that minimizes the time required for gel filling and capillary reconditioning. Preliminary sample runs of 400 - 500 DNA bases per capillary have been achieved, yielding 38,400 - 48,000 bases read per instrument run. Including the time for sample introduction, separation, and capillary reconditioning, the PATCO prototype is capable of one complete run within two hours. This sample analysis throughput is comparable to the demands of the Human Genome Project.
The 96-capillary prototype will be available by the end of the year. In the next stage, the instrument will be scaled up to 384 capillaries for 4X improvement in the instrument throughput.
Courtney Davidson, Joseph Balch, Larry Brewer,
Joe Kimbrough, Steve Swierkowski, David Nelson,
Ramkrishna Madabhushi, Ron Pastrone, Ann Lee,
Paula McCready, Aaron Adamson, Bob Bruce,
Ray Mariella, and Anthony Carrano
We are developing instrumentation for DNA sequencing based on the use of an array of microchannels fabricated on a glass substrate. Arrays with up to 101 microchannels, 48 cm long have been fabricated in plates of borosilicate float glass. Since last reporting we have further improved our microfabrication process to etch channels of arbitrary width and depth. We have etched substrates with 12, 24, and 101 channels per plate varying from 150 - 200 um wide and 30 - 60 um deep. The microchannels are constructed with two plates of glass (nominally 7.5 cm x 58 cm) which are fusion bonded. The channel plate is bonded to a top plate that has the input and output ports for sample introduction and buffer reservoir interconnects. Channels of various cross section sizes have been built and tested using low viscosity solutions of linear poly(dimethylacrylamide). This sieving media dynamically coats the walls of the glass microchannels to significantly reduce the electro-osmotic flow so that sequencing separations can be done in uncoated channels. Also, the relatively low viscosity of the sieving media allows it to be pumped into the channels via a simple syringe pump. This syringe pump is coupled directly to a common output port of the microchannel plate so that all channels can be filled with sieving media in one simple pump operation. A linear scanning confocal PMT-based detection system is used to detect the laser induced four color fluorescence from the DNA fragments. The data acquisition and control system is based on a Pentium class personal computer running LabVIEW. Data compression and transfer, signal analysis, and basecalling routines have been developed in S-PLUS and C on a Sun Microsystems Ultra Enterprise 2 server. Recent experimental results of microchannels 60 um deep by 250 um wide and having a 38 cm load-to-read length resulted in an electrophoretic resolution greater than 400 bases in about 90 min. for a 160 V/cm separation field. A 24 microchannel plate is presently operational providing electrophoretic resolution of 450 to 500 bases. Currently we are building and assembling a 96 channel system based on this technology into an "alpha-phase" DNA sequencing instrument which will be networked with the Sun for data analysis and base calling. We will report on-going results obtained from it for high throughput DNA sequencing.
Work is supported by a grant from the National Center for Human Genome Research, National Institutes of Health and by The Department of Energy Human Genome Program. Work was performed by Lawrence Livermore National Laboratory under auspices of the U. S. Department of Energy under contract no. W-7405-ENG-48.
Steve Swierkowski, Joseph Balch, Courtney
Davidson, and Lisa Tarte
We have developed a process for the production of microchannel arrays on single glass substrates as an alternative electrophoresis technology to arrays of discrete capillaries for DNA sequencing. This technology approach provides a number of advantages for building large arrays of electrophoresis microchannels for DNA sequencing. By fabricating the array of microchannels on a single glass substrate, the arrays of microchannels are very robust mechanically and can be handled without any special care. By means of photolithography and chemical etching techniques the dimensions of rectangular cross-section channels can be optimized by making the channel depth thin to minimize the thermal dispersion of DNA bands while at the same time the channel width can be made large to increase the amount of dye-labeled DNA available for strong fluorescence signal generation and detection. The detection of the fluorescence signal is also made easier by having a flat optical window over the channels through which laser excitation of fluorescence occurs with less scattered light of the primary laser beam to contribute to the overall noise level.
Microchannel arrays for electrophoresis with up to 101 channels, 48 cm long have been fabricated in plates of borosilicate float glass. The channels are constructed with two plates of glass that are 7.6 cm wide and 58 cm long and are fusion bonded at 650C. The channel plate is 5 mm thick and typically has 12, 24, or 101 channels per plate; the channels are 150 - 200 um wide and about 30 - 60 um deep. The channel plate is bonded to a top plate that has the input and output ports in it. Two different types of input ports have been tested. For a 5 mm thick top plate, input ports 1 mm in diameter have been ultrasonically milled through the top plate and registered with the channels before the bonding process. For a 1.2 mm thick top plate, input ports as small as 150 um have been fabricated(about the same as the channel diameter) and these have been registered to within 20 um accuracy to the channels before bonding.
The patterning of these plates employs simple contact printing with flat panel display industry type photomasks onto standard photoresists. Special apparatus was constructed to coat the plates with photoresist and also to expose them with a simple contact printing method. A critical procedure was developed to eliminate microscopic damage to the glass before processing begins and to clean the glass at the beginning of the processing. This special procedure was essential to reduce the microchannel etching defects by many orders of magnitude that would have otherwise rendered the plates useless for high resolution genome sequencing. Extensions of this fabrication technology should enable very high(400 - 500) channel count plates to be made that would, in turn, greatly increase the throughput and efficiency of sequencing instruments.
Work is supported by a grant from the National Center for Human Genome Research, National Institutes of Health and by The Department of Energy Human Genome Program. Work was performed by Lawrence Livermore National Laboratory under auspices of the U. S. Department of Energy under contract no. W-7405-ENG-48.
Laurence R. Brewer, Joseph Kimbrough,
Courtney Davidson, and Joseph Balch
We have developed a technique for automatically aligning a microchannel plate with a scanning fluorescence detector in a high throughput DNA sequencer. An optical signal from each microchannel can be used to dramatically reduce the amount of data collected while further eliminating the effects of hysteresis and velocity variation of the scanning motor. While the system can be run in a high resolution mode (~2800 points per scan) for diagnostic purposes such as evaluating band shape, the described technique can be adapted for the collection of a single data point per channel per scan. Such a reduction is prerequisite for practical application to high throughput DNA sequencing. The technique makes use of the difference in reflection of the air-glass, gel-glass, and bonded glass-glass interfaces present in the microchannel plate. A thin piece of glass is used to collect back reflected light from laser illuminated microchannels and sent to a photodiode for detection. This signal delineates the position of the microchannels with a high signal to noise ratio and is used as an electronic trigger for data collection.
Work is supported by a grant from the National Center for Human Genome Research, National Institutes of Health and by The Department of Energy Human Genome Program. Work was performed by Lawrence Livermore National Laboratory under auspices of the U. S. Department of Energy under contract no. W-7405-ENG-48.
Mark A. Quesada, Harbans S. Dhadwal, David J.
Fisk, Janine S. Graves and F. William Studier
Capillary electrophoresis through a replaceable polymer matrix has great promise for improving the speed and efficiency of DNA sequencing if many different capillaries can be analyzed simultaneously. However, illumination of the interiors of multiple capillaries and collection of the emitted fluorescence from each is complicated by the cylindrical shape and small radii of curvature of the capillaries. We have solved this problem by using the capillaries themselves as optical elements in a waveguide.
Refraction of light at the surfaces of a capillary depends on the radii of curvature of the capillary walls and the change in refractive index in crossing each surface. With appropriate dimensions and refractive indices, refractive effects can confine a beam of light to pass through the interior of each successive capillary in a parallel array. The illuminating beam must be in the plane of the array and normal to the first capillary, have minimal divergence, and have a radius comparable to or smaller than the internal radius of the capillary. This condition is readily achieved by delivering the beam through an integrated fiber optic transmitter.
Losses of light due to reflection at each successive capillary surface can be minimized by reducing the differences in refractive index across the surfaces while still satisfying the conditions for beam confinement. Illuminating with coaxial beams from opposite ends of the array also improves the uniformity of illumination. Using commercially available materials, it would be feasible to make 96-capillary waveguide sequencers that would be expected to have only a few percent variation in the intensity of illumination of each capillary across the entire array. The fluorescence can be collected by an array of optical fibers whose spacing is identical to that of the capillaries and whose ends are positioned normal to the capillaries. This configuration allows simultaneous alignment of all of the collection fibers.
A 12-capillary prototype sequencing apparatus was fabricated and tested, validating the key elements of this design. The capillaries are illuminated efficiently with only 30 mWatts of laser power, and the fluorescence is efficiently collected by the matched optical fibers with little cross-talk between channels. The collected light is delivered to a spectrograph and the full fluorescence spectrum of all the capillaries in parallel is displayed on the surface of a CCD and read into a computer about 3 times per sec. Replacement of capillaries and alignment of the system is very simple.
Samples are electrokinetically injected simultaneously into all 12 capillaries from a single row of a microtiter plate. A dimethylpolyacrylamide polymer matrix is used in uncoated capillaries. Our current preparations can be used for about 30 sequencing runs per capillary before performance begins to degrade. We are developing base calling software based on Bayesian analytical methods. The system currently reads almost 500 bases per capillary when run at room temperature, with a cycle time of about 2 hr.
E.R. Mardis, L. Hillier, A. Chinwalla, M. Cook, M. Holman, G. Marth, R.
McGrane, D.A. Panussis, D.C. Peluso, L.L. Rifkin, J.E. Snider, J. Strong, E. Stuebe, R.H.
Waterston, M. Wendl, R.K. Wilson
Efficient performance of high throughput, large-scale genome sequencing projects depends upon improvements to techniques and devices for their performance that streamline data production, as well as computer tools to complement these improvements. Thus, the ability to study processes and apply the appropriate modifications, devices and/or informatics tools is critical to our ability to increase productivity and efficiency. Several examples of technology development and informatics contributions to our processes will be highlighted. These include efforts to increase sample capacity on DNA sequencers, improve sample loading processes, and streamline liquid transfer steps, as well as scripts and algorithms to reduce time spent on gel retracking, to facilitate sequence data entry and to examine assembled projects.
Hong Cai, Peter M. Goodwin, James H. Jett,
Richard A. Keller, Nicholas P. Machara, Susan
L. Riley, and David J. Semin
Our flow-cytometry based approach to DNA sequencing involves: (1) labeling DNA fragments with base-specific fluorescent tags; (2) suspending a single labeled fragment in a flow stream; (3) digesting with an exonuclease that sequentially cleaves the end nucleotide and releases it into the flow stream; (4) detecting and identifying the individual cleaved nucleotides as they pass in order of cleavage, through a focused laser beam.
We have implemented an optical trap to suspend the DNA laden microsphere upstream from the detection laser. This resulted in: reduced background; improved detection efficiency; and simplified sample handling. With this system, our detection efficiency of labeled nucleotides is ~ 90% and false positive signals from the background are a few per second. Enzymatic cleavage rates with Exo III are ~ 5 per second at 37 C on fluorescently labeled substrates. Progress towards a two color sequencing demonstration will be described.
Hong Cai, # Susan L. Riley, # Kristina
Kommander, * John Nolan, * Richard A. Keller#
The limitation of current automated sequencers is 1800 base pairs/day, which results in the need for 4600 machine years to create the first finished sequence of one human genome ( ~3 x 109 bases pairs). A rapid laser-based technique for sequencing of 10 kb or larger fragments of DNA at a rate of 100 to 1000 bases per second is being developed in our laboratory. Successful completion of this would greatly reduce the time and effort needed in the sequencing of the human genome and other genomes. This new method relies on the attachment of fluorescent labeled DNA to a microsphere, introduction of this microsphere into a flowing sample stream, and detection of the individual labeled nucleotides as they are cleaved from the DNA fragment by an exonuclease. In order to increase analysis rates, an exonuclease with a fast digestion rate is required.
Extensive testing of commercially available exonucleases in our laboratory has not revealed a suitable exonuclease capable of rapidly cleaving fluorescently-labeled DNA. This has lead us to search for either mutant forms of exonucleases or exonucleases derived from other bacterial strains. In order to screen the numerous exonucleases, a rapid screening assay based on flow cytometry has been developed. Compared to conventional techniques, this new assay is sensitive, rapid, and requires no radiation labeling or separation. This will allow the screening of hundreds of samples per day.
Zhengping Huang, Yongseong Kim, Jonathan L.
Longmire, Nancy C. Brown, James H. Jett and
Richard A. Keller
Our flow-cytometry based approach to DNA fragment sizing involves: (1) staining a restriction digest of DNA with a dye that intercalates stoichiometrically with the fragments such that the number of incorporated dye molecules is proportional to the fragment length; (2) diluting the sample to ~ 10-13 M; passing the sample through our single molecule detection apparatus and (3) measuring the magnitude of the fluorescence from individual, stained fragments. A histogram of the fluorescence intensities gives the size distribution of the DNA fragments, i.e. a DNA fingerprint. Samples containing less than a femtogram of DNA are sized in minutes with an accuracy of ~ 98%. We have demonstrated the applicability of this technique for sizing DNA fragments as small as 212 bp and as large as 340 kbp. In comparison with pulsed-field gel electrophoresis for the sizing of large DNA fragments, this approach is more accurate, much faster, requires much less DNA, and is independent of the DNA conformation. Applications to the characterization of PAC and BAC clones for DNA library construction and identification of bacteria strains by their DNA fingerprint will be described.
Brian B. Haab and Richard A. Mathies
We have shown that single-molecule fluorescence burst counting is a highly sensitive method for detecting electrophoretic separations of ds-DNA fragments.1 In previous work, methods for optimizing dye labeling, laser power and data analysis were developed, which enabled detection of single DNA fragments as small as 100 bp in capillary electrophoresis separations.2 These separations were detectable when only 50-100 molecules passed through the probe volume. To further enhance the applicability of this method to low level pathogen and mutation detection, we have now successfully performed single molecule detection of DNA separations in microfabricated glass capillary electrophoresis (CE) chips. By fabricating CE chips with a 280 mm thick top cover plate and by using a 40X NA 1.3 immersion microscope objective, the S/N ratio for single molecule detection was enhanced by more than two-fold over conventional capillaries. By constricting the sample in the detection region to a ~10 mm wide by ~5 mm deep cross section, approximately 10% of the molecules could be probed by the ~2 mm wide focused laser beam. Cross channel and separation channel dimensions were systematically varied to optimize the injection of the DNA sample. We have now achieved a detection limit of 500 molecules on-column or 200 attograms/ml for 500 bp DNA fragments.3 This accomplishment is important because DNA-based methods are becoming increasingly important in environmental monitoring and in health care diagnostics to detect trace levels of pathogen contamination or DNA mutation.
Supported by the Director, Office of Energy Research, Office of Health and Environmental Research of the U.S. Department of Energy under Contract DE-FG-91ER61125.
1 B. B. Haab and R. A. Mathies, "Single molecule fluorescence burst detection of DNA fragments separated by capillary electrophoresis," Anal. Chem. 34, 3253-3260 (1995).
2 B. B. Haab and R. A. Mathies, "Optimization of single molecule fluorescence burst detection of ds-DNA: Application to capillary electrophoresis separations of 100-1000 bp fragments," Appl. Spec., 51, N10 (1997).
3 B. B. Haab and R. A. Mathies, in preparation.
W. Henry Benner
Mass spectrometry has important potential applications in the measurement of molecular masses relevant to the goals of the Human Genome Project. Thin gels, capillary gels and DNA chip technology appear to offer improvements in DNA analysis rates over slab gels but "mass spectrometry has perhaps demonstrated the greatest near-term potential [to increase through-put]." To the extent that direct instrumental measurements of molecular size could replace gel electrophoresis as a routine tool for DNA-sequencing separations and general molecular biology experimentation, a significant saving in time could be effected. Electrophoresis gels take hours to run but mass spectra are acquired in seconds to a few minutes. Additionally, mass spectrometers produce a mass measurement compared to gel electrophoresis which separates ions according to mobility, a relative measurement.
Electrospray mass spectrometry is an important type of mass spectrometry applicable to DNA analysis because it does not break apart DNA molecules. For relatively small electrospray DNA ions, the different charge states provide a way to calculate mass but for large ions with numerous charge states the calculation of ion mass is precluded because the charge states are not resolved in most mass spectrometers. We recently demonstrated the direct detection of charge on large electrospray ions as a way to solve this problem so that the mass of large DNA ions could be measured.
A patent application has been submitted for the invention of charge-detection-mass-spectrometry. More recently, we have significantly improved the measurement capability of this technique by implementing this detection system in a new type of ion trap. The gated electrostatic trap consists of a detector tube mounted between two sets of ion mirrors. The mirrors define symmetrically-opposing potential valleys which guide axially-injected ions to cycle back and forth through the charge-detection tube. A low noise charge-sensitive amplifier, connected to the tube, reproduces the image charge of individual ions as they pass through the detector tube. Ion mass is calculated from measurement of ion charge and velocity following each passage through the detector. Individual ions carrying more than 250 charges at an energy of 200 eV/charge have been trapped for 10 ms corresponding to 450 cycles through the detector tube. At this level of trapping time, a theoretical precision for charge measurement as small as 2 electrons RMS can be achieved. The operation of the system was demonstrated by trapping a 4.3 kilobase long circular DNA molecule of bacterial plasmid pBR322. The sodium form of this molecule has a molecular weight of 2.88 MDa. A mass value of 2.79 +/- 0.09 (ave +/- s.d.) MDa was determined. The accuracy of the mass measurements and the speed of this technique suggest that this measurement approach could be applied to the routine sizing of cloning vectors for the purpose of quality control of the cloning process.
Christopher S. Martin, Jing Ying Lee, Betty Liu,
Jeffrey Shumway, John C. Voyta and Irena Bronstein
Nylon membrane is the preferred support for a multitude of molecular biology applications due to its robustness and retension of high levels of bioanalytes. Chemiluminescent detection with 1,2-dioxetane substrates on nylon membrane is facilitated by the enhancement properties of nylon, but limited by high levels of non-specific binding of enzyme labeled reagents. We have developed polymers of quaternary amines for enhancement of chemiluminescence intensity from 1,2-dioxetanes in solution. These polymers also improve membrane assays when added to the substrate buffer. Poly(benzyltributyl)ammonium chloride(TBQ)(SapphireAE II) and Poly(benzyldimethylvinylbenzyl)ammonium chloride (BDMQ) (SapphireAE I) are current Tropix chemiluminescent enhancer products. Various quaternary amine polymers were screened to determine if membranes coated with such polymers would exhibit superior performance compared to commercially available nylon, PVDF, or nitrocellulose supports. Due to the superior chemiluminescent enhancement properties of THQ (Polyvinylbenzyltrihexylammonium chloride), this polymer was used to overcoat different membrane supports. After overcoating with THQ, biotinylated DNA was detected with a chemiluminescent dot blot assay. Polyethersulfone membranes exhibited the best performance in these assays. Recently, we have collaborated with an outside membrane manufacturer to bench cast membranes with THQ polymer. These membranes were subsequently tested by performing chemiluminescent detection of biotin or fluorescein labeled DNA in dot blot and Southern assays. The results indicate that sensitive detection and low background signal are attained on these membranes. Development of a superior membrane for chemiluminescent assays is of great benefit, enabling more rapid imaging of signals with x-ray film and electronic imaging devices (i.e. CCD cameras).
This work was funded by the DOE Genome Program.
Contract No. DE FG05 92ER81389
Adam T. Woolley, Peter C. Simpson, Shaorong
Liu, Kaiqin Lao, Stephen J. Williams, Mary X.
Tang, Lester Hutt, Alexander N. Glazer and Richard A. Mathies
The microfabrication of DNA sample preparation, electrophoretic analysis and detection devices is making possible a new generation of high-speed, high-throughput DNA analysis systems. Our early work showed that high-quality fragment sizing as well as DNA sequencing separations could be performed on microfabricated capillary electrophoresis (CE) chips.1,2 We also demonstrated that PCR amplification could be directly performed on our CE chips to make the first integrated DNA sample preparation and analysis devices.3 It was also possible to increase the throughput of these microdevices by making capillary array electrophoresis (CAE) chips that could carry out parallel genotyping separations of up to 12 samples on a single chip in under 160 seconds.4 Recent advances in the use of replaceable denaturing separation matrices, in injection methodology, and in channel fabrication now enable DNA sequencing separations on chips with single base resolution to =500 bases in only 12 minutes. Improvements in fabrication permit the construction of larger defect-free devices on 10 cm diameter glass wafers.5 We have also developed (i) novel injection modules for serially introducing multiple (2-4) samples onto the same capillary, (ii) elastomer sample well arrays for the facile loading of up to 96 samples, and (iii) electrode arrays for addressing up to 96 samples. These improvements have led to the development of a CAE chip that can separate 96 DNA fragment samples in less than 8 minutes using 48 parallel capillaries, each capable of analyzing two different samples.6 These analyses have all been performed by using high-sensitivity, laser-excited confocal fluorescence detection. However, to produce truly portable high-speed microdevices it is desirable to eliminate expensive and bulky optical components and laser systems. Toward this end we have been working on the development of integrated electrochemical detection systems for CE chips.7 In these devices, the working electrode, formed by RF sputter deposition of Pt (2500 Å) on a 200 Å Ti adhesion layer, was photolithographically placed =30 um outside the end of the separation channel to avoid interference from the separation field. Electrophoretic separations of neurotransmitters were performed in under 100 seconds demonstrating the speed, resolution and attomole detection sensitivity of these devices. Indirect electrochemical detection of DNA fragment separations was performed by using the redox-active intercalator Fe(1,10-phenanthroline)32+ in the separation buffer; transient dips in the constant background current from free intercalator indicated the migration of DNA-intercalator complexes through the detection region. On-chip electrochemical detection provided excellent sensitivity (=103 molecules) for rapid (=200 s) and high-quality separations of DNA restriction fragments and PCR products. This work is the harbinger of a paradigm shift in the application of CE chips to genomic sequencing and analysis.
Supported by the Director, Office of Energy Research, Office of Health and Environmental Research of the U.S. Department of Energy under Contract DE-FG-91ER61125.
1 Woolley, A. T.; Mathies, R. A. Proc. Natl. Acad. Sci., USA 91, 11348-11352 (1994).
2 Woolley, A. T.; Mathies, R. A. Anal. Chem. 67, 3676-3680 (1995).
3 Woolley, A. T.; Hadley, D.; Landre, P.; deMello, A. J.; Mathies, R. A.; Northrup, M. A. Anal. Chem. 68, 4081-4086 (1996).
4 Woolley, A. T.; Sensabaugh, G. F.; Mathies, R. A. Anal. Chem. 69, 2181-2186 (1997).
5 Simpson, P. C.; Woolley, A. T.; Mathies, R. A. BioMEMS 1, in press (1997).
6 Simpson, P. C.; Roach, D.; Thorsen, T.; Johnston, R.; Sensabaugh, G. F.; Mathies, R. A. manuscript in preparation (1997).
7 Woolley, A. T.; Lao, K. Glazer, A. N.; Mathies, R. A. submitted for publication (1997).
Kenneth Beattie, Mitchel Doktycz, Ming Zhan,
Gabriel Betanzos, William Bryan, John Turner
A flowthrough genosensor instrument for ultrahigh throughput analysis of gene structure and function is being developed. The core of this microscale instrumentation is a microchannel hybridization array, containing a library of thousands of specific DNA sequences, immobilized within microporous cells in a thin layer of silicon or glass. A nucleic acid sample is passed through the microchannel genosensor at precisely controlled temperature and flow rate, and each porous hybridization cell binds nucleic acid sequences that are complementary to the immobilized DNA probe. The quantitative binding pattern reflects the base sequence of the nucleic acid strands present in the analyte and reveals the relative abundance of different sequences. The porous glass configuration has several important advantages over the flat surface DNA chip being developed by others: greatly improved hybridization kinetics, superior detection sensitivity, the ability to analyze dilute solutions of nucleic acids, and direct detection of heat-denatured PCR fragments without prior isolation of single strands.
The prototype genosensor system includes a temperature-controlled fluidics module and a CCD imaging system for quantitation of hybridized fluorescent-labeled strands. A key objective in the project is to develop important applications of the flowthrough genosensor for analysis of gene structure and function and DNA diagnostics. Feasibility studies for key genosensor applications are being pursued in collaboration with the mouse genetics program at ORNL and with several other organizations. In one application area, a series of miniature "genochips" containing arrays of genomic DNA fragments are being prepared for use in gene discovery and mapping. Another application area, being pursued via a CRADA with Gene Logic, Inc. (Columbia, MD) and in collaboration with Dr. Jeffrey Trent (NHGRI) employs flowthrough arrays of DNA probes for transcriptional profiling, facilitating the discovery of genes that function in specific biological processes. A third application of the flowthrough genosensor, led by Dr. Mitch Doktycz, involves model hybridization studies with defined nucleic acid sequences, aimed at providing a more complete understanding of the specificity of oligonucleotide hybridization, which will facilitate intelligent selection of probes and valid interpretation of hybridization patterns. A fourth application of the flowthrough genosensor, being pursued in collaboration with Dr. James Weber at the Marshfield Medical Research Foundation, is high throughput genotyping. In this work miniature flowthrough genosensors will be fabricated to simultaneously analyze thousands of biallelic single nucleotide polymorphisms.
In another approach, the ultrahigh surface area of channel glass is being exploited to create arrays of "microreactor cells" containing immobilized BAC DNAs, for use in repetitive reactions needed for genome mapping and sequencing. These reactions will include: (1) Cycle sequencing reactions - Sequencing primers, nucleotides and Taq polymerase are flowed into each of the porous glass "microreaction cells," then the wafer will be sealed and placed into a thermocyler to carry out the cycle sequencing reactions. Products will be eluted and analyzed by Dr. Joe Balch at LLNL, using his parallel microcapillary array electrophoresis apparatus. This process will be carried out in numerous successive cycles, each time with a new set of primers, to acquire a large amount of sequence information from each BAC. The sequencing primers can be selected from oligonucleotide libraries as suggested by Ulanovsky and others, to achieve rapid primer walking along the entire set of BACs. (2) Mapping of expressed sequences in BAC libraries - Libraries of BACs immobilized in the channel glass array will be hybridized with individual cDNA clones or ESTs to localize each expressed sequence across the BAC array. This process will be repeated with numerous expressed sequences, to achieve rapid assignment of expressed sequences to their genomic clones.
Damir Sudar, Steve Clark, Ian Poole, Rick
Segraves, Stephen Lockett, Arthur Jones, Donna
Albertson, Joe Gray, Daniel Pinkel
We have developed a method of performing comparative genomic hybridization (CGH) to microarrays of genomic DNA clones (cosmids, P1s, BACs, etc.) that permits high resolution detection and mapping of DNA copy number variations in the human genome (Albertson et al., this meeting). In array CGH, spots of cloned DNA are arrayed onto a microscope slide and hybridized with total genomic DNA from a test specimen, labeled with one fluorochrome, and reference genomic DNA labeled with a spectrally different fluorochrome. The ratio of the fluorescence intensities on each target clone is proportional to the relative copy number of those sequences in the test and reference genomes. Efficient implementation of array CGH requires overcoming several major technical challenges including production of high density arrays, and rapid quantitative readout and analysis of the fluorescence signals. We have made considerable progress in both of these areas.
We are developing a robotic arraying system with a multi-pin tool to print DNA target solutions from 864 well plates onto fused silica slides mounted on a precision X-Y stage. The pins, each made from a segment of capillary electrophoresis tubing are spring mounted on 3 mm centers. The upper end of each pin is connected to a manifold by flexible tubing to permit pressurization for printing and suction for cleaning. We have demonstrated the feasibility of printing targets on 100 um centers using this approach, providing a density of 10,000 targets / cm2.
Our imaging system was designed to provide: (a) large field-of-view (1cm2, to match the print format), (b) sufficient resolution (10 pixels per target diameter), and (c) high sensitivity (accurate quantitation down to single-copy DNA quantities). A digital CCD camera and 2 high-speed lenses in a back-to-back configuration image the sample at 1x magnification. Fluorescence is excited using a mercury arc lamp with filter wheel for wavelength selection. A fiber-optic delivery system illuminates the sample at a 45 degree angle through the back side of the slide. A right-angle fused-silica prism is used in "total internal reflection" conditions to efficiently excite the fluorescent dyes without allowing excitation light to enter the imaging optics. A multi-band barrier filter was used in the emission path so it did not have to be changed when imaging different fluorochromes. We acquire a DNA counterstain (DAPI) image and the test and reference images with an exposure of several seconds or less. Software automatically identifies the spots using the DAPI counterstain and measures the background-corrected fluorescence intensities and intensity ratios of the spots.
We evaluated the performance of the system and analysis software using test samples made by spiking 200 ng of total human genomic DNA with 0, 1, 2, 20, 200, and 2000 pg of lambda DNA and a reference sample consisting of 200 ng of total genomic DNA spiked with 20 pg of lambda DNA. In this situation 3 pg of lambda DNA is equivalent to a single copy sequence. We found that the changes in the fluorescence ratios were detectable from below single copy equivalent level, and were quantitatively proportional to DNA sequence copy number over three orders of magnitude.
Partially supported by the Director, Office of Energy Research, Office of Health and Environmental Research of the U.S. Department of Energy under contract NO. DE-AC03-76SF00098.
T. Vo-Dinh*, D.L. Stokes1, G.D. Griffin1,
N. Isola1, J.P. Alarie1, U.J. Kim2, M.I.
Simon1, T. Bunde1
We describe a new type of DNA gene probe based on surface-enhanced Raman scattering (SERS) label detection. Raman spectroscopy is a useful tool for chemical analysis due to its excellent capability of chemical group identification. One limitation of conventional Raman spectroscopy is its low sensitivity, often requiring the use of powerful and costly laser sources for excitation. However, a renewed interest has recently developed among analytical spectroscopists as a result of observation that Raman scattering efficiency can be enhanced by factors of up to 10(8) when a compound is adsorbed on or near special metal surfaces. The technique associated with this phenomenon is known as Surface-Enhanced Raman Scattering (SERS) spectroscopy. The surface-enhanced Raman gene (SERGen) probes described here do not require the use of radioactive labels and have great potential to provide both sensitivity and selectivity for DNA sequencing. The SERGen probes can be used to detect DNA biotargets (e.g., gene sequences, bacteria, viral DNA fragments) via hybridization to DNA sequences complementary to that probe.
The use of stable clone resources containing large human DNA insets has opened new possibilities to contig building for the Human Genome Project. The objective of this research is to apply the SERS multi-label technique for use in DNA mapping and bacterial artificial chromosomes (BAC) colony hybridization. The technology is based on a system that will integrate several concepts including: i) multi-label SERS detection, ii) spectral multiplex mapping, and iii) BAC colony hybridization.
Emphasis is on detection techniques that minimize the time, expense and variability of preparing samples by combining the BAC mapping approach with SERS "label multiplex" detection. Large numbers of DNA samples can be simultaneously prepared by automated devices. With this device, multiple samples can be separated and directly analyzed using multiple SERS labels simultaneously. The results demonstrate the feasibility of the SERGen approach in the detection of two gene probes simultaneously.
Technologies for robotic manufacturing of Microarrays of Gel-Immobilized Compounds on a chip (MAGIC chip) have been developed and the MAGIC chips are being tested for various applications. These microchips are polyacrylamide gel pads fixed and separated on a glass surface by hydrophobic spacers. The microchips can be used like an array of micro test tubes in which chemical and biochemical reactions can be carried out separately in each gel pad. Different oligonucleotides, DNA antibodies, and proteins have been immobilized in specified gel pads to produce oligonucleotide, DNA, and protein microchips. The usual size of the gel pads is 100x100x20 m. However, microchips with gel element sizes as small as 10x10x10 m can be produced by photopolymerization. Applications of oligonucleotide microchips have been demonstrated for detection of mutation, identification of microorganisms, HLA allotyping, DNA fractionation, DNA enzymatic phosphorylation and ligation on specified or all microchip elements, and thermodynamic analysis of DNA duplexes. Generic microchips containing all 4,096 possible 6 mers have been manufactured, and their use for de novo sequencing and DNA sequence analysis will be described.
John P. Nolan, Hong Cai, Kristina Kommander, and P. Scott White
Single-base polymorphism analysis of the human genome on a large scale requires robust and sensitive screening methods which are amenable to automation and high throughput analysis. We are developing a suite of microsphere-based approaches employing fluorescence detection by flow cytometry to screen for and analyze single base polymorphisms. One approach being developed for polymorphism detection is to immobilize on microspheres proteins which recognize specific DNA structures. When a fluorescently labeled DNA molecule binds to the immobilized protein, the binding can be measured by flow cytometry. An example is the recognition of heteroduplex DNA by the mutS protein as a means to detect single base mismatches. To analyze nucleotide base frequencies at a polymorphic site, we are developing an approach based on minisequencing in which immobilized primers designed to interrogate a specific site are used to bind the region of interest in an unknown sample. The primers are then extended by polymerase using fluorescent ddNTPs and flow cytometry is used to read the frequency of each differently colored base. Alternately, the ligation of fluorescent oligonucleotides containing base variation at the site of interest is used to detect the base frequency at that site. Apart from the advantages of sensitivity and low sample consumption, the flow cytometric approaches have the advantages of the potential for multiplexed analysis using different color or size beads and automated sample handling. Multiplexed analysis could enable simultaneous analysis of base frequencies at dozens of different loci, which combined with automated sample handling could provide a powerful tool for high throughtput screening of single base polymorphisms. Supported by NIH, DOE, and LDRD.
David S. Wunschel, Ljiljana Pasa Tolic, David. C.
Muddiman, James E. Bruce, Steve A. Hofstadler
and Richard D. Smith
Mass spectrometry offers the potential for high speed DNA sequencing and other applications. In addition to the development of sequencing approaches, ongoing work in the laboratory is exploring applications using Fourier transform ion cyclotron resonance (FTICR) mass spectrometry. These efforts include the characterization of polymerase chain reaction (PCR) products, enzymatically produced oligonucleotide mixtures, modified DNA and the development of methods for the analysis of DNA large fragments. The analysis time required is on the order of seconds, and is made possible by isotopic resolution of each component's charge states obtained using FTICR. High mass accuracy measurements for PCR products have been achieved for products up to 114 base pairs in length. As an example, the mass accuracy allowed single base substitutions to be detected with de novo identification of an unreported base substitution (1,2). This approach was extended to examine a multi-component reaction from a single primer pair where a base pair deletion was identified with the putative identification of inter- variability within a single bacterial strain (3). Recent efforts have focused on increasing the size of products amenable to analysis with a goal of providing comparable "read-lengths" to traditional sequencing methods, and have involved improvements to sample preparation methods and the exploitation of improved methods for dynamic range expansion. We are also exploring the use of collision induced dissociation methods with PCR products to provide sequence information. This would allow for direct selection and analysis of individual components from within mixtures that may share a high degree of similarity without cloning. This alternative would not only eliminate that time-consuming step, but also potentially allow identification of low abundance products without an intensive screening process.
In this presentation the recent advances at our laboratory will be described. These include the development of new interfacing methods for realizing greatly improved sensitivity and the implementation of new high performance FTICR that has been designed to achieve greater sensitivity as well as resolution and mass measurement accuracy.
1. "Characterization of PCR products from bacilli using electrospray ionization FTICR mass spectrometry", D. C. Muddiman, D. S. Wunschel, C. L. Liu, L. Pasa Tolic, K. F. Fox, A. Fox, G. A. Anderson and R. D. Smith, Anal. Chem. 68, 3705-3712 (1996)
2. "Length and Base Composition of PCR-Amplified Nucleic Acids Using Mass Measurements from Electrospray Ionization Mass Spectrometry", D. C. Muddiman, G. A. Anderson, S. A. Hofstadler and R. D. Smith, Anal. Chem. 69, 1543-1549 (1997)
3. "Inter-Operon Variability in B. cereus by Normal PCR using ESI-FTICR Mass Spectrometry", D. S. Wunschel, D. C. Muddiman, K. F. Fox, A. Fox and R. D. Smith, submitted.
This research was supported by the Office of Biological and Environmental Research, U.S. Department of Energy. Pacific Northwest National Laboratory is operated by Battelle Memorial Institute through Contract No. DE-AC06-76RLO 1830.
Gregory B. Hurst, Kristal Weaver, and Michelle V. Buchanan
While a wealth of biological and genetic information can be gleaned from properly- designed polymerase chain reaction (PCR) assays, currently-used technologies for analysis of the resulting oligonucleotides all suffer from limitations in speed, accuracy, safety, or convenience. Matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS) offers considerable potential for rapid and accurate molecular mass determination of biopolymers, such as proteins and DNA. In order to achieve this potential, we are working to improve the utility of MALDI-MS for measurement of PCR product size.
The resolution (ability to distinguish products of similar molecular mass) of MALDI-MS is determined by both chemical and instrumental factors. The presence of reaction components necessary for polymerase activity, particularly metal ions, causes broadening of the observed peaks in MALDI mass spectra of PCR products. Post-PCR removal of these metal ions and other interferences can be performed efficiently using reverse-phase cartridges in syringe-mounted or microtiter plate formats. The wide energy distribution imparted to biomolecules by the laser desorption process also broadens mass spectrometric peaks. Delayed ion extraction has greatly improved the resolution of MALDI-MS by compensating for this energy spread. Combining these techniques, PCR products differing in length by a single base can be resolved up to a total length of 60 bases or more, and larger oligonucleotides can be detected if single- base resolution is not required. The design of PCR products that are shorter than typically used for gel electrophoresis is thus a high priority in improving the applicability of MALDI-MS.
Other important practical considerations are the reproducibility and throughput of MALDI-MS. Commercially-available instrumentation allows robotic loading onto multiple-sample plates followed by automated analysis. However, samples outside narrow constraints of concentration, size, and purity still require human intervention because of the inhomogeneity of the dried matrix/sample mixture. We are currently developing methods for combining oligonucleotides with the MALDI matrix to yield a homogenous preparation resulting in a uniform signal at any point interrogated by the desorption laser.
K.W. acknowledges support through an appointment to the Oak Ridge National Laboratory Postgraduate Research Program administered jointly by the Oak Ridge Institute for Science and Education and Oak Ridge National Laboratory. Research supported by the Environmental Management Science Program and Office of Biological and Environmental Research, U.S. Department of Energy, and the Oak Ridge National Laboratory Director's Research and Development Funds. Oak Ridge National Laboratory is managed for the United States Department of Energy by Lockheed Martin Energy Research Corp. under contract DE-AC05-96OR22464.
C. H. Winston Chen, N. I. Taranenko, Y. F. Zhu,
S. L. Allman, V. V. Golovlev and N. R. Isola
During the past two years, we have used laser desorption mass spectrometry (LDMS) to obtain the following major achievements. They are (1) LDMS sequencing of ss-DNA longer than 100 nucleotides with DNA ladders (2) Direct DNA sequencing without ladders (3)LDMS for hybridization detection (4) STR detection for forensic applications and (5) Rapid disease diagnosis.
For conventional gel electrophoresis for DNA sequencing, major steps include (1) DNA ladders preparation (2) Separation of different sizes of DNAs and (3) detection by either autoradiogram or laser-induced fluorescence. Laser desorption mass spectrometry (LDMS) can be used to enhance DNA sequencing speed. One approach is to use a mass spectrometer as a detector only. In general multiplexing is used to increase the sequencing speed. However, gel electrophoresis and DNA ladders preparation are still required. Another approach is to use LDMS for both separation and detection. With this approach, gel electrophoresis is not needed but the preparation of DNA ladders is required. Recently, we succeeded in using LDMS in sequencing single-stranded DNA with the size longer than 100 nucleotides. With primers for both strands, a double-stranded DNA segment with the size up to 260 base pairs can be sequenced. Both cycle sequencing and standard Sanger's sequencing have been tried with successful results.
Since the preparation of DNA ladders is somewhat time consuming, it is very desirable to be able to sequence DNAs without the need of ladder preparation. We recently found that preferred bond cleavage can be obtained during the laser desorption process if adequate matrices and laser fluences are used. We took this approach and recently succeeded in sequencing an oligonucleotide with 35 bases. This direct sequencing by MALDI without the need to prepare DNA ladders can be conveniently used to sequence primers and short DNA probes.
In addition to the DNA sequencing, we also apply LDMS for the detection of DNA probes for hybridization. Preliminary results indicate that LDMS as detector for hybridization process can reduce the time and cost for DNA sequencing by hybridization (SBH). LDMS was also used to obtain short tandem repeats (STR) for forensic applications. Clinic applications for disease diagnosis such as cystic fibrosis due to the base deletion and point mutation have also been demonstrated. Different schemes for resolution and detection efficiency improvements will be pursued in the future to increase the sequencing and/or analysis speed. Experimental details will be presented in the meeting.
Research has been sponsored by the Office of Health and Environmental Research, U.S. Department of Energy under contract DE-AC05-84OR21400 with Lockheed Martin Energy System, Inc.
George Church, Martha Bulyk, Sonali Bose,
Chris Harbison, Linxiao Xu, Poguang Wang,
Laura Kutney, T. O'Keeffe, Dereth Phillips
Together with Bruker Instruments Inc., Genome Therapeutics Corp., and Northeastern University, we have developed a laser-desorption electron capture mass spectrometry method to quantitate up to 400 "Mass-tags" every 10 to 200 milliseconds. Both DNA and proteins have been mass-tagged in ways analogous to fluorescent-tags. The tags are laser-heat-releasable electrophores. Advantages over fluorescence include more numerous cleanly separated spectral peaks, better detection, and higher throughput. Applications to capillary electrophoretic (CE) assays, DNA microarray chips, in situ hybridization, and in situ PCR imaging are under evaluation. The electrophoretic applications are designed (but not yet tested) to collect 100 sequence base pair equivalents per second. To assess the capacity of our CE (75 micron inner diameter), we have obtained dideoxy sequence data from a multiplex PCR sequence of a mixture of 47 ds-templates. Another feature of the Mass-tag is the 400 internal standards, which should enhance the ability to computationally align and quantitate 400 images for chip and in situ applications. For the new complete microbial genome sequences, we have developed chip and CE technologies for systematic analyses of DNA-protein interactions and competitive growth phenotypes.
L. Xu, N. Bian, Z. Wang, S. Abdel-Baky, S. Pillai, D. Magiera, V. Murugaiah, R.W. Giese, P. Wang, T. O'Keeffe, H. Abushamaa, L. Kutney, G.M. Church, S. Carson, D. Smith, M. Park, J. Wronka, F. Laukien. Electrophore Mass Tag Dideoxy DNA Sequencing. Analytical Chemistry in press.
Tetsuyoshi Ono, Mark Scalf, Lloyd M. Smith
DNA fragmentation is a major factor limiting mass range and resolution in the analysis of oligonucleotides by Matrix-Assisted Laser Desorption/Ionization Mass Spectrometry (MALDI-MS). Protonation of the nucleobase leads to base loss and backbone cleavage by a mechanism similar to the depurination reactions employed in the chemical degradation method of DNA sequencing. In a previous study, the stabilizing effect of substituting the 2' hydrogen with an electronegative group such as hydroxyl or fluorine was investigated. These 2' substitutions stabilized the N-glycosidic linkage, blocking base loss and subsequent backbone cleavage. For such chemical modifications to be of practical significance, it would be useful to be able to employ the corresponding 2'-modified nucleoside triphosphates in the polymerase-directed synthesis of DNA. This would provide an avenue to the preparation of 2'-modified PCR fragments and dideoxy sequencing ladders stablilized for MALDI analysis. In this paper methods are described for the polymerase-directed synthesis of 2'-fluoro modified DNA, using commercially available 2'-fluoronucleoside triphosphates. The ability of a number of DNA and RNA polymerases as well as reverse transcriptase to incorporate the 2'-fluoro analogs was tested. Four thermostable DNA polymerases (Pfu (exo-), Vent (exo-), Deep Vent (exo-) and UlTma) were found that were able to incorporate 2'-fluoronucleotides with reasonable efficiency. In order to perform Sanger sequencing reactions, the enzymes' ability to incorporate dideoxy terminators in conjunction with the 2'-fluoronucleotides was evaluated. UlTma DNA polymerase was found to be the best of the enzymes tested for this purpose. MALDI analysis of enzymatically produced 2'-fluoro modified DNA using the matrix 2,5-dihydroxy benzoic acid showed no base loss or backbone fragmentation, in contrast to the extensive fragmentation evident with unmodified DNA of the same sequence.
Wei Yu and Richard A. Gibbs
Baylor College of Medicine Human Genome Sequencing Center
The Baylor College of Medicine Human Genome Sequencing Center (BCM-HGSC) is scaling with the aim of completing 15 Mb of human genomic DNA in the period until April 1998, using a modified shotgun strategy. In order to make this process more efficient, innumerable aspects of the shotgun sequencing process have been optimized and streamlined, providing a good test of this sequencing approach for analyzing large fragments (>50 kb). We routinely now complete genomic fragments using less than 20 reads per kilobase of DNA.
In order to efficiently sequence short (1 - 5 kb) DNA fragments we have devised another shotgun based strategy entitled Concatenation cDNA Sequencing (CCS). CCS has been applied to the joining of multiple independent cDNA molecules to form long concatemers, and then generation of shotgun libraries. The libraries are then sequenced in a similar manner to cosmid, BAC or PAC libraries and the sequences of the cDNA molecules are resolved as individual contigs in the final computer assemblies.
More than 500 cDNA full insert sequences from Soares libraires have now been initiated in libraries averaging 50 inserts. More than 150 cDNAs have been finished and submitted to GenBank, and 250 are in closure. The complete sequences show that the method is as efficient as when sequencing comparable clones with the same total length as the combined cDNAs, and libraries have been competed with as few as 14 reads/kb. Simplified methods for pooling cDNA preparations have been developed, and problems associated with individual cDNAs that evade concatenation have been solved. Our aim is to complete 1,000 cDNAs insert sequences within 1998.
Trey Ideker, Richard Karp, and Leroy Hood
From the perspective of the data analyst, current DNA array technology is in the early stages of development and its data hard to reproduce across multiple experiments. A comprehensive array system for the determination and interpretation of gene transcription rates is under development with a focus on obtaining well-characterized data for transcriptional network analysis. The acquisition scheme includes a robot with a custom-designed modular print head which deposits cDNA onto glass slides and a commercial fluorescent imaging machine which detects array hybridizations. Software to locate and quantitate samples in the image has also been developed. The software generates a list of expression levels for each gene and produces an estimate of the fluorescence background local to each sample spot on the array. We have statistically characterized the performance of the array process so that each measured expression level also has an associated confidence metric. This metric reflects the measurement error in the expression levels of each gene and includes the deviation in identical experiments performed several times. Variation is due to error in the array robotics, DNA hybridization and attachment chemistry, mRNA sample preparation, fluorescence detection, and naturally occurring expression differences between separate RNA samples. A series of known test samples was deposited and analyzed using the array system in order to characterize each area of variation. Future work will focus on an in-depth analysis of the complex expression data. This analysis includes formulation of a general computer model of gene transcription and implementation of an algorithm to predict biochemical pathways from transcription rate data.
Trevor L. Hawkins1, Laurent Jacotat2 , Mary Pat
Reeve1 and Kevin McKernan2
Over the last three years we have developed and put into full production a number of automated systems for DNA purification, DNA sequencing set-up and processes involved in high throughput genome analysis. Much of our recent work was described in Hawkins et al Science 276: 1887- 1889.
Methods and Protocols. We have continued to develop applications for Solid-Phase Reversible Immobilization1 (SPRI) to extend to M13 and PCR2 product isolation and purification. We have also demonstrated the use of SPRI for the clean-up/ desalting of Dye-Primer and Dye-terminator sequencing chemistries, an approach which not only helps with automation but also has led to improved results with capillary and thin slab gel electrophoresis systems.
The Sequatron Systems. The initial Sequatron system consisted of a CRS A255 arm running along a 4M track to service a number of devices such as XYZ robots, thermal cyclers and plate washers. This Sequatron was in full production at the Whitehead Institute Genome Center for one year and purified and sequenced over 400,000 M13 and PCR products. This system was then taken out of commission in favor of a smaller, modular system. The Sequatron II systems consisted of a CRS A455 arm on a 5 foot x 5 foot table3. Two systems were put into full production, one for DNA purification from M13 and PCR products, and the other for the set-up, thermal cycling and pooling of DNA sequencing reactions. Both systems had throughput of over 16,000 samples per 24-hour period.
Next Generation Systems. The latest work has centered on the development of integrated systems based on nano-liter reaction volumes. Working with the Packard BioChip Processor, we have shown feasibility of using piezo electric methods for small-scale PCR and Sequencing reactions. These reactions can then be automatically loaded into the CuraGen Niagara DNA sequencer or onto a MALDI-TOF Mass Spec system. This new system is currently being designed.
This work was funded by DOE DE-FG02-95ER62099 to TLH.
Stefan Burde and Babetta L. Marrone
Comparative genomic hybridization (CGH) has become a valuable technique for identification of gross changes in DNA copy number (duplications or deletions) in solid tumor samples. Using current data analysis methods, red and green fluorescence profiles are obtained for all chromosomes, and green:red ratio profiles are computed from these. An increase or decrease in this ratio is scored as a duplication or deletion respectively.
Here, we present an alternative approach to CGH data analysis. By converting digital images to Flow Cytometry Standard (FCS) file format, it was possible to take advantage of the wide variety of data analysis methods which have been developed for flow cytometry data. By generating pixel-by-pixel bivariate displays of image data, populations of pixels containing increased or decreased green fluorescence could be readily identified. By gating on these subpopulations, a reprocessed image was generated, showing only those areas of the chromosomes where aberrations are present. Duplications were false-colored in green and deletions in red. It was thus only necessary to analyze those chromosomes showing red or green areas on the reprocessed image. By examining these chromosomes using second-derivative analysis of the DAPI intensity profile, it was possible to assess the cytogenetic band locations of these aberrations. This approach can be easily implemented in laboratories performing CGH analysis, and represents a less computation-intensive and potentially more rapid method for screening of chromosome aberrations.