US 20060252061 A1
The present invention is in the field of nucleic acid-based diagnostic assays. More particularly, it relates to methods useful for the “diagnostic sequencing” of regions of sample nucleic acids for which a prototypic or reference sequence is already available (also referred to as “re-sequencing”), or which may be determined using the methods described herein. This diagnostic technology is useful in areas that require such re-sequencing in a rapid and reliable way: (i) the identification of the various allelic sequences of a certain region/gene, (ii) the scoring of disease-associated mutations. (iii) the detection of somatic variations, (iv) studies in the field of molecular evolution, (v) the determination of the nucleic acid sequences of prokaryotic and eukaryotic genomes; (vi) identifying one or more nucleic acids in one or more biological samples; (vii) and determining the expression profile of genes in a biological sample and other areas.
1. A method for sequence analysis of one or more target nucleic acids for which a reference nucleic acid sequence is known, the method comprising the steps of:
(a) deriving from one or more biological samples the one or more target nucleic acids;
(b) subjecting the one or more target nucleic acid to complementary cleavage reactions with one or more cleavage reagents thereby generating cleavage products;
(c) performing mass spectroscopical analysis on the cleavage products generated in step (b) to obtain one or more mass spectra; and
(d) comparing the one or more mass spectra of the cleavage products of the one or more target nucleic acids obtained in step c) with a known or predicted mass spectra for the reference nucleic acid sequence, and deducing therefrom, by systematic computational analysis, all or part of the nucleotide sequence of the one or more target nucleic acids, and comparing the deduced nucleic acid sequence with the reference nucleic acid to determine whether the one or more target nucleic acids have the same sequence or a different sequence from the reference nucleic acid.
2. The method of
3. The method of
4. The method of
5. The method of
6. The method of
(a) linking operatively an expression control sequences to the one or more target nucleic acids; and
(b) transcribing one or both strands of the one or more target nucleic acid of step a) using one or more RNA polymerases that recognize the transcription control sequence on the one or more target nucleic acids.
7. The method of
8. The method of
9. The method of
10. The method of claims 4, 5, 6, 7, 8 or 9 wherein the target nucleic acid is derived using one or more modified nucleoside triphosphates selected from the group consisting of mass modified deoxynucleoside triphosphates, mass modified dideoxynucleoside triphosphates, and mass modified ribonucleoside triphosphates.
11. The method of
12. The method of
13. The method of
14. The method of
15. The method of
16. The method of claims 11, 12, 13 or 14 wherein the modifications alter the specificity of cleavage by the one or more cleavage reagents and/or the mass and/or the length of the cleavage products.
17. The method of
18. The method of
19. The method of
20. The method of
21. The method of
22. The method of
23. The method of
24. The method of
25. The method of
26. A method for scoring known nucleotide sequence variations of one or more target nucleic acids for which a known reference nucleic acid sequence is available, the method comprising the steps of:
(a) deriving from one or more biological samples the one or more target nucleic acids;
(b) subjecting the one or more target nucleic acids to complementary cleavage reactions with one or more cleavage reagents thereby generating cleavage products;
(c) performing mass spectroscopical analysis on the cleavage products obtained in step (b) to obtain two or more mass spectra; and
(d) comparing the one or more mass spectra of the cleavage products of the one or more target nucleic acids obtained in step c) with the known or predicted mass spectra for the reference nucleic acid sequence, and scoring therefrom, by systematic computational analysis the presence or absence of the known nucleotide sequence variations of the one or more target nucleic acids.
27. The method of
28. The method of
29. The method of
30. The method of
31. The method of
(a) linking operatively a transcription control sequences to the one or more target nucleic acids; and
(b) transcribing one or both strands of the one or more target nucleic acids using one or more RNA polymerases that recognize the transcription control sequence on the one or more target nucleic acids.
32. The method of
33. The method of
34. The method of
35. The method of
36. The method of
37. The method of
38. The method of
39. The method of
40. The method of
41. The method of
42. The method of
43. The method of
44. The method of
45. The method of
46. The method of
47. The method of
48. The method of
49. The method of claim or claim wherein the one or more target nucleic acids are phosphorothioate- modified single-stranded DNA or RNA and the endonuclease is nuclease P1.
50. The method of
51. A method for determining the sequence of one or more target nucleic acids, the method comprising the steps of:
(a) deriving from a biological sample material the one or more target nucleic acids;
(b) subjecting the one or more target nucleic acids to complementary cleavage reactions with one or more cleavage reagents thereby generating cleavage products;
(c) performing mass spectroscopical analysis on the cleavage products obtained in step (b) to obtain a mass spectra of the nucleic acid fragments; and
(d) deducing therefrom the nucleotide sequence of the one or more target nucleic acids by systematic computational analysis on the mass spectra of the cleavage products obtained in step (c).
52. A method for genome wide genotyping of one or more known or unknown target nucleic acids, the method comprising the steps of:
(a) deriving from one or more biological samples the one or more target nucleic acids;
(b) subjecting the one or more target nucleic acids to one or more specific cleavage reactions with one or more cleavage reagents thereby generating cleavage products;
(c) performing mass spectroscopical analysis on the cleavage products obtained in step (b) to obtain one or more mass spectra; and
(d) comparing the mass spectra of the cleavage products obtained in step c) with reference mass spectra, and diagnosing therefrom genetically relevant nucleic acid sequence variations of the one or more known or unknown target nucleic acids.
53. A method for identifying one or more target nucleic acids in one or more biological samples, the method comprising the steps of:
(a) deriving from one or more biological samples the one or more target nucleic acids;
(b) subjecting the one or more target nucleic acids of step (a) to one or more specific cleavage reactions with one or more cleavage reagents thereby generating cleavage products;
(c) performing mass spectroscopical analysis on the cleavage products obtained in step (d) to obtain one or more mass spectra; and
(d) comparing the one or more mass spectra obtained in step c) with each other or with a plurality of mass spectra of reference nucleic acids, and deducing therefrom the identity of the one or more target nucleic acids.
54. The method of
55. The method of
56. A kit for sequence analysis of one or more target nucleic acids using mass spectroscopy, the kit comprising:
(a) one or more sets of reference nucleic acids for which the nucleic acid sequence is known;
(b) one or more nucleotide triphosphates;
(c) one or more polymerases;
(d) one or more nucleic acid cleaving agents; and
(e) computer software for comparing the mass spectra of the one or more target nucleic acid with the mass spectra of the reference nucleic acid and deducing therefrom the nucleic acid sequence of the target nucleic acid.
The present invention is in the field of nucleic acid-based diagnostic assays. More particularly, it relates to methods useful for the “diagnostic sequencing” of regions of sample nucleic acids for which a prototypic or reference sequence is already available (also referred to as ‘re-sequencing’), or which may be determined using the methods described herein. This diagnostic technology is useful in areas that require such re-sequencing in a rapid and reliable way: (i) the identification of the various allelic sequences of a certain region/gene, (ii) the scoring of disease-associated mutations, (iii) the detection of somatic variations, (iv) studies in the field of molecular evolution, (v) the determination of the nucleic acid sequences of prokaryotic and eukaryotic genomes; (vi) identifying one or more nucleic acids in one or more biological samples; (vii) and determining the expression profile of genes in a biological sample and other areas.
Complete reference genome sequences for a number of model organisms as well as humans are currently available or are expected to become available in the near future. A parallel challenge is to characterize the type and extent of variation in the sequences of interest because it underlies the heritable differences among individuals and populations. In humans, the vast majority of sequence variation consists of nucleotide substitutions referred to as single nucleotide polymorphisms (SNPs). DNA sequencing is the most sensitive method to discover polymorphisms [Eng C. and Vijg J. et al., Nature Biotechnol. 15: 422-426 (1997)]. A growing panel of such sequence variants, together with powerful methods to monitor them [Landegren U. et al. Genome Res. 8: 769-776 (1998)], is useful in linkage studies to identify even the most subtle disease susceptibility loci [Lander E. and Schork N., Science 265: 2037-2048 (1994); Risch N. and Merikangas K., Science 273: 1516-1517 (1996)]. Also, the identification of all (functional) allelic variants will require the re-sequencing of particular regions in a large number of samples [Nickerson D. et al., Nature Genet. 19: 233-240 (1998)]. Although a number of methods to monitor known SNPs have been developed [Landegren U. et al., Genome Res. 8: 769-776 (1998)], re-sequencing is likely to be routinely applied to secure diagnoses of patients. Indeed, in a significant number of disease-associated genes that have been surveyed thus far, literally hundreds or even thousands of different mutations have been identified and catalogued. Consequently, sequence determination represents the ultimate level of resolution and may be the preferred method to monitor which mutation or combination of mutations, out of a large number of mutations of known clinical relevance, is present.
It would appear that the field of biomedical genetics will rely heavily on sequencing technology. Hence, there is a need for advanced sequencing methods that are time- and cost-competitive, and at the same time accurate and robust. Recent developments in this area include improvements to the basic dideoxy chain termination sequencing method [Sanger et al. Proc. Natl. Acad. Sci. USA 74: 5463-5467 (1977); reviewed by Lipshutz R and Fodor S. et al., Current Opinion in Structural Biology 4: 376-380 (1994)], as well as new approaches that are based on entirely new paradigms. Two such novel approaches are sequencing-by-hybridization (SBH) [Drmanac R. et al., Science 260: 1649-1652 (1993)] and pyro-sequencing [Ronaghi M. et al., Science 281: 363-365 (1998); Ronaghi M. et al., Anal. Biochem. 242: 84-89 (1996)]. While the concepts of these approaches have been experimentally validated, their ultimate acceptance and usage may depend on the type of application—e.g. de novo sequencing, re-sequencing, and genotyping of known SNPs.
Recently, progress has also been made in the use of mass spectroscopy (MS) to analyze nucleic acids [Crain, P. F. and McCloskey, J. A., Current Opinion in Biotechnology 9: 25-34 (1998), and references cited therein]. One promising development has been the application of MS to the sequence determination of DNA and RNA oligonucleotides [Limbach P., Mass Spectrom. Rev. 15: 297-336 (1996); Murray K., J. Mass Spectrom. 31: 1203-1215 (1996)]. MS and more particularly, matrix-assisted laser desorption/ionization MS (MALDI MS) has the potential of very high throughput due to high-speed signal acquisition and automated analysis off solid surfaces. It has been pointed out that MS, in addition to saving time, measures an intrinsic property of the molecules, and therefore yields a significantly more informative signal [Köster H. et al., Nature Biotechnol., 14: 1123-1128 (1996)].
Sequence information can be derived directly from gas-phase fragmentation [see for example Nordhoff E. et al., J. Mass Spectrom., 30:99-112 (1995); Little D. et al., J. A. Chem. Soc., 116: 4893-4897 (1994); Wang B. et al., WO 98/03684 and WO 98/40520; Blöcker H. et al., EP 0 103 677; Foote S. et al., WO 98/54571]. In contrast, indirect methods measure the mass of fragments obtained by a variety of methods in the solution phase, i.e., prior to the generation of gas phase ions. In its simplest form, mass analysis replaces the gel-electrophoretic fractionation of the fragment-ladder (i.e., a nested set of fragments that share one common endpoint) generated by the sequencing reactions. The sequencing reactions need not necessarily be base-specific because the base-calling may also be based on accurate mass measurement of fragments that terminate at successive positions and that differ from one another by one nucleotide residue. The fragment-ladder can be generated by the Sanger method [Köster H. et at, Nature Biotechnol., 14: 1123-1128 (1996); Reeve M. A., Howe R. P., Schwarz T., U.S. Pat. No. 5,849,542; Köster H., U.S. Pat. No. 5,547,835; Levis R. and Romano L., U.S. Pat. No. 5,210,412 and U.S. Pat. No. 5,580,733; Chait B. and Beavis R., U.S. Pat. No. 5,453,247], by base-specific partial RNA digestion [Hahner S. et al., Nucleic Acids Res., 25: 1957-1964 (1997); Köster H., WO 98/20166] or by chemical cleavage [Isola N. et al., Anal. Chem., 71: 2266-2269 (1999); references cited in Limbach P., Mass Spectrom. Rev., 15: 297-336 (1996)]. An alternative method consists of analyzing the ladder generated by exonuclease digestion from either the 3′- or 5′-end [Pieles U. et al., Nucleic Acids Res., 21: 3191-3196 (1993); Köster H., U.S. Pat. No. 5,851,765; Engels J. et al., WO 98/45700; Tarr G. and Patterson D., WO 96/36986; Patterson D., U.S. Pat. No. 5,869,240].
A severe limitation of both the direct and indirect MS methodologies under the current performance conditions is the poor applicability to chain lengths beyond ˜30-50 nucleotides. As a consequence, it has been suggested that the prospects for MS lie with DNA diagnostic assays, rather than large-scale sequencing [Smith L., Nature Biotechnol., 14: 1084-1087 (1996)]. Given the fact that MS represents an exquisite means to analyze short nucleotide fragments, the various MS-based processes that have been described for nucleic acid based diagnostic purposes generally involve the derivation and analysis of such relatively short fragments [see for example Köster H., WO 96/29431; Köster H. et al., WO 98/20166; Shaler T. et al., WO 98/12355; Kamb A., U.S. Pat. No. 5,869,242; Monforte J. et al., WO 97/33000; Foote S. et al., WO 98/54571].
Some of the MS-based assays have been used for the scoring of defined mutations or polymorphisms. Other processes derive multiple oligonucleotide fragments and yield a ‘mass-fingerprint’ so as to analyze a larger target nucleic acid region for mutations and/or polymorphisms. The latter MS analyses are however considerably less informative in that they are essentially restricted to the detection of sequence variations. The methods cannot be applied to diagnostic sequencing of nucleic acids, where the term diagnostic sequencing means the unequivocal determination of the presence, the nature and the position of sequence variations. At best, the measurements confirm the base composition of small fragments whose masses are determined with sufficient accuracy to reduce the number of possible compositional isomers. Also, it will be realized that only certain changes in composition (as revealed by shifts in the mass spectrum) can be unambiguously assigned to a polymorphism or mutation. A match between the spectrum of the interrogated sequence and a reference-spectrum obtained from wild-type sequence or sequences known to contain a given polymorphism, is assumed to indicate that the interrogated nucleic acid region is wild-type or incorporates the previously known polymorphisms, thereby disregarding certain other possible interpretations.
While most methods in the art do yield sequence-related information, they do not disclose that a combination of several different mass spectra, obtained after complementary digestion reactions, allows for the effective survey of a nucleic acid region and provides an unambiguous assignment of both known as well as previously unknown sequence variations that occur relative to a reference nucleic acid with a known nucleotide sequence.
In view of the limitations of the methods described above, the art would clearly benefit from a new procedure for the diagnostic sequencing of nucleic acids that would overcome the shortcomings of the processes discussed above.
In comparison with conventional sequencing technology, i.e., the gel-electrophoretic analysis of fragment ladders, the methods of the present invention are more suited for the simultaneous analysis of multiple target sequences. In general, each particular sequence or sequence variant is associated with a distinct set of mass peaks. Consequently, the sequencing reactions according to the methods of the present invention lend themselves readily to (i) multiplexing (i.e., the analysis of two or more target non-contiguous target regions from a single biological sample), (ii) the analysis of heterozygous samples, as well as (iii) pooling strategies (i.e., the simultaneous sequencing of the analogous regions derived from two or more different biological samples).
Because of the multiplex capacity, the present methods can be adapted as a tool for the genome-wide discovery and scoring of polymorphisms (e.g., SNPs) useful as markers in genetic linkage studies. The unambiguous identification/diagnosing of a number of variant positions is less demanding than full sequencing and, consequently, a considerable number of target genomic loci can be combined and analyzed at the same time, especially when their lengths are kept relatively small. The number of markers that can be scored in parallel will depend on the level of genetic diversity in the species of interest and on the precise method used to prepare and analyze the target nucleic acids, but may typically be in the order of a few tens to up to 100 with current MS capabilities. The addition of multiplexing to the high-precision and high-speed characteristics of MS constitutes a new marker technology that enables the large-scale and cost-effective scoring of several (tens of) thousands of markers. Some aspects of the application of the present methods to genome-wide genotyping are described in Example 5.
Sequencing reactions according to the methods of the present invention yield, in principle, a discrete set of fragments for each individual sequence or sequence variant whereas conventional sequence ladders stack on top of one another. Therefore, such sequences or sequence variants can be analyzed even when present as a lesser species. This is a useful quality for the analysis of clinical samples which are often genetically heterogeneous because of the presence of both normal and diseased cells or in itself (e.g., cancerous tissue, viral quasi-species). Additionally, the ability to detect mutations at a low ratio of mutant over wild-type allele makes it practicable to pool individual biological samples, a strategy which should permit a more cost-effective search for genomic sequence variations in a population.
The present invention rests in part on the insight that integration of the data obtained in a set of complementary fingerprints produced by an appropriate set of complementary cleavage reactions of the invention represents a level of characterization of a sample nucleic acid essentially equal to sequence determination. The present invention is also directed to the use of cleavage protocols that result in the generation of cleavage products that range from mono- and dinucleotides to fragments of a few tens of nucleotides that are particularly suited for analysis by MS. At the same time, the present method is distinct from the other fragmentation processes that are limited to screening target nucleic acids for a wide range of potential mutations. According to the present invention, a combination of several different mass spectra, obtained after complementary digestion reactions, coupled with systematic computational analysis allows the survey of a selected target nucleic acid or region thereof and leads to the unambiguous assignment of both known and previously unknown sequence variations. In certain aspects of the present invention, knowledge of the reference sequence in combination with the methods disclosed herein allows modeling of the experimental approach, anticipation of potential ambiguities, and the design of an adequate resolution.
The present invention is directed to a mass spectroscopic method for detecting or analyzing a particular nucleic acid sequence. The present invention is useful for de novo sequencing or re-sequencing nucleic acid in a rapid and reliable way which permits, for example, the identification of the various allelic sequences of a certain region/gene, the identification and scoring of disease-associated mutations, the detection of somatic variations, determining genetic diversity in molecular evolution, and the determination of the genomic sequences e.g., of viral and bacterial isolates. The present invention is also useful for identification of all nucleic acid molecules in one or more biological samples including for expression profiling i.e., identification of all the mRNA species that are expressed in a given cell at a given time, by rapidly determining the sequence of the mRNA that is expressed.
In one embodiment, the present invention is directed to methods for sequence analysis of one or more target nucleic acids for which a known reference nucleic acid sequence is available. In this method, one or more target nucleic acids are derived from one or more biological samples, and a reference nucleic acid are each subjected to complementary cleavage reactions, and the products of the cleavage reactions are analyzed by mass spectroscopic methods. The mass spectra of the one or more target nucleic acids are then compared with the mass spectra of the reference nucleic acid sequence, and the nucleotide sequence of the one or more target nucleic acids is deduced by systematic computational analysis.
In one aspect of this embodiment, multiple targets, such as cDNA clones, are prepared from the mRNA of the same biological sample, and are separately analyzed as above in parallel experiments. In a second aspect, multiple targets are derived from the same biological sample and are analyzed simultaneously, for example in genome-wide genotyping.
The one or more target nucleic acids may be selected from the group consisting of a single stranded DNA, a double stranded DNA, a cDNA, a single stranded RNA, a double stranded RNA, a DNA/RNA hybrid, and a DNA/RNA mosaic nucleic acid.
In a second embodiment, the one or more target nucleic acids are selected from the group consisting of an amplified nucleic acid fragment, a cloned nucleic acid fragment, and a series of non-contiguous DNA fragments from the genome. In one aspect of this invention, the amplified one or more target nucleic acids are derived by one or more consecutive amplification procedures selected from the group consisting of in vivo cloning, the polymerase chain reaction (PCR), reverse transcription followed by the polymerase chain reaction (RT-PCR), strand displacement amplification (SDA), and transcription based processes.
In a preferred embodiment, the amplified nucleic acid fragments are RNA transcripts generated from one or more target nucleic acids or a reference nucleic acid by a process comprising the steps of: (a) amplifying the one or more target nucleic acids or the reference nucleic acid using one or more primers corresponding to a region that is complementary to the one or more target nucleic acids or the reference nucleic acid and encoding an expression control sequence using any one of the amplification procedures described above, and (b) generating RNA Transcripts from the amplified one or more target nucleic acids or reference nucleic acid using one or more RNA polymerases that recognize the transcription control sequence on the target or reference nucleic acid. The RNA generated by the above process is then subjected to complementary cleavage reactions to generate nucleic acid fragments, which are then analyzed by MS. The transcription control sequence may be selected from the group consisting of an eukaryotic transcription control sequence, a prokaryotic transcription control sequence, and a viral transcription control sequence. The prokaryotic transcription control sequence may be selected from the group consisting of T3, T7, and SP6 promoters. The cognate RNA polymerases may be either a wild-type or a mutant form capable of incorporating non-canonical substrates with a 2′-substituent other than a hydroxyl group.
In a third embodiment, the one or more target nucleic acids are amplified using modified nucleoside triphosphates. The mass modified nucleoside triphosphates may be selected from the group consisting of a mass modified deoxynucleoside triphosphate, a mass modified dideoxynucleoside triphosphate, and a mass modified ribonucleoside triphosphate. The mass modified nucleoside triphosphate may be modified on the base, the sugar, and/or the phosphate moiety, and are introduced through an enzymatic step, chemically, or a combination of both. In one aspect the modification may consist of 2′-substituents other than a hydroxyl group on transcript subunits. In another aspect, the modification may consist of phosphorothioate internucleoside linkages or phosphorothioate internucleoside linkages further reacted with an alkylating reagent. In yet another aspect, the modification may consist of a methyl group on C5 of the uridine-5′-monophosphate subunits. Such modifications may alter the specificity of cleavage by certain reagents, and/or the mass of the cleavage products, and/or the length of the cleavage products.
In one aspect of the invention, the one or more target nucleic acids and reference nucleic acid are subjected to complementary cleavage reactions using enzymatic cleavage, chemical cleavage, and/or physical cleavage reactions. In a preferred embodiment, the one or more target nucleic acids and the reference nucleic acid are subjected to enzymatic cleavage reaction using one or more enzymes selected from the group consisting of endonucleases and exonucleases. In a more preferred embodiment, the target nucleic acid is a double-stranded RNA and the endonuclease used is a ribonuclease. The ribonuclease may be selected the G-specific T1 ribonuclease, the A-specific U2 ribonuclease, the A/U specific phyM ribonuclease, the U/C specific ribonuclease A, the C-specific chicken liver ribonuclease (RNaseCL3), and cusativin. In one aspect of this preferred embodiment, the target nucleic acid is a phosphorothioate-modified single-stranded DNA or RNA and the endonuclease is nuclease P1.
In another aspect, the mass spectroscopical analysis of the nucleic acid fragments is performed using a mass spectrometer selected from the group consisting of Matrix-Assisted Laser Desorption/Ionization-Time-of-flight (MALDI-TOF), Electrospray-Ionization (ESI), and Fourier Transform-Ion Cyclotron Resonance (FT-ICR). In a preferred embodiment the mass spectrometer used for the analysis of the cleavage fragments is MALDI-TOF
In a fifth embodiment, the method of the present invention can be used for diagnosing nucleic acid sequence variations in one or more target nucleic acids derived from a biological sample, for which a known reference nucleic acid sequence is available. In this method, one or more target nucleic acids derived from a biological sample, and a reference nucleic acid whose sequence has been predetermined are subjected to complementary cleavage reactions, and the products of the cleavage reactions are analyzed by mass spectroscopic methods. The mass spectra of the one or more target nucleic acids is then compared with the mass spectra of the reference nucleic acid, and the nucleotide sequence variations in the one or more target nucleic acids is then deduced by systematic computational analysis of the sequence variations between the one or more target nucleic acids and the reference nucleic acid. A variety of acid sequence variations including deletions, substitutions and/or insertions in a target nucleic acid can be determined using the method of the present invention.
In a sixth embodiment, the method of the present invention can be used for scoring known nucleotide sequence variations in one or more target nucleic acids derived from a biological sample, for which a known reference nucleic acid sequence is available. In this embodiment, one or more target nucleic acids derived from a biological sample, and a reference nucleic acid are subjected to complementary cleavage reactions, and the products of the cleavage reactions are analyzed by mass spectroscopic methods. The mass spectra of the one or more target nucleic acid is then compared with the mass spectra of the reference nucleic acid sequence, and the nucleotide sequence variations/mutations in the one or more target nucleic acids are scored by comparing the nucleic sequence between the one or more target nucleic acid and reference nucleic acid by systematic computational analysis.
In a seventh embodiment, the method of the present invention can be used for determining the nucleotide sequence (de novo sequencing) of one or more target nucleic acids derived from a biological sample for which no reference sequence is available. In this method, target nucleic acid, derived from a biological sample is subjected to complementary cleavage reactions, and the products of the cleavage reactions are analyzed by mass spectroscopic methods. The mass spectra of the one or more target nucleic acids coupled with a systematic computational analysis is then used to deduce the sequence of the one or more target nucleic acids.
In an eighth embodiment, the method of the present invention can be used for genome-wide genotyping of one or more known or unknown target nucleic acids. In this method, one or more target nucleic acids, derived from a biological sample, are amplified and then subjected to complementary cleavage reactions. In one aspect, multiple targets are derived from a single sample and are analyzed simultaneously. The products of the cleavage reactions are then analyzed by mass spectroscopic methods. The mass spectra of the one or more known or unknown target nucleic acid is compared with the mass spectra of a reference nucleic acid. This comparison is then used to infer the genotype of an organism from which the biological sample is derived and to determine therefrom the genetically relevant nucleic acid sequence variations of the one or more known or unknown nucleic acids.
In a ninth embodiment, the method of the present invention can be used to identify one or more target nucleic acids in one or more biological samples. In this method, one or more target nucleic acids, derived from a biological sample, are amplified and then subjected to complementary cleavage reactions. In one aspect, multiple targets are derived from a single sample and are analyzed simultaneously. The products of the cleavage reactions are then analyzed by mass spectroscopic methods. The identity of one or more target nucleic acids is deduced by comparing the mass spectra of the one or more known or unknown target nucleic acid with each other or by comparison with a plurality of mass spectra of reference nucleic acids.
In one aspect, the method of the present invention can be used for expression profiling, i.e. identifying the various mRNA expressed in one or more biological samples.
Also encompassed by the present invention is a kit for sequence analysis of one or more target nucleic acids using mass spectroscopy, the kit comprising a container having one or more sets of reference nucleic acids for which the nucleotide sequence is known, one or more nucleic acid cleaving agents, and computer algorithm/software for comparing the mass spectra of the one or more target nucleic acids with the mass spectra of the reference nucleic acid and deducing therefrom the nucleic acid sequence of the one or more target nucleic acids. In one embodiment, the nucleic acid cleaving agent in the kit is a chemical agent. In an alternate embodiment, the nucleic acid cleaving agent is an enzyme selected from a group of enzymes consisting of endonucleases and exonucleases. In a preferred embodiment, the endonuclease is a ribonuclease selected from the group consisting of the G-specific T1 ribonuclease, the A-specific U2 ribonuclease, the A/U specific phyM ribonuclease, the U/C specific ribonuclease A, the C-specific chicken liver ribonuclease (RNaseCL3), and cusativin.
With current capabilities in mass spectroscopy, it is impractical to sequence nucleic acids greater than ˜50 bases in length. Consequently, an impractical and cumbersome number of independent sequencing reactions is necessary to cover the thousands of bases of a gene or other genetic region of interest. The methods of the present invention described below overcome this limitation. At the same time, the present method is distinct from the other fragmentation processes that are limited to screening target nucleic acids for a wide range of potential mutations. Indeed, the appropriate choice of complementary cleavage reactions as described herein allows the determination of the exact location and nature of a genetic variation. Also, it is demonstrated herein that computational protocols are an integral part of the described method. The methods and algorithms are required to deduce, on the basis of the reference sequence(s), the relation between (i) the spectral changes associated with one or more cleavage reactions of a given nature, and (ii) the uniquely defined sequence variations.
Sequencing reactions according to the methods of the present invention can be multiplexed, i.e. used for the simultaneous analysis of multiple non-contiguous target regions [supra]. Therefore, the methods can be adapted as a tool for the genome-wide discovery and/or scoring of polymorphisms (e.g. SNPs) useful as markers in genetic linkage studies. Indeed, it will be recognized that the unambiguous identification/diagnosing of a number of variant positions is less demanding than full sequencing and that consequently a considerable number of target genomic loci can be combined and analyzed in parallel, especially when their lengths are kept relatively small. The number of markers that can be scored in parallel will depend on the level of genetic diversity in the species of interest and on the precise method used to prepare and analyze the target nucleic acids, but may typically be in the order of a few tens to up to 100 or more with current MS capabilities. The addition of multiplexing to the high-precision and high-speed characteristics of MS constitutes a new marker technology that enables the large-scale and cost-effective scoring of several (tens of) thousands of markers. Some aspects of the application of the present methods to genome-wide genotyping are described in Example 5.
The present invention provides a mass spectroscopy (MS) based nucleic acid sequencing method that overcomes some of the drawbacks inherent in the prior art. In contrast to the previously described methods, the methods of the present invention do not require the generation of a ladder, i.e. an ordered set of nested nucleic acid fragments characterized by a common end. Rather, the disclosed methods rely on a combination of complementary fragmentation reactions and the analytical resolution power of MS to improve mass resolution and mass accuracy. The present invention is directed to the use of enzymatic cleavage protocols that result in the generation of cleavage products that range from mono- and dinucleotides to fragments of a few tens of nucleotides that are particularly suited for analysis by MS. According to the present invention, a combination of several different mass spectra, obtained after complementary digestion reactions, coupled with systematic computational analysis allows the survey of a selected nucleic acid or region thereof and leads to the unambiguous assignment of both known and previously unknown sequence variations.
The present invention is also directed to methods for the diagnostic sequencing (also referred to as re-sequencing) of all or part of a sample nucleic acid, i.e. the determination of the presence, the nature and the location of the sequence variations that occur relative to a related known reference sequence. The sequence variations may either be previously identified or hitherto unknown. Diagnostic sequencing according to the present invention may focus on particular positions in a nucleic acid sequence, e.g. when scoring previously known mutations or polymorphisms.
The term “mapping”, as used herein, will be understood to include both the characterization, i.e. determination of the nature, and the position of the sequence variations.
The terms “target DNA”, “target sequence”, “target nucleic acid” and the like, as used herein, refer to the sequence region which is to be sequenced or re-sequenced entirely or in part as well as to the nucleic acid material that is actually subjected to one or more complementary cleavage reactions.
The terms “reference nucleic acid sequence”, “related sequence”, “previously known sequence”, and the like, refer to a nucleic acid region, the sequence of which has previously been determined which corresponds to the target. The reference and target sequences may be found to be identical or may differ. The reference sequence need not derive from the same species. In many applications, several different sequence variants will be available as reference. The differences between a target sequence and its reference sequence may be simple (e.g., single nucleotide substitutions, deletions and insertions; microsatellite polymorphisms) or complex (e.g., substitution, insertion, and deletion of multiple nucleotides). In certain situations, one may not know in advance to what reference sequence, if any, the target nucleic acid corresponds. In such situations the interrogated target sequence typically corresponds to a portion of a (much) larger reference sequence and/or to one out of a plurality of different references.
The terms “unambiguous”, “unique”, “unequivocal”, and the like, are used to indicate that only a single sequence variation or combination of sequence variations can explain the observed mass spectral changes.
The terms “complementary (cleavage) reactions”, “complementary cleavages” and the like, as used herein, refer to target nucleic acid digestions characterized by varying specificity [e.g., stringent or relaxed mono- and di-nucleotide specificity; digestion with a combination of reagents; partial cleavage] and/or to digestion alternative forms of the target sequence [e.g., the complementary (+) and (−) strands; incorporation of modified subunits; analysis of variable portions of the target sequence].
The terms “transcript” and “transcription”, as used herein, refer to the synthesis of a nucleic acid polymer by means of an RNA polymerase. In addition to canonical subunits (having a 2′-OH group), a transcript may incorporate non-canonical substrates (having any other substituent than a hydroxyl group at the 2′-position). Canonical and non-canonical substrates may contain additional modifications.
The term “genotyping,” as used herein, refers to determining the genetic constitution, which is the particular set of alleles inherited by the organism as a whole, or the type of allele found at a particular locus of interest.
The term “expression profiling,” as used herein, refers to method(s) for determining the mRNA expression profile of a given cell or a population of cells at a given time under a given set of conditions.
Nucleotides are designated as follows. A ribonucleoside triphosphate is referred to as NTP or rNTP; N can be A, G, C, U or m5U to denote specific ribonucleotides. Likewise, deoxynucleoside triphosphate substrates are indicated as dNTPs, where N can be A, G, C, T, or U. Throughout the text, monomeric nucleotide subunits are denoted as A, G, C, or T with no particular reference to DNA or RNA. When necessary, the nature of the nucleoside monophosphates is clarified by the use of more specific abbreviations such as U, m5U, CMP, and UMP to refer to ribonucleotides and dC, dU, dCMP, dUMP and dTMP to indicate deoxynucleotides. Note that T is not an alternative designation for m5U.
Sequencing via Non-Ordered Sets of Specific Cleavage Fragments
The methods of the present invention allow the interrogation every position in a given target sequence without creating a fragment-ladder, i.e. a nested set of fragments that share one common endpoint. The method comprises, in part, subjecting one or more target nucleic acids to a set of complementary mononucleotide- and/or dinucleotide-specific cleavages, the products of which are analyzed by mass spectroscopy (MS). A preferred method according to the invention includes the specific cleavage of the one or more target nucleic acids at each nucleotide by way of two or more separate reactions. The digestion products obtained in mononucleotide- and dinucleotide-specific cleavage reactions such as those described herein range from mononucleotides to fragments of a few tens of nucleotides and are particularly well suited for analysis by MS. This aspect of the invention overcomes the technical limitation of the short read lengths encountered when analyzing fragment-ladders under the current MS performance. The mass spectra obtained with the methods do not provide a simple readout of the sequence. Computational approaches provided herein allow the comparative analysis of the obtained spectra with those known or predicted for the related reference sequence.
The ability to detect and map sequence variants based on the non-ordered set of cleavage fragments according to the present invention resides in part in the combination of the various complementary site-specific reactions. For example, one cleavage scheme useful in the practice of the present invention makes use of the mononucleotide-specific ribonuclease-T1 (RNase-T1, G-specific) and RNase-U2 (A-specific; the limited specificity of this enzyme is recognized and will be dealt with below). Both purines (A/G) and pyrimidines (C/T) in a target nucleic acid can be examined by cleaving an RNA copy of the two complementary strands of a target nucleic acid with both enzymes. MS analysis of the fragments generated by only a single mononucleotide-specific reaction would detect the presence of most sequence variations but only a minority of the mutations—in essence those affecting the nucleotide that is recognized—would also be localized. Since the methods of the present invention examine each of the four bases in a given sequence, each of the twelve possible nucleotide substitutions result in the loss of one cleavage site and the concomitant gain of another cleavage site. This principle is illustrated in Table I for the RNase-T1 and RNase-U2 cleavage reactions on the two complementary transcripts of a hypothetical target nucleic acid. Transitions affect both the RNase-T1 and RNase-U2 cleavage patterns of either the (+) or the (−) strand. As can be seen in Table 1, all transversions change the cleavage pattern of both strands of the transcript: they affect either one of the RNase digests on both strands, or the T1 digest of one strand and the U2 digest of the complementary strand. In addition to altering two cleavage patterns, each single nucleotide substitution also affects the molecular mass of one fragment in each of the remaining two digestion reactions (Table I). In conclusion, complementary cleavage reactions of the present invention results in a high degree of built-in redundancy. Each nucleotide substitution is potentially associated with a maximum of ten differences (data points) with respect to the reference spectrum. The loss and gain of a cleavage site are associated with both the disappearance and appearance of three peaks; two additional peaks undergo a shift as a result of a mass difference. In practice, the 1 Da mass difference between C and U(T) may result in the loss of a significant amount of information (Table I). More particularly, in G- and A-specific cleavage reactions, the C/U transitions may go unnoticed while the observed mass difference may not be unambiguously assigned to a certain transversion. However, in preferred methods of the present invention directed to the analysis of RNA target sequences the method makes use of C and/or U analogs that exhibit more favorable mass differences, thus allowing the unambiguous assignment of the mass difference to a particular transversion. Example 1 and Table I illustrate that 5-methyluridine is an example of such a useful analog [m5U; R.I. Chemical, Orange, Calif.; see also to Hacia J. et al., Nucleic Acids Res. 26: 4975-4982 (1998) for the incorporation of m5UTP during in vitro transcription reactions].
The methods of the present invention are therefore largely, yet not completely, sequence-independent and permits the re-sequencing of virtually any variation. Computer simulations of diagnostic sequencing by the present methods, more particularly those involving digestion of RNA copies of each strand with the RNases T1 and U2, have shown that for target sequences of up to three hundred base-pairs˜90% or more of all possible single nucleotide substitutions are associated with ≧4 data points. Fewer than 1% of the substitutions do not result in spectral changes. More than 95% of all possible single nucleotide substitutions give rise to unique spectral changes and can therefore be unambiguously identified (see Example 1 and
In summary, deduction of the sequence according to the methods of the present invention is based on the integration of the information that resides in a complementary set of ‘mass-fingerprints’ as well as the previous knowledge about a related reference sequence. The relationship between this multitude of data allows inferring the presence, nature and position of sequence variations in an unambiguous way. It is illustrative of the method that the derivation of the sequence is not critically dependent on the accuracy, i.e., the absolute values of the mass measurements. It is rather the coherent ensemble of mass-shifts and appearances/disappearances of cleavage sites that uniquely define the sequence. The computer simulations, described herein, assumed a resolution of 5 Da or 0.1%, a figure which is well above what can be achieved with state-of-the-art equipment. Also, it should be pointed out that the determination of the correct base composition is limited anyway to short fragments, even in the case of high-precision measurements [e.g., 5-mers in the case of unrestrained sequences and if the measurement has an accuracy of 0.01% or better; Limbach P., Mass Spectrom. Rev. 15: 297-336 (1996)]. Other methods in the art, which involve the accurate mass determination to assign the correct base composition to one or more fragments, will generally permit the detection of most sequence variations but not their unequivocal mapping. In these experiments it is generally assumed that a certain experimental observation relates to one particular previously known sequence variation, ignoring the fact that alternative sequence variations can explain the same result.
The present invention encompasses several additional embodiments and aspects described hereinafter and certain other embodiments will be readily apparent to one of ordinary skill in the art.
Target Nucleic Acid Preparation and Fragmentation
(a) Derivation of Target Nucleic Acid and Approaches to Cleaving with Base-Specificity
Nucleic acid molecules can be isolated from a particular biological sample using any of a number of procedures, which are well-known in the art, the particular isolation procedure chosen being appropriate for the particular biological sample. To obtain an appropriate quantity of isolated target nucleic acid on which to perform the methods of the present invention, amplification of the target nucleic acid may be necessary. Examples of appropriate amplification procedures for use in the invention include but are not limited to: cloning [Sambrook et al., Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press (1989)], polymerase chain reaction (PCR) [Newton C. R. and Graham A., PCR, BIOS Publishers (1994)] and variations such as RT-PCR [Higuchi et al., Bio/Technology 11: 1026-1030 (1993)] and allele-specific amplification (ASA), strand displacement amplification (SDA) [Terrance Walker G. et al., Nucleic Acids Res. 22: 2670-77 (1994)], and transcription based processes.
One embodiment of the present invention is directed to methods for sequencing (re-sequencing, etc.) Nucleic acid comprising the digestion of an RNA copy of each strand of the target nucleic acid with the RNases T1 and U2. One of the advantages of the method is the use of RNA, which exhibits higher sensitivity and better stability in MALDI-MS compared to DNA [Hahner S. et al., Nucleic Acids Res. 25: 1957-1964 (1997)]. Typically, the first stage of this aspect of the invention involves the amplification of the target nucleic acid by PCR or reverse-transcription followed by PCR. (RT-PCR) This can be achieved with a pair of dedicated primers that incorporate promoter sequences as non-annealing 5′-extensions. In a second stage, these promoters are used for the specific transcription of the adjacent sequences including the target sequences. Preferably, the promoter sequences are small and permit the in vitro transcription by a single subunit cognate RNA polymerase such as those deriving from bacteriophage T7, T3 and SP6. Preferred for use in this aspect of the invention are C and/or U analogs that can be incorporated during transcription and that exhibit favorable mass differences [e.g. m5U; supra]. The use of PCR primers that carry different promoter sequences permits the generation of an RNA copy of both strands in two parallel strand-specific transcription reactions. Both strands may also be transcribed from the same promoter sequence: this requires two parallel amplification reactions with only one promoter tagged primer. Alternatively, the in vitro transcripts may also be produced from sequences cloned in special purpose vectors such as the pGEM-type vectors available from Promega (Madison, Wis.) which contain appropriate promoters. The third step further comprises the treatment of the resultant RNA transcripts with one or more complementary mononucleotide-specific RNases (e.g RNase-T1 and RNase-U2), such that each desired position in the target sequence is interrogated. The final step in the process consists of the mass-spectrometric analysis of the RNA fragments resulting from the complementary cleavage reactions and the comparison of the spectra obtained with those of the known reference sequence.
Alternative schemes to prepare target nucleic acid obtained from a biological sample and to subject the target sequence to a set of complementary mononucleotide-specific cleavage reactions are also within the scope of the invention. The target nucleic acid can be DNA, cDNA, any type of RNA, DNA/RNA hybrid, or of mosaic RNA/DNA composition [depending on the ratio of ribo- and deoxyribonucleoside triphosphates (rNTP/dNTP) in the synthesis reaction; Sousa R. and Padilla R., EMBO J. 14: 4609-4621 (1995): Conrad F. et al., Nucleic Acids Res. 23: 1845-1853 (1995)]. The target sequence may also include modifications that are either introduced during or after enzymatic synthesis.
In general, different forms of each target sequence will be prepared so as to be able to perform a complementary set of mono-specific cleavage reactions. The cleavage reactions may be performed enzymatically and/or chemically. The mononucleotide-specificity of the digestion reactions may reside in the cleaving agent (e.g RNase T1), in the structure of the target nucleic acid, or in a combination of both. For example, RNase A (specific for both C- and U-residues) can be made monospecific by modifications of the substrate sequence that block the ribonucleolytic action at C or U residues. RNase A cleavage at U residues can in theory be prevented by chemical modification [Simoncsits A. et al., Nature 269: 833-836 (1977)]. The enzymatic incorporation of nucleotide analogs, most notably those modified at the 2′-hydroxyl group of the ribose is particularly preferred in the practice of the invention. A variety of such analogs have been demonstrated to be substrates for T7 RNA polymerase; e.g. 2′-fluoro, 2′-amino [Aurup H. et al., Biochemistry 31: 9636-9641 (1992)], 2′-O-methyl [Conrad F. et al., Nucleic Acids Res. 23: 1845-1853 (1995)], as well as 2′-deoxy NTPs [Sousa R. and Padilla R., EMBO J. 14: 4609-4621 (1995); Conrad F. et al., Nucleic Acids Res. 23: 1845-1853 (1995)]. The above strategy may also be used to improve the specificity of certain RNases such as RNase U2 which is said to cleave GpN phosphodiester bonds in extensive digests [Brownlee G., in “Laboratory Techniques in Biochemistry and Molecular Biology” (Work T. S. and Work E., eds.), North-Holland, Amsterdam, pp 199-200 (1972)]. Mosaic DNA/RNA target sequences that incorporate only one specific rNTP and that can be obtained quite efficiently with particular mutant polymerases [Sousa R. and Padilla R., EMBO J. 14: 4609-4621 (1995); Gao G. et al., Proc. Natl. Acad. Sci. USA 94: 407-411 (1997); Bonnin A. et al., J. Mol. Biol. 290: 241-251 (1999)], may allow mono-specific cleavages by alkaline treatment or by digestion with a non-specific RNase such as RNase-I [Meador J. et al., Eur. J. Biochem. 187: 549-553 (1990)].
Alternative strategies to obtain selective cleavage of target sequences make use of phosphorothioate chemistry. DNA and RNA polymers with phosphorothioate internucleoside linkages in the Rp stereo-configuration are readily synthesized [see Eckstein F., Ann. Rev. Biochem. 54: 367-402 (1985) and references cited therein]. Such phosphorothioate linkages can be specifically hydrolyzed following alkylation [Gish G. and Eckstein F., Nucleic Acids Symp. Ser. pp 253-256 (1987); Gish G. and Eckstein F., Science 240: 1520-1522 (1988)]. Mono-nucleotide 25 specific fragmentation according to this aspect of the invention would require the synthesis of targets making use of one particular α-thio nucleotide triphosphate substrate. Some nucleases (e.g. nuclease P1) cannot hydrolyze Rp phosphorothioate diesters; indirect selective cleavage (at a natural phosphodiester) may thus be obtained with target sequences that incorporate three different αS-dNTPs (or αS-rNTPs).
(b) Alternative Complementary Reactions
The performance of the present sequencing methods will be understood by those skilled in the art to be dependent on the following interrelated factors: (1) the length of the region to be sequenced, (2) the resolution of the MS analysis, and (3), to some extent, the sequence itself. The longer the region of interest and, consequently, the larger the number of digestion products, the more important the resolution becomes. Also, the length of the region to be sequenced is directly proportional to the number of single nucleotide substitutions that cannot be unambiguously mapped on the basis of the four base-specific fragmentation patterns only (Example 1;
One such measure consists of dividing the target region and deriving two or more (partially overlapping) segments (e.g., amplicons) from the sample nucleic acid rather than sequencing the target region as a whole. In addition to setting the length, this allows to exert some control over the composition. This would abrogate problems arising when the region of interest contains a duplicated segment. A second measure consists of carrying out one or more alternative or additional reactions involving target fragments that incorporate one or more modified nucleotides that exhibit different molecular masses such as is described above. Those of skill in the art will know of the existence of a wealth of mass-modified nucleotide analogs, many of which are useful and can be reconciled with the enzymatic procedures of the present method. The nucleotide analogs will differentially affect the masses of many of the digestion products and will therefore yield a significantly different spectrum that may reveal the required information. The analogs U and m5U [supra] exemplify this. Simulation studies (which model the present invention) have indicated that the use of U resolves certain sequence ambiguities observed with m5U (data not shown), while overall the latter nucleotide analog results in considerably fewer sequence ambiguities (see Example 1).
Another option consists of performing one or more additional reactions on the complementary strand. Compared to, for example, a G-specific cleavage of one strand, the C-reaction of the complementary sequence will yield a different set of fragments characterized by other mass differences. The effect of including reactions on the complementary strand of the target sequence is therefore similar to the use of nucleotide analogs.
Still another alternative provided by the present invention and which is useful in obviating the potential problems exemplified above includes using reactions with alternative specificities of cleavage. For example, partial base-specific cleavage can be achieved by changing the reaction conditions or by use of a specially prepared target wherein the cleavable and uncleavable (e.g. 2′-modified; supra) forms of one particular nucleotide occur randomly. Alternatively, instead of partial base-specific cleavages, one or more specific digestions characterized by a greater stringency can be performed (e.g. dinucleotide- or relaxed dinucleotide-specificity; see below). The digestion of the target sequence, in double stranded DNA form, with restriction enzymes is still another alternative provided by the present invention. Double digestion (i.e. a combination of two base-specific cleavages) of target nucleic acid alone or in combination with other digestion methods of the present invention also represents an informative alternative within the scope of the present invention.
Another informative option within the scope of the present invention involves the analysis of truncated target sequences. More specifically, cleavage of chain terminated sequences prepared, for example, by incorporation of a particular 3′-deoxy nucleotide substrate, will yield spectra that contain additional fragments when compared to the spectrum of the full target nucleic acid and will consequently provide additional information that will, in certain cases, allow a more unambiguous indemnification of sequence variation. This approach will be particularly useful for the characterization of lengthy digestion products or regions containing complex sequence variations.
(c) Alternative Complementary Reactions: Cleavage Characterized by a Greater than Mononucleotide Specificity
In still another of its embodiments, the method of the present invention also includes nucleolytic processes that are characterized by a dinucleotide- or a relaxed dinucleotide-specificity. Such stringency of cleavage will facilitate the analysis of longer target sequences because the size distribution of the resultant digestion products is even better suited for analysis by MS than fragments with an average length of 4 nucleotides that are generated by mononucleotide-specific cleavage. Useful in this aspect of the invention are, for example, restriction endonuclease reagents capable of cutting DNA at dinucleotide sequences such as those described by Mead D. et al., WO 94/21663 (PCT/US94/03246). RNases that preferentially hydrolyze pyrimidine-adenosine (CA and UA) bonds have also been identified which are useful in the practice of the present invention [E. coli RNase-M, Cannistraro V. and Kennell D., Eur. J. Biochem. 181: 363-370 (1989); as in an endoribonuclease isolated from Saccharomyces cerevisiae, Stevens A. et al., J. Bacteriol. 164: 57-62 (1985); and as is the Enterobacter sp. C-ribonuclease, described by Marotta C. et al., Biochemistry 12: 2901-2904 (1973)]. As disclosed and exemplified in the present invention, the specificity of these enzymes can, if need be, essentially be restricted to CA- or UA-bonds by the use of target nucleic acids that incorporate dUMP (or dTMP) on the one hand and dCMP on the other hand.
Stringent or relaxed dinucleotide-specific cleavage may also be engineered through the enzymatic and chemical modification of the target nucleic acid. By way of non-limiting example, transcripts of the nucleic acid of interest may be synthesized with a mixture of regular and α-thio-substrates and the phosphorothioate internucleoside linkages may subsequently be modified by alkylation using reagents such as an alkyl halide (e.g. iodoacetamide, -iodoethanol) or 2,3-epoxy-1-propanol. The phosphotriester bonds formed by such modification are not expected to be substrates for RNases. Using this procedure, a mono-specific RNase, such as RNase-T1, can be made to cleave any three, two or one out the four possible GpN bonds depending on which substrates are used in the α-thio form for target preparation. The repertoire of dinucleotide-specific reagents useful in the practice of the present invention may be further expanded by using additional RNases, such as RNase-U2 and RNase-A. In the case of RNase-A, the specificity may be restricted to CpN or UpN dinucleotides through the enzymatic incorporation of the 2′-modified form of the appropriate substrates as described above. For example, to make RNase-A specific for CpG dinucleotides, a transcript (target) is prepared using the following substrates: αS-dUTP, αS-CTP, αS-ATP, and GTP. Thus, using the indicated methods described herein, it is possible to engineer all 16 dinucleotide specificities. However, not all dinucleotide-specific reagents described herein would be required if the complementary strand of the target nucleic acid is included in the analysis.
The strategy outlined above makes it possible to prevent cleavage within homopolymer tracts (stretches of A's, G's, C's or T's) by an RNase that is made specific (or is made specific as described above) for the repeated nucleotide. Indeed, incorporation of a particular αS-NTP, followed by alkylation, will selectively prevent cleavage within repeated stretches of that nucleotide, allowing cleavage to occur at the 3′-side of the last nucleotide in the repeat. Simulation studies, similar to those described in Example 1, have identified this as a particularly useful strategy. Sequence analysis by digestion of the two complementary strands with RNase-T1 and RNase-U2 yielded a 5- to 10-fold reduction in the number of ambiguous mutations when αS-GMP and αS-AMP were incorporated in the respective transcripts. These studies also suggest that the selective blockage of cleavage within repeats is accompanied by a relatively small increase in the average length of the digestion products, thereby resulting in considerably less loss of information.
Those of skill in the art will also readily recognize variations or alternatives in certain aspects of the fragmentation methods described herein. Such alternatives or variations encompassed by the present invention include but are not limited to:
In another embodiment, the methods of the present invention are directed to the simultaneous sequence determination of at least two non-contiguous regions in a sample nucleic acid. In contrast to traditional sequencing methods that generate a fragment-ladder (i.e. a nested set of fragments that share a common endpoint), the strategies outlined herein are equally useful for multiplex sequencing. Multiplex sequencing, according to the present invention, generally involves the co-amplification of selected regions of target nucleic acids. This can be achieved by using sets of dedicated primer pairs which flank or are co-terminal with a target nucleic acid to be amplified. Alternatively, the preparation of the multiple target nucleic acids comprises the concomitant amplification of restriction fragments derived from the sample nucleic acid. Some approaches are illustrated and exemplified in Example 5. A special case of multiplex sequencing consists of the simultaneous analysis of the two complementary strands of a double stranded target nucleic acid.
In yet another embodiment, the methods of the present invention can be used for the simultaneous sequence determination of the corresponding target region(s) of at least two biological samples. A sequence variation in one out of a pool of analogous target nucleic acids may go unnoticed when analyzing conventional sequence ladders by means of gel electrophoresis. With the present methods, a sequence variation will, as a rule, yield one or more distinct peaks in the various complementary mass spectra. This feature should allow the detection of mutations at a significantly lower ratio of mutant to wild-type allele and therefore permit the analysis of larger pools. The ability to pool renders the present methods useful for the discovery of sequence variations across particular target regions in a given population. For this application, typically 5-10 samples may be combined. In case the mutations have previously been identified, considerably more samples, e.g. several tens, can be combined. The characteristics that render the present method useful for the analysis of sample pools make the method also effective for the analysis of heterozygous samples (i.e., an equimolar mix of two alleles).
Mass Spectrometric Methods
Mass-spectrometric methods useful in the practice of the present invention include ionization techniques such as matrix assisted laser desorption ionization (MALDI) and electrospray (ES). These ion sources can be matched with various separation/detection formats such as time-of-flight (TOF; using linear or reflectron configurations), single or multiple quadrupole, Fourier transform ion cyclotron resonance (FTICR), ion trap, or combinations of these as is known in the art of mass spectrometry. [Limbach P., Mass Spectrom. Rev. 15: 297-336 (1996); Murray K., J. Mass Spectrom., 31: 1203-1215 (1996)).
Because the present methods generally require the analysis of complex oligonucleotide fragment mixtures, the MALDI approach, mostly resulting in singly charged molecules, is preferred over ES where significant multiple charging will further increase the number of spectral peaks. For the desorption/ionization process, numerous matrix/laser combinations can be employed.
Sequence Determination of Simple Versus Complex Variations
In another embodiment, the methods of the present invention are directed to the diagnostic sequencing of one or more target nucleic acids that, in comparison with a related reference nucleic acid, incorporates a sequence variation other than a single nucleotide substitution. Such a sequence variation can involve the deletion or insertion of one or more nucleotides as well as the substitution of multiple nucleotides.
Similar to single nucleotide substitutions, the insertion or deletion of a single nucleotide represents a simple sequence variation whose analysis using methods of the present invention is straightforward. Both of these types of sequence variations are associated with a characteristic set of (maximum nine) changes in the four complementary mononucleotide-specific fragmentation patterns. It will be understood that the methods of the present invention, similar to other sequencing methods, may not unambiguously locate the point of insertion or deletion when it concerns one nucleotide in a stretch of identical nucleotides. This, however, may be taken into consideration when performing a computer assisted analysis of whether the observed spectra relate in a unique way to a specific sequence variant in accordance with the practice of the present invention.
Analysis of a microsatellite DNA [also referred to as VNTR (variable number tandem repeat) or SSR (simple sequence repeat)] represents a special case whose analysis is readily achieved using the methods of the present invention. Although multiple nucleotides are involved with VNTRs or SSRs, the interpretation of the spectral changes on the basis of the known reference sequence is rather simple and the polymorphism (an altered number of repeat units) may readily be characterized.
The methods of the present invention may also be used to analyze more complex sequence variations such as those where multiple nucleotides are affected either through insertion, deletion, substitution or a combination thereof The analysis of a number of double and triple mutants is described below in Example 3d. Multiple substitutions within a target sequence are also expected to be accompanied by a characteristic number of spectral changes. This number depends on whether the substitutions are adjoining or separated, as well as on the intervening sequence in case the mutations are separated. Single nucleotide substitutions, isolated by a sequence that contains at least one A, G, C, and T, are each associated with 10 spectral differences as outlined above. In general, the analysis of complex sequence variants will require (elaborate) computational approaches. One possible algorithm involves the comparison of the experimentally observed spectra with those generated on the basis of all possible sequences in the short region to which the sequence variation is confined. Such an algorithm will identify the sequence variant or, in case of ambiguities, the different matching sequences. This procedure illustrates that the present methods may be applied to the de novo sequencing of short regions of a target sequence. It will be recognized that, in practice, the experimental observations will not only set the boundaries but will also define the length of the variant region such that the algorithm need not consider insertions or deletions. Additional experimentally derived information, such as the absence of a particular nucleotide, can further limit the sequence space the algorithm has to explore. In particular applications, the complex sequence variants may be previously known and may thus be part of the set of reference sequences. In such cases, the experimentally observed spectra may be directly correlated to those predicted for the reference sequences. There would however still be a need to compute whether such correlation is unique. The advantage of previous knowledge is that the experimental approach can be adapted such that the output information indeed relates uniquely to the potentially occurring complex sequence variations.
The present invention, in part, rests on the insight that computational analysis of the spectra obtained in a set of complementary cleavage reactions, and comparison of these data with the computationally predicted spectral changes from the known reference sequence, as illustrated herein, is an important step in the unambiguous determination of the presence, the nature and the location of sequence variations. More specifically, the computational approaches to simulate the experiment illustrated herein are necessary to determine whether a unique relation exists between the spectra obtained and a particular sequence variation. Accordingly, one aspect of the present invention contemplates a method which utilizes a computer algorithm or method capable of computing the spectral differences resulting from one or more nucleotide differences between the target nucleic acid and the reference nucleic acid, the method and algorithm comprising subjecting the reference nucleic acid and sequence variants thereof (i.e., target nucleic acid having nucleotide differences) to the different base specific cleavages to generate oligonucleotide fragments, computing the mass of each oligonucleotide fragment, generating the mass spectra of the oligonucleotide fragments from the reference nucleic acid and the sequence variants thereof for each of the base specific cleavage reactions, and matching these computationally derived mass spectra with the spectra obtained experimentally in the different base specific cleavage reactions.
In one preferred embodiment the computer algorithm is designed to systematically compute the spectra of all possible simple nucleotide variations of the reference nucleic acid, including but not limited to all possible single nucleotide substitutions, deletions and insertions. Since most of the genetic diversity found in living organisms involves single nucleotide variations, most of the experimentally observed sequence variations can be identified with the methods and algorithms of the present invention, meaning that one or more matches may be found between the observed spectra and the computationally derived mass spectra. In case a unique match is found, the sequence variation in the target nucleic acid is unique. When more than one match is found between spectra, the sequence variation cannot be established unambiguously.
It will be obvious to the person skilled in the art that different approaches may be used for performing the computational analysis, such as, but not limited to, performing the computational analysis on the complete reference sequence, or performing a serial computational analysis on segments of the reference sequence using, for example, a sliding window. The latter approach will enable the identification of different sequence variants occurring in different parts of the reference sequence.
In another embodiment, the methods and computer algorithms of the present invention are designed to explore all possible nucleotide sequences in a limited segment of the reference sequence. Such methods and algorithms may be used when the preceding approach fails to give a match, demonstrating that the sequence variation does not correspond to a simple nucleotide variation in the reference nucleic acid. This may be the case when more than one nucleotide change occurs within a short region, such that one or more cleavage products contain multiple nucleotide alterations. The region corresponding to these cleavage products can then be explored further by computing the spectra for all possible sequence permutations and determining the matching sequence. It is anticipated that given sufficient computing power, such methods and algorithms may be used for de novo sequencing using mass spectral data generated according to the present invention.
Applications of the Present Methods
The methods of the present invention are particularly well suited for rapidly and accurately re-sequencing nucleic acids from a variety of biological sources including, but not limited to, plants, animals, fungi, bacteria and viruses. Re-sequencing implies the detection and mapping of both previously known as well as unknown sequence variations (e.g. mutations and polymorphisms) relative to a related reference sequence. One of the most notable distinctions with respect to conventional gel-electrophoretic analysis of fragment ladders, is that generally each particular sequence (variation) results in a distinct and characteristic set of (mass) peaks. This feature makes the present methods effective for the reliable scoring of heterozygous samples, the simultaneous sequencing of multiple target regions from a single biological sample (i.e., multiplexing), as well as the simultaneous analysis of the analogous regions from different samples (i.e., pooling). The use of pools of individual samples should permit the cost-effective identification of previously unknown sequence variations in a population. This aspect of the invention properties makes the present methods valuable for clinical and public health studies. Very often such studies rely on samples (e.g., saliva, blood, swabs, paraffin-embedded tissue, biopsy material) that are cellularly and genetically heterogeneous and, consequently, require assays that can detect mutations at a low ratio of mutant over wild-type allele.
An additional advantage of the present methodology is that it can be tuned (by reducing the number of complementary cleavage reactions) such that the diagnostic sequencing is limited to particular positions in a target nucleic acid, a feature useful for the unambiguous scoring of previously identified mutations or polymorphisms. The processes described herein can be used, for example, to diagnose any of the more than 3000 genetic diseases currently known (e.g., hemophilias, thalassemias, Duchenne Muscular Dystrophy, Huntington's Disease, Alzheimer's Disease and Cystic Fibrosis) or genetic defects yet to be identified. In addition, certain DNA sequences may predispose an individual to any of a number of diseases or conditions such as diabetes, artherosclerosis, obesity, various autoimmune diseases and cancer (e.g., colorectal, breast, ovarian, lung). Depending on the biological sample, the diagnosis for a genetic disease or genetic predisposition can be performed either pre- or post-natally using the methods of the present invention. Re-sequencing of nucleic acids derived from infectious organisms using the methods of the present invention may reveal the basis of pathogenicity and may also be useful to identify the variation(s) that cause drug-resistance. For example, mutations in the protease/reverse transcriptase region of the human immunodeficiency virus (HIV) have been implicated in the decreased sensitivity towards the antiviral activity of protease and reverse transcriptase (RT) inhibitors. The re-sequencing of the nucleic acid encoding these viral domains is therefore of special interest to monitor disease progression (see Example 1). Similarly, sequencing, according to the present invention, may be useful to determine the antibiotic-resistance phenotype of certain bacteria [e.g. Mycobacterium tuberculosis; Head S. et al., Mol. Cell Probes 13: 81-87 (I 999); Troesch A. et al., J. Clin. Microbiol. 37: 49-55 (1999)].
In other embodiments, the present methods are directed to the identification and classification of target nucleic acids. Analyses according to the present invention characterize nucleic acids at a level essentially equal to sequence determination. Therefore, interrogated unknown sequences may be unambiguously identified by comparison of the obtained mass spectra with those known or predicted for a plurality of reference sequences. In this exercise, novel sequences that have no matching reference database sequence may also be found. The use of the methods for expression profiling (i.e., the analysis of cDNA libraries) as well as whole-genome sequencing is exemplified in Example 6 and 7, respectively. Other applications include the determination of identity or heredity (e.g., paternity or maternity).
Kits for Practicing the Invention
Kits for diagnostic sequencing of one or more target nucleic acids in a sample are also provided. In preferred embodiments, such kits comprise one or more reference nucleic acids, various reagents for sequence specific cleavage protocols, and computer algorithm(s). Such kits may optionally also contain nucleic acid amplification reagents. Additionally, the kits may contain reagents for the preparation of modified nucleic acids, including but not limited to modified nucleotide substrates. The kits may also contain buffers providing conditions suitable for certain enzymatic or chemical reactions. In addition, the kits may contain reagents, such as solid supports, for purposes of isolating certain nucleic acids and preparing nucleic fragments for mass spectrometric analysis.
The foregoing aspects of the invention are illustrative and should not be construed to limit the invention as set out in the appended claims. Variations in some aspects as well as alternative procedures will be readily recognized by one of ordinary skill in the art.
Example 1 describes modeling the diagnostic sequence analysis of a 1200 base-pair region of HIV-1 using methods of the present invention.
Example 2 describes methods for base-specific cleavage by modifying the nucleic acid template to be cleaved.
Example 3 illustrates the diagnostic sequencing of the RNase-T1 coding region according to the methods of the present invention
Example 4 illustrates the analysis of a ˜1000 base-pair nucleic acid.
Example 5 illustrates the use of the present invention for genotyping, including multiplex genotyping.
Example 6 illustrates the use of the present invention for transcription profiling.
Example 7 illustrates the use of the present invention for whole genome resequencing.
The methods of the present invention have been utilized on a 1200 base-pair sequence derived from human immunodeficiency virus type 1 (HIV-1; HXB2 isolate; Genbank accession number K03455; position 2161 to 3360). This sequence was used as a model in computer simulations to examine the overall performance of the method, as well as the occurrence of ambiguities. The selected region encompasses the entire protease gene and the first ˜270 codons of reverse transcriptase [compare with Hertogs K. et al., Antimicrob. Agents Chemother. 42: 269-276 (1998)]. The genotyping/re-sequencing of this domain of clinical isolates of HIV is of special interest in order to monitor the emergence of drug resistance-associated mutations. Single as well as multiple changes have been implicated in the decreased sensitivity towards the antiviral activity of protease and RT inhibitors [Hertogs K. et al., Antimicrob. Agents Chemother. 42: 269-276 (1998); Schinazi R. et al., Int. Antivir. News 4: 95-107 (1996) and references cited therein].
The principal objective of the computer simulation was to examine the performance of the re-sequencing method for detecting and mapping SNPs. To this end we have performed computational simulation analyses in which we have systematically mutated each nucleotide one by one in the 1200 base-pair sequence. For each mutation we have calculated the molecular masses of the cleavage products that would be generated from a given segment of the sequence in the different four RNase digestion reactions, namely upon RNase-T1 and RNase-U2 cleavage of the (+) and (−) strands. The comparison of these masses with those of the reference cleavage products from the original sequence identifies the masses of the diagnostic fragments associated with each mutational change, i.e., fragments that either appear or disappear as a result of the mutation. The underlying assumption in this analysis was that in order to be measurable, the fragment must have a molecular mass different from those of the other cleavage products generated in the same reaction. Furthermore, we have assumed that the resolution of the mass spec analysis is limited to mass differences larger than either 5 Da or 0.1%. In other words fragments whose mass difference with other fragments in the same digest is smaller than 5 Da or 0.1% were not scored in the analysis. The quantitative aspects of a mass spectrum (i.e. peak heights) were not considered in the present simulation study. For each mutational change we have computed the number of fragments that are diagnostic for the presence of the mutation. Mutational changes were scored as detectable when there was at least one diagnostic fragment (showing a spectral change). In addition, we have examined whether the mutational changes can also be mapped unambiguously. To this end we have compared the sets of diagnostic fragments associated with each mutation. Mutations that yield unique sets of fragments can be mapped unambiguously, while mutations that give the same sets cannot be distinguished from one another.
In a first simulation analysis we have computed the fraction of SNPs that may be detected and mapped using respectively 1, 2 and 4 RNase digestion reactions. To this end we have performed a systematic single nucleotide substitution simulation on a 200-base-pair segment of the HIV sequence. For each of the four different RNase digestion reactions (RNase-T1 and RNase-U2 cleavage of the (+) and (−) strands] we have calculated the number of detectable diagnostic fragments and have analyzed whether these fragments are unique for each mutation. The results summarized in
It will be obvious that the quality of the sequences obtained with the methods of the invention will be strongly influenced by the size of the sequence segments that are examined. Indeed, the larger the size of the segment, the larger the statistical chance that certain relevant diagnostic fragments may coincide with other cleavage products generated in the same reaction. We have therefore performed a systematic single nucleotide substitution simulation analysis on the 1,200 base-pair HIV sequence using different size segments, namely 100, 200, 300 and 600 base-pairs. In each simulation a total of 3,600 single mutational substitutions was analyzed. For each of the four different RNase digest reactions both the number and the patterns of the measurable diagnostic fragments were computed using the detection limits described above.
In each simulation we determined both the number of detectable SNPs as well as the fraction of SNPs that can be mapped unambiguously. The results of the computational simulations summarized in
In conclusion, the results of the simulations show that the methods of the invention are effective for re-sequencing and that even large segments may be used when only a limited number of positions need to be analyzed. Also, it appears that in most cases a computer-aided simulation study will be essential in the experimental design as well as the data interpretation when using the methods of the present invention. Most importantly, the simulations will indicate whether spectral changes are unambiguously linked to particular sequence variations.
The present example illustrates that the specificity of cleavage by a nucleolytic reagent may be further confined through the modification of the target template such that particular phosphodiester bonds resist cleavage. More particularly, it is demonstrated that RNase-A, which normally cleaves at the 3′-side of both C- and U-residues, becomes mononucleotide-specific when the target incorporates the 2′-deoxy analog of one of these nucleotides. A region of the plasmid vector pGEM3-Zf(+) (Promega, Madison, Wis.), encompassing the multi-cloning site as well as the phage T7 promoter sequences, was used as a model (see
The first step towards the sequence analysis according to the present invention involved the amplification of the 158 base-pair test sequence. The reaction was carried out in a total volume of 50 1 using 12.5 pmol each of the forward and reverse primer, 200 M of each dNTP, 0.25 1 Taq DNA polymerase (5 U/1; Promega, Madison, Wis.), 1.5 mM MgCl2 and a buffer supplied with the enzyme. After an initial incubation at 94° C. for 2 min, 40 cycles of the following temperature program were performed: 94° C. for 30 sec, 50° C. for 30 see, and 72° C. for 15 sec. The sample was kept an additional 15 min at 72° C. and then chilled. The PCR reaction product was purified (High Pure PCR Product Purification Kit; Roche Diagnostics Belgium, Brussels, Belgium) and subsequently used for transcription of one specific strand. A mutant T7 RNA polymerase (T7 R&DNA™ polymerase; Epicentre, Madison, Wis.) with the ability to incorporate both dNTPs and rNTPs was used in the transcription reactions. In addition to a transcription with the regular ribonucleotide substrates, one reaction was performed where CTP was replaced by dCTP, while in two more separate transcriptions either dUTP or dTTP replaced UTP. The transcription reactions were run in a 50 1 volume containing: 40 mM Tris-Ac (pH 8.0), 40 mM KAc, 8 mM spermidine, 5 mM dithiothreitol, 15 mM MgCl2, 1 mM of each rNTP, 5 mM of dNTP (in these cases the appropriate NTP was excluded), ˜40 nM DNA template (˜2 pmol), and 250 units T7 R&DNA™ polymerase. Incubation was performed at 37° C. for 2 hours. After transcription, the full-length T7 in vitro transcripts (118 nucleotides) were purified by allowing them to anneal to the 5′-biotinylated form of the complementary reverse PCR primer (
The RNase-A cleavage products predicted for each of the four transcripts are shown in Table II. Note that the mass calculation of the predicted fragments assumes a 3′-phosphate group and not the 2′,3′-cyclic phosphate intermediate of the cleavage reaction. Overall, the experimentally obtained spectra (
The protocol described in the present Example is illustrative and certain modifications and variations will occur to the skilled artisan. The immobilization of the transcripts represents an easy means to prepare the material for MS analysis, e.g., removal of all other reaction components and exchange of Na+ and K+ counter-ions for (NH4)+ (note that the subsequent RNase digestion does not require any reagents that are ‘incompatible’ with MS). While other methods, such as chromatography, may be 20 used to prepare the transcripts or the derived digestion products for MS analysis, the present method is favorable in that it is readily amenable to automation and high-throughput analysis. In repeat experiments, yielding essentially the same results as described herein, the transcripts were digested in water and ˜15 nanoliter of these digests was directly applied onto a Spectrochip™ (Sequenom Inc., San Diego, Calif.) for analysis by MALDI-TOF-MS.
The present example illustrates the application of the methods of the invention to the re-sequencing of a portion of the RNase-T1 coding region. We selected the RNase-T1 coding region because of the availability of a collection of site-directed mutants [Steyaert J., Eur. J. Biochem. 247: 1-11 (1997)] which had previously been sequenced using the classical dideoxy chain termination method. The wild-type and mutant sequences, used in the present example, are shown in
a. Analysis of the Wild-Type RNase-T1 Sequence
The experiments were performed essentially as described in Example 2. First, the selected wild-type RNase-T1 target sequences were amplified by PCR with the following primers:
The resultant amplicon was subsequently re-amplified using either a forward or a reverse primer that incorporates the T7 promoter site as a 5′ non-annealing extension (see
A graphical representation of the spectra is shown in
b. Analysis of Selected RNase-T1 Single Point Mutations
Four single nucleotide substitutions were chosen (mutant #1, #2, #3, and #4 in
c. Analysis of a Mixture of Wild-Type and Mutant RNase-T1 Sequences
The analyses shown in Table IV can be used to simulate experiments where equimolar mixtures of the wild-type RNase-T1 sequence and one of the single nucleotide substitutions are examined. In such cases, which mimic heterozygotic genotypes, the spectra contain a number of novel fragments in addition to all those derived from the (wild-type) reference sequence. The characterization and location of the mutation/polymorphism is therefore necessarily based on the novel fragments only. Unambiguity requires that the novel fragments are sufficient to uniquely define the mutation. Those of skill in the art will realize that zygosity determination is straightforward using the present methods because each allele is associated with a distinct set of peaks.
We performed a number of experiments where on particular single nucleotide mutant (e.g., mutant #3;
d. Analysis of RNase-T1 Multiple Mutants
The methods of the present invention are not limited to the analysis of single nucleotide substitutions. Complex variations can also be sequenced. Table IV lists the spectral changes that are predicted to be associated with a number of RNase-T1 multiple mutants, more particularly double and triple mutants (mutant #5, #6, #7, and #8 in
The methods of the invention arc designed to overcome the limitation of the short read lengths encountered with current MS-based sequencing methodologies that involve the analysis of fragment-ladders. One can envision that, depending on the application, target regions of several hundred or even a few 1000 base-pairs can be analyzed. The present example demonstrates that a large number of oligonucleotide fragments can be analyzed simultaneously by the methods of the present invention and that, consequently, the detection platform does not impose a limit on the methodology.
Following the scheme presented in Example 2, a 1012 base-pair region of the plasmid vector pGEM3-Zf(+) (Promega, Madison, Wis.) was amplified and the resultant amplicon, subsequently, used for preparation of a 972 nucleotides long in vitro T7 transcript (see
The methods of the present invention are also useful for the diagnostic sequencing of multiple non-contiguous regions of a sample nucleic acid. This renders the present methods useful for the genome-wide discovery as well as the routine scoring of polymorphisms (e.g. SNPs) and mutations at multiple loci in genomic DNA. Such multiplex genotyping is conceptually no different than re-sequencing; both require that alterations are characterized and positioned unequivocally. Similar to experiments involving a single target sequence described above, a computer simulation can be performed to find out which ones of the observed spectral changes is uniquely linked to particular genomic alterations. Since multiplex genotyping only requires the identification/diagnosing of a number of variant positions, it will be recognized by those of skill in the art that (i) the complexity (i e the combined length) of the multiple target sequences may be significantly greater than in the case of fall re-sequencing, and (ii) a single specific cleavage reaction may often suffice for both allele and zygosity identification. Applications which involve the use of two sequence-specific cleavages that each positively identify one of the two alternative forms of a series of bi-allelic SNPs are also possible using the methods of the present invention. For example, many C to T transitions, the most common type of point mutations and polymorphisms in human, may be easily scored by a combination of C- and T(U)-specific reactions. It is worth mentioning that heterozygous samples analyzed using gel-electrophoretic sequencing are often difficult to identify with confidence. With the methods described herein, the detection of heterozygosity is unambiguous because of the presence of both the wild-type and the mutation specific set of mass spectral peaks.
Multiplex genotyping will generally involve the co-amplification of genomic regions. In the case of previously known SNP genetic markers, co-amplification of selected loci can be achieved by using dedicated primer pairs [Wang et al., Science 250: 1077-1081 (1998)]. Alternatively, a more generic approach can be adopted for both the discovery and the subsequent routine scoring of a set of SNPs where the preparation of target sequences comprises the concomitant amplification of multiple short restriction fragments derived from the sample nucleic acid. This ‘random sampling’ method may be particularly useful with organisms that have a high polymorphism content (e.g., more than 1 SNP in 100 base-pairs). This co-amplification can be achieved by ligating to the ends of the restriction fragments adaptor sequences that incorporate the target sites for a single PCR primer pair. In this approach, the average size of the amplicons must be small such that the majority incorporates ≧1 SNP while, additionally, the total number of the amplicons must be sufficiently small so that their combined length is amenable to analysis by the present methods. These requisites can be met by the appropriate choice of restriction enzymes and the use of methods that permit the selective amplification of discrete subsets of restriction fragments [Vos P. et al., Nucleic Acids Res. 23: 4407-4414(1995); Zabeau M. and Vos P., EP 0534858 (1993); Kikuya Kato, Nucleic Acids Res. 23: 3685-3690 (1995)] and as described herein. For example, a first restriction enzyme that cleaves rarely in the genome under study can be combined with a second reagent that generates fragments with an average size of about 100 base-pairs (e.g., a combination of two enzymes with tetra-nucleotide recognition sites). The number of fragments edged by the two different restriction sites should preferably be less than 100,000; a suitable subset of these can readily be amplified by the use of selective primers [Vos P. et al., Nucleic Acids Res. 23:4407-4414 (1995)]. In addition, a PCR protocol, characterized by a highly shortened elongation time, can be used such that the amplification of short fragments is strongly favored thereby further reducing the number and the average size of the amplicons. During the selective co-amplification of genomic fragments or in a subsequent PCR step, a first primer can be used that attaches a full promoter sequence (e.g., one deriving from bacteriophage T7, T3 or SP6; supra) to the amplicons. The second strand may be synthesized by extension of a primer that contains a ribonucleotide residue at, for example, the penultimate position. Following PCR amplification, the primer sequences can be removed from this second strand by RNase digestion, and the resultant truncated strand transcribed with the aid of the first primer. This procedure minimizes the common sequences that are connected to the target restriction fragments.
Diagnostic sequencing will, generally, be performed on a defined nucleic acid, i.e. one will know to what reference sequence the target nucleic acid corresponds. However, the re-sequencing methods according to the present invention can also be used to identify or classify certain sequences. In such experiments, the interrogated nucleic acid (e.g. a random clone of DNA) will typically correspond to an unknown portion of a (much) larger sample sequence or represent one out of a plurality of nucleic acids present in a biological sample, or a combination of both. The mass spectra derived from the unknown nucleic acid are compared to those known or predicted for the related reference sequence(s), or portions thereof. Note that, in this type of experiments, some of the interrogated target sequences need not necessarily have their counterparts in the reference sequences, and vice versa. It will be realized that sequence identification according to the present methods may, at the same time, reveal possible sequence variations. Interrogated sequences may thus be classified as identical to one of the database sequences, as a variant of such as a reference sequence or as novel in case no matching sequence is found.
It should be recognized that analyses that involve at least the four complementary mono-nucleotide specific cleavage reactions identify unknown sequences with a resolution essentially equal to sequence determination. At the same time, the MS-based methods described herein allow fast data acquisition and are amenable to high-throughput. Therefore, the present methods are useful to identify and catalogue nucleic acids at an unprecedented scale and speed. One application consists of the analysis of cDNA libraries for the purpose of: (i) the assembly of unigene libraries (i.e. the identification/removal of replicate clones), (ii) the identification of novel genes or novel variants of previously identified genes, and (iii) transcription profiling. The speed and throughput of the present method should permit the processing of more clones and, hence, a more in depth analysis of a cDNA library.
A variety of methods are known in the art for transcription profiling, i.e. the analysis of the transcription in both qualitative and quantitative terms. In one method, the expressed-sequence-tag (EST) approach, the mRNA population is assessed by partial sequencing of randomly selected cDNA clones. Global changes in gene-expression patterns are deduced from the EST ratios among two compared cDNA libraries [Lee N. et al., Proc. Natl. Acad. Sci. USA 92: 8303-8307 (1995)]. The methods described herein may be used to catalogue expressed genes with a similar level of resolution but considerably higher speed and throughput. First, a library of unidirectionally cloned cDNAs is constructed in a vector that permits transcription of the inserted sequences. Preferably, the 3′-end of the cDNAs is located adjacent to the promoter. Template for transcription can be prepared by amplification of the promoter-cDNA cassette using a pair of vector-specific primers. Alternatively, vector DNA is prepared and cleaved at a restriction site within the vector close to the 5′-end of the inserted cDNA (e.g. ˜25 base-pairs). Preferably, the restriction site at which the templates are cleaved should have a low occurrence frequency within the cDNAs under study. Run-off transcripts, synthesized from PCR product or digested vector, are characterized by a common 3′-end, consisting of vector sequences, which allows the isolation of full-length transcripts as described in Example 2. An alternative strategy involves treatment of the vector DNAs with a restriction reagent such that not only all templates are digested at the cDNA 5′-end but that a vast majority is also cleaved within the cDNA at some distance from the 3′-end (e.g. a few hundred base-pairs). The restriction reagent may be a single enzyme or a combination of two or more restriction enzymes. Ligation of an adaptor to the digestion product(s) [see Vos P. et al., Nucleic Acids Res. 23: 4407-4414 (1995)] can be considered so as to obtain full-length transcripts with a common 3′-end enabling their isolation as described in Example 2. However, transcripts that incorporate a biotin group at the 5′-end may also be prepared [Hahner S. et al., Nucleic Acids Res., 25: 1957-1964 (1997)], providing an alternative means for their immobilization. Digestion within the cDNAs is an attractive option in that different partial cDNAs deriving from the same transcript are made congruent by this procedure and thereby facile to identify. The full-length run-off transcipts are finally subjected to complementary sequence-specific cleavage reactions, and the resultant digestion products analyzed by MS as disclosed herein.
Those of skill in the art will recognize the advantages of the transcript profiling method outlined above. Comparable to the EST approach, cDNAs are identified at the sequence-level, i.e. the ultimate level of resolution. Thus, while the method involves fragmentation of the interrogated nucleic acid, its level of resolution far exceeds that attained by fingerprinting techniques [Prashar Y. and Weissman S., Proc. Natl. Acad. Sci. USA 93: 659-663 (1996); Bachem C. et al., The Plant Journal 9: 745-753 (1996); Ivanova N. and Belyavsky A., Nucleic Acids Res. 23: 2954-2958 (1995); Liang P. and Pardee A., Science 257: 967-971 (1992)]. In contrast to hybridization-based approaches [Schena M. et al., Science 270: 467-470 (1995); Wodicka L. et al., Nature Biotechnology 15: 1359-1367 (1997)] the method can identify both known and previously unknown sequences. Also, it should prove faster then methods requiring gel-electrophoretic fractionation.
In the past couple of years the technology for sequencing entire genomes, especially those of microorganisms, has come to maturity. More than 50 microbial genomes are scheduled to be completed by the year 2000, and the benefits emerging from this vast body of knowledge are rapidly becoming clear [Clayton R. et al., Curr. Opinion Microbiol. 1: 562-566 (1998)]. It seems clear that sequencing entire microbial genomes is becoming routine and that microbial genetics is entering the era of ‘comparative genomics’. Knowledge of the complete genome sequence is the ultimate tool in phylogenetic analyses, allows gene/functional diversity studies, and fundamentally changes the manner in which research is conducted in an organism. At the present time, a substantial portion of each new genome sequence has no database match. One may expect to see a greater proportion of orthologous genes in the future, when the microbial species diversity is better represented. At that point, when most of the sequences generated will be similar to already known sequences, global genome analyses could be performed rapidly, accurately, and cost-effectively using a re-sequencing strategy as described herein rather than by de novo sequencing methods. Similar evolutions may be anticipated outside the bacterial genetics field where genome projects for many (model) organisms are ongoing or have already been finished (e.g., Drosophila melanogaster, Caenorhabditis elegans, human, mouse, Arabidopsis thaliana, and rice). The methods of the present invention may be readily adapted to the re-sequencing of entire (bacterial) genomes or megabase nucleic acid regions. This may be accomplished with the use of a shotgun approach that involves the sequence analysis of unselected subclones that harbor random fragments according to the methods of the present invention. The assembly of all the independent, random sequences is fundamentally different from that in a de novo sequencing project [Fleischmann R et al. Science 269: 496-512 (1995)] because of the availability of a reference sequence that serves as a scaffold. The assembly into a single complete sequence comes down to matching each set of experimentally obtained spectra with a portion of the reference sequence. The computational approaches required to accomplish this are similar to those that are needed for the analysis of cDNA libraries, outlined in Example 6. In both cases one does not know in advance the reference sequence, if at all existing, for a given interrogated target region. It should be noted, however, that the present shotgun approach might be even more demanding in terms of computational power because of the undefined ends of the segments. At the same time, the algorithms must be capable of mapping the variations that occur between the target and the reference sequence. It is expected that a shotgun approach with its built-in redundancy (i.e., most sequences will be covered several-fold) should prove useful for the comprehensive comparison of a pair of related genomes. An alternative for the shotgun approach strategy consists of the analysis of clones from one or more libraries of restriction enzyme fragments or the analysis of defined amplicons generated with locus specific primer pairs.
While the present invention has been described in terms of the preferred embodiments, it is understood that variations and modifications will occur to those skilled in the art. Therefore, it is intended that the appended claims cover all such equivalent variations which come within the scope of the invention as claimed. All of the references cited herein are expressly incorporated by reference.