CA2342838A1

CA2342838A1 - Molecular cloning using rolling circle amplification

Info

Publication number: CA2342838A1
Application number: CA002342838A
Authority: CA
Inventors: Paul M. Lizardi
Original assignee: Individual
Current assignee: Yale University
Priority date: 1998-09-15
Filing date: 1999-09-15
Publication date: 2000-03-23
Also published as: WO2000015779A3; EP1114184A2; US20020048761A1; JP2002525049A; WO2000015779A2; AU770993B2; US6287824B1; AU5925099A

Abstract

Disclosed are reagents and a method for efficient in vitro molecular cloning of nucleic acid molecules of interest. Because the method is entirely in vitro, it can be automated and scaled-up in ways that are not possible in ce ll- based molecular cloning. The method involves insertion of a nucleic acid molecule of interest in a linear vector to form a circular vector where one strand is continuous and the other strand is discontinuous. The continuous strand of the circular vector is then amplified by rolling circle replicatio n, amplifying the inserted nucleic acid molecule in the process. The amplification is rapid and efficient since it involves a single, isothermic reaction that replicates the vector sequences exponentially. The amplificati on process is amenable to automation where multiple reactions are carried out simultaneously in a small area. The amplified nucleic acid can be used for a ny purpose and in any manner that nucleic acid cloned or amplified by known methods can be used. This includes sequencing, probing, restriction analysis , subcloning, transcription, hybridization or denaturation analysis, further amplified, and storage for future use or analysis.

Description

MOLECULAR CLONING USING

ROLLING CIRCLE AMPLIFICATION
BACKGROUND OP THE INVENTION
The disclosed invention is generally in the field of molecular cloning 5 and nucleic acid amplification, and specifically involves rolling circle replication of nucleic acid molecules inserted into circular vectors.
DNA molecular cloning is routinely carried out using plasmid, phage, or viral vectors that replicate inside cells. A method, in which individual DNA molecules are cloned in solution by serial dilution and subsequent PCR
10 amplification from tubes containing single molecules has been described (Lukyanov et al., Nucleic Acid Research 24:2194-2195 (1996)). A method has also been described for cloning RNA populations derived from single RNA molecules in an immobilized medium (Chetverina and Chetverin, Nucleic Acids Research 21:2349-2353 (1993)). While both of these methods 15 allow in vitro cloning, neither is practical for high throughput cloning.
Velculescu et al., Science 270:484-487 (1995), have described a method for the quantitative cataloguing and comparison of expressed genes in normal, developmental, and disease states. The method, termed serial analysis of gene expression (SAGE), is based in the use of relatively short 20 sequence tags for the unique identification of cDNAs derived from mRNA
transcripts. While this method is very powerful, the study of low-abundance mRNAs can require several months of work in order to obtain sufficient sequence information for a complete SAGE analysis of one tissue sample.
Thus, there is a need for a method to obtain the sequence of sequence tags 25 more rapidly.
It is therefore an object of the present invention to provide a more efficient method of in vitro molecular cloning.
It is also an object of the present invention to provide vectors and kits useful for in vitro cloning.
30 It is also an object of the present invention to provide an automated method molecular cloning.
It is also an object of the present invention to provide a more efficient method of sequential analysis of gene expression.

BRIEF SUMMARY OF THE INVENTION
Disclosed are reagents and a method for efficient in vitro molecular cloning of nucleic acid molecules of interest. Because the method is entirely in vitro, it can be automated and scaled-up in ways that are not possible in 5 cell-based molecular cloning. The method involves insertion of a nucleic acid molecule of interest in a linear vector to form a circular vector where one strand is continuous and the other strand is discontinuous. The continuous strand of the circular vector is then amplified by rolling circle replication, amplifying the inserted nucleic acid molecule in the process.
The amplification is rapid and efficient since it involves a single, isothermic reaction that replicates the vector sequences exponentially. 'the amplification process is amenable to automation where multiple reactions are carried out simultaneously in a small area. The amplified nucleic acid can be used for any purpose and in any manner that nucleic acid cloned or amplified 15 by known methods can be used. This includes sequencing, probing, restriction analysis, subcloning, transcription, hybridization or denaturation analysis, further amplified, and storage for future use or analysis.
The insertion reaction involves insertion of a double-stranded nucleic acid molecule into a double-stranded linear vector to produce a double-20 stranded circular vector. The use of circular vectors facilitates the selection of molecules that have successfully incorporated inserts. The amplification reaction involves rolling circle replication of a single-stranded circular nucleic acid molecule.
A key feature of the method, which facilitates double-stranded 25 insertion followed by single-stranded amplification, is formation of the circular vector in such a way that one of its strands is a closed circular strand (that is, continuous) while the other strand is not a closed circular strand (that is, it has a nick, a gap, an overlap, or is otherwise discontinuous). This feature is most useful, and most effectively accomplished, when, by 30 operation of the method, the closed strand and the open strand are predetermined; that is, when a panCicular strand of the vector is selectively left discontinuous.

With rolling circle replication, amplification takes place not in cycles, but in a continuous, isothermal replication. This makes amplification less complicated and much more consistent in output. A single round of rolling circle replication results in a large amplification of the circular vector, orders of magnitude greater than a single cycle of PCR replication and other amplification techniques in which each cycle is limited to a doubling of the number of copies of a target sequence.
Following amplification, the amplified nucleic acid can be used for any purpose. Numerous methods for the use and manipulation of cloned or 10 isolated nucleic acid are known and can be applied to nucleic acid amplified in the present method. For example, the nucleic acid can be sequenced, probed, subjected to restriction analysis, subcloned, transcribed, subjected to hybridization or denaturation analysis, further amplified, or stored.
Diagnostic methods, such as sequencing and probing for specific sequences, 15 are preferred.
Libraries of cloned nucleic acids formed by the disclosed method can be screened using any of the methods used for screening conventional libraries. For example, cDNA libraries made using the disclosed method can be analyzed using conventional screens. Libraries can also be used for in 20 situ transc:iption to generate RNA colonies, which can then be analyzed (in situ or in replicas) by appropriate screens, such as aptamer screens or ribozyme activity screens. Libraries can also be screened by in situ translation on array replicas (see. for example, Saris et al., Nucleic Acids Res.
10:4831-4843 (1982)). Libraries can also be screened by in situ coupled 25 transcription-translation systems, and subsequent catalytic activity assays for the analysis of mutagenized enzymes. Libraries can be screened and cataloged by sequencing and use of the data for the analysis of cDNA
abundancies, which is useful for RNA profiling and serial analysis of gene expression (SAGE; Velculeseu eml., Science 270:484-487 (1995)).

Figure 1 is a diagram of examples of various forms of linear vectors and the circular vectors, having one continuous strand and one discontinuous strand, that result upon insertion of a nucleic acid molecule.

Figure 2 is a diagram of an example of the disclosed method where the linear vector includes an non-ligatable nick in one of its strands. Upon ligation to form a circular vector, the discontinuous strand is separated from the continuous strand by binding the biotin moiety at the nick to an immobilized streptavidin moiety.
Figure 3 is a diagram of an example of the disclosed method where the linear vector has an overlap. Upon ligation to form a circular vector (with a Y tail), the discontinuous strand is separated from the continuous strand by ligating one end of the discontinuous strand to an immobilized 10 nucleic acid probe. This ligation is mediated by hybridization between the single-stranded extension of the discontinuous strand and the probe.
Figure 4 is a diagram of an example of the disclosed method where the linear vector is a linker having sticky ends compatible with sticky ends formed by restriction digestion of PCR primer sequences incorporated at the 15 ends of PCR amplified DNA. 'The linker facilitates circularization of~the PCR amplified DNA. 'the linker has one non-ligatable end which, upon ligation, results in a circular molecule with one continuous strand and one discontinuous strand. The discontinuous strand is separated from the continuous strand by binding the biotin moiety at the nick to an immobilized 20 streptavidin moiety.
DETAILED DESCRIPTION OF THE INVENTION
Disclosed are reagents and a method for efficient in vitro molecular cloning of nucleic acid molecules of interest. Because the method is entirely in vitro, it can be automated and scaled-up in ways that are not possible in 25 cell-based molecular cloning. The method involves insertion of a nucleic acid molecule of interest in a linear vector to forni a circular vector where one strand is continuous and the other strand is discontinuous. The continuous strand of the circular vector is then amplified by rolling circle replication, amplifying the inserted nucleic acid molecule in the process.
30 The amplification is rapid and efficient since it involves a single, isothermic reaction that replicates the vector sequences exponentially. The amplification process is amenable to automation where multiple reactions are carried out simultaneously in a small area. The amplified nucleic acid can be used for any purpose and in any manner that nucleic acid cloned or amplified by known methods can be used. This includes sequencing, probing, restriction analysis, subcloning, transcription, hybridization or denaturation analysis, further amplified, and storage for future use or analysis.
~fhe insertion reaction involves insertion of a double-stranded nucleic acid molecule into a double-stranded linear vector to produce a double-stranded circular vector. The use of circular vectors facilitates the selection of molecules that have successfully incorporated inserts. The ampliFcation reaction involves rolling circle replication of a single-stranded circular 10 nucleic acid molecule. In its most useful forms, the disclosed method involves insertion of nucleic acid molecules of interest into a vector and separate amplification of the resulting recombinant vectors to produce separate nucleic acid "colonies," each representing a clonal population of nucleic acid sequences present in founding vector of that "colony."
15 A key feature of the method, which facilitates double-stranded insertion followed by single-stranded amplification, is formation of the circular vector in such a way that one of its strands is a closed circular strand (that is, continuous) while the other strand is not a closed circular strand {that is, it has a nick, a gap, an overlap, or is otherwise discontinuous). This 20 feature is most useful, and most effectively accomplished, when, by operation of the method, the closed strand and the open strand are predetermined; that is, when a particular strand of the vector is selectively left discontinuous.
With rolling circle replication, amplification takes place not in cycles, 25 but in a continuous, isothermal replication. This makes amplification less complicated and much more consistent in output. A single round of rolling circle replication results in a large amplification of the circular vector, orders of magnitude greater than a single cycle of PC1~. replication and other amplification techniques in which each cycle is limited to a doubling of the 3(> number of copies of a target sequence.
the present method is an alternative to traditional molecular cloning involving clonal amplification in cells harnessing the natural nucleic acid replication in, and growth and division of, the cells. The present method has several advantages over this traditional method. First, it is much more rapid.
Traditional molecular cloning usually requires at least twelve hours of cell growth in the best cell-based cloning methods to produce a million copies of a single vector. In contrast, the present method allows production of millions or billions of copies of a single vector in only 60 to 90 minutes.
Once a clonal culture or colony of cells is grown, there is still the problem of separating the amplified nucleic acid from the cell and all the cellular components. Although numerous methods have been devised over the years for such purification, they remain both time consuming and 10 ineffective (that is, cellular contaminants remain with the isolated nucleic acid). Generally, the amount of time for the purification and the level of~
purity obtain from nucleic acid isolation methods is proportional: more time, more purity; less time, less purity. In contrast, the present method accomplishes clonal amplification in an uncomplicated mix of just a few 15 well-defined components: the vector, one or two types of primers, nucleotides, and polymerase. The purification of the amplified nucleic acid is correspondingly simplified. Most significantly, the amplification reaction in present method need not result in a complex mixture of nucleic acids as is true of cell-based molecular cloning.
20 'fhe present method also has advantages over cyclic amplification methods such as the polymerase chain reaction (PCR). Rolling circle replication is more rapid and has higher yields than PCR. Significantly, the products of roiling circle replication arc tandemly repeated amplicons of double-stranded DNA. 'These amplicons, if carried as contaminants to any 25 surface or vessel, are unable to seed new rolling circle replication reactions.
In other words, the amplified DNA is non-contaminating because it is a replication dead-end. 13y contrast, PCR or SDA amplicons are potentially contaminating.
Following amplification, the amplified nucleic acid can be used for 30 any purpose. Numerous methods for the use and manipulation of cloned or isolated nucleic acid are known and can be applied to nucleic acid amplified in the present method. I~or example, the nucleic acid can be sequenced, probed, subjected to restriction analysis, subcloned, transcribed, subjected to hybridization or denaturation analysis, further amplified, or stored.
Diagnostic methods, such as sequencing and probing for specific sequences, are preferred.
The nucleotide sequence of the amplified sequences can be 5 determined either by conventional means or by primer extension sequencing of amplified target sequence. One preferred form of sequencing for use with amplified sequences produced with the disclosed method is nanosequencing or single-nucleotide extension sequencing. Nanosequencing methods are described below and by Jalanko et al., Clinical Chemistjy 38:39-43 (1992);
10 Nikiforov et al., Nucleic Acids Research 22:4167-4175 (1994); and Kobayashi et al., Molecular and Cellular Probes 9:175-182 (1995).
'Two forms of sequencing that can be used with the disclosed method are described in PCT Application WO 97/20948. One is single nucleotide primer extension sequencing involving interrogation of a single nucleotide in 15 an amplified target sequence by incorporation of a specific and identifiable nucleotide based on the identity of the interrogated nucleotide. The other is degenerate probe primer extension sequencing involving sequential addition of degenerate probes to an interrogation primer hybridized to amplified target sequences.
20 Libraries of cloned nucleic acids formed by the disclosed method can be screened using any of the methods used for screening conventional libraries. 1~or example, cI~NA libraries made using the disclosed method can be analyzed using conventional screens. Libraries can also be used for in situ transcription to generate RNA colonies, which can then be analyzed (in 25 situ or in replicas) by appropriate screens, such as aptamer screens or ribozyme activity screens. Libraries can also be screened by in .situ translation on array replicas (see, for' example, Saris et crl., Nucleic Acids Res.
10:4831-4843 (1982)). Libraries can also be screened by in situ coupled transcription-translation systems, and subsequent catalytic activity assays for 30 the analysis oFmutagenizcd enzymes. Libraries can be screened and cataloged by sequencing and use of the data for the analysis of cUNA
abundancies, which is useful for RNA profiling and serial anal~~sis of gene expression (SAGE: Velculescu et al., Science 270:484-487 (1995)).

One embodiment of the disclosed method is a method of isolating and amplifying a nucleic acid molecule, where the method involves:
(a) ligating a nucleic acid molecule into a linear vector to form a circular vector including the vector and the nucleic acid molecule, where the linear vector is a double-stranded linear nucleic acid including two nucleic acid strands, where the second strand of the circular vector is discontinuous, and where the first strand in the circular vector is a closed circular strand, and (b) amplifying the first strand by rolling circle replication to form tandem sequence DNA, where the amplification results in amplification of the nucleic acid molecule in the first strand.
The method can be practiced and expanded in several ways. For example, the second strand of the linear vector can contain at least one nick, where the nick cannot be ligated. The linear vector can be designed such that 15 either the 5' or the 3' end of the second strand of the linear vector cannot be ligated. 'The linear vector can be designed such that the second strand of the linear vector contains at least one gap or overlap. The method can be extended to include, following ligation and prior to amplification, separation of the first strand from the second strand. The second strand of the vector 20 can include an affinity tag. In this case, the first strand can be separated from the second strand by binding the affinity tag to a substrate, denaturing the lust and second strands prior to, simultaneous with, or following binding, and separating the first strand from the substrate.
The second strand of the linear vector can also be designed to contain 25 at least one overlap, where part of the overlapping portions of the second strand are complementary, and where the 3' end of the overlap extends beyond the part of the overlapping portions that are complementary. In this case, the first strand can be separated from the second strand by ligating one end of the second strand to a nucleic acid molecule coupled to a substrate, 30 denaturing the first and second strands following ligation of the second strand. and separating the first strand from the substrate.
The method can also be practiced such that step {a) involves ligating a plurality of nucleic acid molecules into a plurality of linear vectors in a single reaction to form a plurality of circular vectors, each circular vector containing at least one nick, gap, or overlap in the second strand, such that step (b) involves amplifying the first strand of the plurality of circular vectors, and such that, prior to amplification, the ligation reaction is divided to produce a plurality of separate amplification reactions. The method can be extended to include making a replica of the amplification reactions. The replica of the amplification reactions can be made by contacting the amplification reactions with a surface to which nucleic acids can bind. The replica of the amplification reactions can also be made by transferring part of each amplification reaction to form a replica amplification reaction.
In the method, the ligation reaction can be divided by spreading the ligation reaction onto a surface to form a spread, where the separate amplification reactions are the locations of circular vectors on the surface after spreading. In this case, a replica of the amplification reactions can be made by contacting the spread with a second surface to which nucleic acids can bind.
Any number or all of the amplification reactions can be ordered as an array of reaction droplets or in an array of reaction vessels. In this case, following amplification, all or part of the contents of any number or all the individual reaction droplets or reaction vessels are transferred by one to one mapping to a new set of reaction droplets or reaction vessels.
'rhe method can be further extended by, following amplification, determining the presence of amplified nucleic acid in the amplification reactions, and transferring all or a part of the contents of the amplification reactions containing amplified nucleic acid reaction to a new set of reaction droplets or reaction vessels.
Replicas of the amplification reactions can be made by contacting the amplification reactions with a surface treated with an affinity target capable of binding an affinity tag, where the amplified nucleic include affinity tags ~0 incorporated during amplification, such that a portion of each amplification reaction is transferred to the surface. l he affil~ity tag is preferably biotin and the affinity target is preferably streptavidin. 'the affinity tag is also preferably a reactive moiety and the affinity target is preferably a corresponding reactive moiety, such that a chemical reaction between the affinity tag and the affinity target results in the amplified nucleic acid being covalently coupled to the surface. In this case, it is preferred that the affinity target is phenylene diisothiocyanate, disuccinimidylcarbonate, disuccinimidyloxolate or dimethylsuberimidate and the affinity tag is a reactme amore.
A replica of the amplification reactions can also be made by transferring part of each amplification reaction to form a replica amplification reaction. Following amplification, all or part of the contents of l 0 any number or all of the reaction droplets or reaction vessels can be transferred and combined to create one or more sets of pooled reactions. The amplification reactions can also be arranged on the surface of a substrate.
Preferred substrates include acrylamide, cellulose, nitrocellulose, polystyrene, polyethylene vinyl acetate, polypropylene, polymethacrylate, 15 polyethylene, polyethylene oxide, glass, polysilicates, polycarbonates, teflon, fluorocarbons, nylon, silicon rubber, polyanhydrides, polyglycolic acid, polylactic acid, polyorthoesters, polypropylfumerate, collagen, glycosaminoglycans, polyamino acids, chemical resistant metals, or corrosion resistant metals.
20 The ligation reactions can also be diluted prior to division of the ligation reaction into amplification reactions such that, on average, each amplification reaction contains a single circular vector. A sample of each amplification reaction can also be collected. Nucleic acid molecules in the amplification reactions or in the collected samples can also be detected or 2~ sequenced.
In the disclosed method, rolling circle replication can be primed by the second strand on the circular vector or by a rolling circle replication primer. The tandem sequence DNA formed after amplification can itself be amplified by strand displacement replication to form secondary tandem 30 sequence DNA. This secondary tandem sequence DNA can also be amplified by strand displacement replication to form tertiary tandem sequence DNA. Strand displacement replication of the tandem sequence can be primed by a strand displacement primer.

~hhe method can also include detection of one or more amplified nucleic acid molecules in one or more of the amplification reactions. In a preferred embodiment, the nucleic acid molecules can be derived from cDNA generated by suppression subtractive hybridization. In another 5 preferred embodiment, the plurality of nucleic acid molecules can be all derived from the same source.
Detection can be accomplished by, following amplification, creating a replica of the amplification reactions, contacting the amplification reactions with a first set of labeled nucleic acid probes and the replica amplification reactions with a second set of labeled nucleic acid probes, and comparing the pattern of hybridization of the first set of probes to the pattern of hybridization of the second set of probes, where differences in the patterns of hybridization indicate differences in the probe sets. Following detection amplification reactions that hybridize to the first set of probes but not to the I 5 second set of probes, amplification reactions that hybridize to the second set of probes but not to the first set of probes, amplification reactions that hybridize to the both sets of probes, or amplification reactions that do not hybridize to either set of probes can be selected for isolation or further analysis.
20 Another embodiment of the disclosed method is an in vitro method of cloning nucleic acid molecules, where the method involves {a) dividing a nucleic acid sample to produce a plurality of separate amplification reactions, (b) amplifying nucleic acid molecules in the amplification reactions, 25 (c) making a replica of the amplification reactions, (d) testing nucleic acid molecules in either the amplification reactions or the replica amplification reactions to identify nucleic acid molecules of interest, and (e) retrieving the identified nucleic acid molecules of~ interest from 30 the corresponding amplification reactions or replica amplification reactions that were not tested.
In this embodiment, the nucleic acid sample can be divided by spreading the sample onto a surface to form a spread, such that the separate amplification reactions are the locations of circular vectors on the surface after spreading. The replica of the amplif canon reactions can be made by contacting the spread with a second surface to which nucleic acids can bind.
Another embodiment of the disclosed method is a method of isolating and amplifying nucleic acid molecules, where the method involves (a) ligating a plurality of nucleic acid molecules into a plurality of linear vectors in a single reaction to form a plurality of circular vectors, each circular vector including a vector and a nucleic acid molecule, where the linear vectors are double-stranded linear nucleic acid comprising two nucleic acid strands, where the circular vectors each contain at least one nick, gap, or overlap in the second strand, and where the first strand in each circular vector is a closed circular strand, (b) separating the first strands from the second strands, (c) diluting and dividing the first strands to produce a plurality of separate amplification reactions that, on average, each contain a single circular vector, (d) amplifying the first strands of the plurality of circular vectors by rolling circle replication to form tandem sequence DNA, where the amplification results in amplification of the nucleic acid molecules in the first strands.
~hhe tandem sequence DNA formed after amplification can itself be amplified by strand displacement replication to form secondary tandem sequence DNA. This secondary tandem sequence DNA can also be amplified by strand displacement replication to form tertiary tandem sequence DNA.
Another embodiment of the disclosed method is a method of isolating and amplifying a nucleic acid molecule, where the method involves (a) ligating a nucleic acid molecule into a linear vector to form a circular vector comprising the vector and the nucleic acid molecule, where the linear vector is a double-stranded linear nucleic acid comprising two nucleic acid strands, where the circular vector contains at least one nick in the second strand, and whcue the first strand in the circular vector is a closed circular strand, (b) amplifying the first strand, where the amplification results in amplification of the nucleic acid molecule in the first strand.
Also disclosed is a kit for isolating and amplifying nucleic acid molecules, where the kit includes (a) a linear vector where the linear vector is a double-stranded linear nucleic acid comprising two nucleic acid strands, and where ( 1 ) the linear vector contains at least one nick, where the nick cannot be ligated, (2) either the 5' or the 3' end of the second strand of the linear vector cannot be ligated, (3) the second strand of the linear vector contains at least one gap, (4) the second strand of the linear vector contains at least one overlap, or (5) any combination of (1), (2), (3) or (4);
(b) a rolling circle replication primer, where the rolling circle replication primer is complementary to a portion of the f rst strand of the linear vector; and (c) a strand displacement primer, where the strand displacement primer matches a portion of the first strand of the linear vector.
Also disclosed is a linear vector where the linear vector is a double-stranded linear nucleic acid made up of two nucleic acid strands, where the second strand of the linear vector contains an affinity tag, and where (1 ) the linear vector contains at least one nick, where the nick cannot be ligated, (2) either the 5' or the 3' end of the second strand of the linear vector cannot be ligated, (3) the second strand of the linear vector contains at least one gap, (4) the second strand of the linear vector contains at least one overlap, or (~) any combination of (1), (2), (3) or (4).

Also disclosed is a linear vector where the linear vector is a double-stranded linear nucleic acid made up of two nucleic acid strands, where the second strand of the linear vector contains at least one overlap, part of the overlapping portions of the second strand are complementary, and the 3' end 5 of the overlap extends beyond the part of the overlapping portions that are complementary.
I. Materials 'I"hc disclosed method makes use of linear vectors in which nucleic acid molecules of interest can be inserted to form circular vectors. The circular vectors contain one continuous strand and one discontinuous strand.
The discontinuous strand may include an affinity tag which, by interaction with an affinity substrate, can facilitate separation of the continuous strand from the discontinuous strand. The continuous strand of the circular vector is amplified by rolling circle replication to form tandem sequence DNA (TS-I 5 DNA). Rolling circle replication is primed by a rolling circle replication primer complementary to a sequence in the continuous strand. 'hhe tandem sequence DNA itself may be amplified by strand displacement replication, to form secondary tandem sequence DNA, using strand displacement primers (complementary to a sequence in the tandem sequence DNA). The secondary tandem sequence DNA may also be amplified by strand displacement replication, to form tertiary tandem sequence DNA, using strand displacement primers or the rolling circle replication primers (complementary to a sequence in the secondary tandem sequence DNA).
Multiple amplification reactions can be carried out in parallel, preferably in arrays or as spreads of diluted vectors on surfaces or embedded in agarose.
The resulting "colonies" of amplified DNA represent molecular clones of the progenitor circular vectors with an inserted nucleic acid molecule.
Collectively, such colonies form a library of cloned nucleic acid molecules that can be replica plated or arrayed, stored. and screened. These materials are described in detail below.
A. Nucleic Acid Molecules The disclosed method can be used to clone or amplify any nucleic acid molecule of interest. The nucleic acid molecules can come from any l4 source such as a cellular or tissue nucleic acid sample, a subclone of a previously cloned fragment, mRNA, chemically synthesized nucleic acid, genomic nucleic acid samples, nucleic acid molecules obtained from nucleic acid libraries, specific nucleic acid molecules, and mixtures of nucleic acid 5 molecules. The disclosed method is particularly suited to producing libraries of cloned nucleic acid molecules starting with a complex mixture of nucleic acid molecules to be represented in the library. For example, cDNA can be produced from all of the mRNA in a cellular sample and used to make a cDNA library, on a library of genomic DNA can be produced from a genomic nucleic acid sample.
In the method, the nucleic acid molecule is inserted into a double-stranded linear vector. Preferably the insertion is accomplished by ligation, although any suitable coupling mechanism can also be used. Thus, the only requirement for nucleic acids molecules to be used in the disclosed method is that they can be coupled to the ends of a double-stranded nucleic acid molecule (that is, the linear vector). Single-stranded nucleic acid molecules, such as RNA, can be used by converting the molecule to be double-stranded.
In the case of RNA molecules, this can be accomplished, for example, by producing a cDNA molecule of the RNA. Numerous methods are known for preparing and inserting nucleic acid molecules into vectors and any of these can be used to prepare nucleic acid molecules for use in the disclosed method {see, for example, Sambrook el ul., Molecular Cloning: A Laboratory Manual, 2nd Edition (Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y., 1989)). Preferably, the nucleic acid molecule is prepared by generating sticky ends to facilitate insertion in the linear vector. This can be accomplished, for example by cleaving a nucleic acid molecule of interest, or a nucleic acid sample, with a restriction enzyme, or by adding linkers to the ends of nucleic acid molecules of interest that have, or can be processed to have sticky ends. One or both of the ends of the nucleic acid molecule can 30 also be left blunt ended. 'I~hc two ends of nucleic acid molecules to be used in the disclosed method can also be made different to allow directional insertion. For example, the to ends can have different sticky ends. or have one sticky end and one blunt end.
1~

B. Linear Vectors Linear vectors for use in the disclosed method are double-stranded nucleic acid molecules that can be circularized when its ends are coupled to a nucleic acid molecule. 'The characteristics of the linear vector are limited 5 only by the requirements for the circular vector that results upon insertion of the nucleic acid molecule. Thus, the linear vector is designed such that, when the nucleic acid molecule is inserted and the linear vector is circularized, the resulting circular vector has one continuous, circular strand and one discontinuous strand.
The use of the term vector is not meant to indicate that the linear vector is required to have any characteristics beyond these, such as promoters, selectable markers, origins of replication, and other features present on traditional vectors for cloning in cells. I-lowever, the linear vector may contain these or any other features that do not interfere with the 15 disclosed method. Such additional characteristics may be useful, for example, to allow transfer of the vector to cells to obtain expression, or to allow in vitro expression. 'Thus, the linear vector can be as simple as a short linker that facilitates circularization of a nucleic acid molecule of interest.
The linear vector is a double-stranded nucleic acid molecule where one of the strands will become part of the continuous strand of the circular vector and the other strand will become part of the discontinuous strand of the circular vector. For identification, the strand of the linear vector which will become part of the continuous strand of the circular vector is referred to as the first strand of the linear vector, and the strand of the linear vector which will become part of the discontinuous strand of the circular vector is referred to as the second strand of the linear vector. The first strand of a linear vector includes a sequence complementary to a rolling circle replication primer. This sequence is referred to as the primer complement portion of the linear vector. This facilitates amplification of the circular 30 vector formed from the linear vector by rolling circle replication primed by the rolling circle replication primer. n separate primer complement portion is not required ii the second strand of the circular vector is to serve as the rolling circle replication primer. It is preferred that the primer complement portion of the linear vector is near the 5' end of the first strand of the linear vector.
The production of a circular vector in the disclosed method is facilitated by giving the first strand and the second strand of the linear vector 5 different characteristics. Although the linear vector is referred to as having two strands, each of these "strands" may be made up of more than one linear nucleic acid strands. That is, the first strand of the linear nucleic acid molecule may be made up of multiple nucleic acid molecules lying end to end and hybridized to the second strand. Similarly, the second strand of the 10 linear nucleic acid molecule may be made up of multiple strands hybridized to the first strand. The use of the terms first "strand" and second "strand"
are used as a convenience to refer to all of the physical nucleic acid strands that make up one side of the linear vector. The relationship of the physical strands in the linear vector to the collective first and seconds strands can be 15 seen in Figure 1. The first strand of the linear vector is preferably composed of one strand. The second strand is preferably composed of more than one strand, and most preferably composed of two strands.
All of the ends present in the first strand of t:he linear vector, including internal ends, if present, should be ligatable. Ligatable ends are 20 elldS that can be ligated to compatible ends by ligase, or which can otherwise be coupled to compatible ends. Preferred ligatable ends arc nucleotides having a 3' hydroxyl or a 5' phosphate. Internal ends are ligatable only if compatible ends are adjacent. For example, a nick with a 3' hydroxyl on one end and a ~' phosphate on the other end is a ligatable nick and the ends are 25 ligatable. Nick has its usual meaning. Specifically, a nick is a break in a strand hybridized to another strand where there are no unpaired nucleotides in the other strand opposite the nick.
'I~o result in a discontinuous second strand in the circular vector, the second strand of the linear vector should contain at least one non-ligatable 30 end or at least one gap or overlap. Non-ligatable ends are ends that cannot be ligated to compatible ends by ligase, or which cannot otherwise he coupled to compatible ends. Preferred non-ligatable ends are nucleotides having a blocking group at the 3' or 5' position. For example, the second strand of the linear vector can include a 3'-terminal or 5'-terminal biotin residue (either at the end of a continuous second strand or at a nick in a discontinuous second strand). This residue renders the terminus non-ligatable, causing all vectors to contain a nick after cloning of inserts by ligation. This biotin residue can 5 then used as a handle to remove the second strand of the circularized vector, generating single-stranded circles for amplification. 'thus, the biotin is both a blocking group and an affinity tag.
Internal ends are also non-ligatable if, for example, compatible ends are not adjacent. For example, a nick with a 3' hydroxyl on both ends is an 10 unligatablc nick and the ends are unligatable. A nick with a blocking group on one of the ends is also an unligatable nick and the end with the blocking group is an unligatable end. Both gaps and overlaps is not ligatable even if the ends would otherwise be compatible since the ends are not close enough to be coupled. Gap has its usual meaning. Specifically, a gap is a break in a 15 strand hybridized to another strand where there is at least one nucleotide on the other strand opposite the gap that is unpaired. A gap can also occur at the end of the linear vector in that a nucleic acid molecule, when hybridized to sticky ends of the linear vector, can fail to extend to the end of one of the strands of the linear vector.
20 An overlap occurs where the adjacent ends of two strands hybridized adjacent to each other on another strand extend beyond the region of hybridization. A preferred form of overlap is where the two overlapping strands hybridize to each other in the overlapping region. This type of overlap in a linear vector produces a Y shaped molecule such as the one 25 illustrated in Figure 1 and in Figure 3.
The second strand of the linear vector may contain multiple nicks, gaps, and overlaps in any combination. Any number of such nicks may be ligatable or non-ligatable. Ail that is required is at least one feature that prevents the second strand of the circular vector from being continuous 30 following insertion of the nucleic acid molecule. For this purpose, a single non-ligatable end or other non-ligatable feature is all that is required. The first strand oh the linear vector contain multiple nicks so long as they are all is ligatable; that is, so long as the first strand of the circular vector will be continuous following insertion of the nucleic acid molecule.
The second strand of the linear vector can also contain one or more affinity tags to facilitate separation of the first and second strands of the 5 circular vector formed from the linear vector. It is preferred that linear vectors include either pre-formed sticky ends or one or more restriction enzymes sites near the ends of the linear vector to facilitate insertion of nucleic acid molecules into the vectors. Multiple cloning sites (MCS) are particularly preferred. Such MCSs facilitate both insertion of a nucleic acid 10 molecule of interest into linear vectors and removal of the nucleic acid molecule from the amplified nucleic acid. It is preferred that the ends of the linear vector, when ready for ligation, do not contain compatible ends that can be ligated. This will prevent the circularization of linear vectors in the absence of insertion of a nucleic acid molecule.
15 C. Circular Vectors A circular vector is a double-stranded circular nucleic acid molecule that is a combination of a linear vector and one or more inserted nucleic acid molecules. One of the strands of the circular vector, termed the first strand, is continuous. That is, the first strand of the circular vector is a closed 20 circular nucleic acid strand. 'hhe other strand of the circular vector, termed the second strand, is discontinuous. That is, the second strand of the circular vector is not a closed circular nucleic acid strand. The second strand can include, for example, nicks, gaps, and overlaps. The discontinuity of the second strand al lows the separation of the f rst and second strands following 25 denaturation. The second strand of the circular vector can also contain one or more affinity tags to facilitate separation of the first and second strands of the circular vector.
'('he first strand of a circular vector includes a sequence complementary to a rolling circle replication primer. This sequence is 30 referred to as the primer complement portion of the circular vector. This facilitates amplification of the circular vector by rolling circle replication primed by the rolling circle replication primer. A separate primer complement portion is not required if the second strand of the circular vector is to serve as the rolling circle replication primer.
D. Affinity Tags An affinity tag is a molecule that interacts specifically with a particular molecule or moiety. The molecule or moiety that interacts specifically with an affinity tag is referred to herein as an affinity target.
Together, an affinity tag and affinity target make up a binding pair. Either member of a binding pair can be used as an affinity tag and either member can be used as an affinity target. An affinity tag is the member of the binding pair coupled to the linear or circular vector. A preferred binding pair is biotin and streptavidin. It is to be understood that the term affinity target refers to both separate molecules and to portions of molecules, such as an epitope of a protein, that interacts specifically with an affinity tag.
Antibodies, either member of a receptor/ligand pair, and other molecules I 5 with specific binding affinities are examples of affinity tags, useful as the affinity portion of a reporter binding molecule. By coupling an affinity tag to the second strand of a linear vector, binding of the affinity tag to its affinity target allows separation of the first and second strands of the circular vector. An affinity tag that interacts specifically with a particular afCnity target is said to be specific for that affinity target. For example, an affinity tag which is an antibody that binds to a particular antigen is said to be specific for that antigen. The antigen is the affinity target. Complementary nucleotide sequences can be used as binding pairs. An example of this is illustrated with the immobilization of Y shaped circular vector in Figure 3.
I:. Affinity Substrates Affinity substrates are solid-state substrates or supports to which affinity targets have been coupled. Generally, an affinity substrate is used to facilitate separation of first and second strands of circular vectors by immobilizing the second strands of circular vectors to a solid-state substrate 30 or support via an affinity tag. Solid-state substrates for use in affinity substrates can include any solid material to which aft7nity targets can be coupled. This includes materials such as acrylamide, cellulose.
nitrocellulose, polystyrene, polyethylene vinyl acetate, polypropylene, polymethacrylate, polyethylene, polyethylene oxide, glass, polysilicates, polycarbonates, teflon, fluorocarbons, nylon, silicon rubber, polyanhydrides, polyglycolic acid, polylactic acid, polyorthoesters, polypropylfumerate, collagen, glycosaminoglycans, polyamino acids, chemical resistant metals, 5 and corrosion resistant metals. Solid-state substrates can have any useful form including thin films or membranes, beads, bottles, dishes, fibers, woven f bers, shaped polymers, particles and microparticlcs. A preferred form for a solid-state substrate is a bead or surface.
Affinity targets immobilized on a solid-state substrate allow capture of the second strand of circular vectors on a affinity substrate. Such capture provides a convenient means of separating the seconds strands from the first strands of circular vectors (which are to be amplified).
Methods for immobilizing proteins, such as antibodies, to solid-state substrates are well established. Immobilization can be accomplished by attachment, for example, to aminated surfaces, carboxylated surfaces or hydroxylated surfaces using standard immobilization chemistries. Examples of attachment agents are cyanogen bromide, succinimide, aldehydes, tosyl chloride, avidin-biotin, photocrosslinkable agents, epoxides and rnaleimides.
A preferred attachment agent is glutaraldehyde. These and other attachment agents, as well as methods for their use in attachment, are described in Protein immobilization: fundamentals and applications, Richard F. 'haylor, ed. (M. Dekker, New York, 1991 ), Johnstone and Thorpe, Immunochemistry In Practice (Blackwell Scientif c Publications, Oxford, England, 1987) pages 209-216 and 241-242, and Immobilized f1 ffinity Lid>crnds, Craig T.
Hermanson et al., eds. (Academic Press, New York, 1992). Proteins, and other affinity targets having free amino groups, can be attached to a substrate by chemically cross-linking a free amino group on the protein to reactive side groups present within the solid-state substrate. For example, proteins may be chemically cross-linked to a substrate that contains free amino or 30 carboxyl groups using glutaraldehyde or carbodiimides as cross-linker agents. In this method, aqueous solutions containing free proteins are incubated with the solid-state substrate in the presence of glutaraldehyde or carbodiimide. For crosslinking with glutaraldehyde the reactants can be incubated with 2% glutaraldehyde by volume in a buffered solution such as 0.1 M sodium cacodylate at pl-I 7.4. Other standard immobilization chemistries are known by those of skill in the art.
Methods for immobilization of oligonucleotides to solid-state substrates are well established. Oligonucleotides, including affinity targets, can be coupled to substrates using established coupling methods. For example, suitable attachment methods are described by Pease et al., Proc.
Natl. Acad. Sci. USA 91 ( I 1 ):5022-5026 ( 1994), and Khrapko et al., Mol Biol (Mosk) (USSR) 25:718-730 ( 1991 ). A method for immobilization of 3'-amine oligonucleotides on casein-coated slides is described by Stimpson et al., Proc. Natl. Acad. Sci. USA 92:6379-6383 (1995). A preferred method of attaching oligonucleotides to solid-state substrates is described by Guo et al., Nucleic Acids Res. 22:5456-5465 ( 1994).
F. Rolling Circle Replication Primer A rolling circle replication primer (RCRP) is an oligonucleotide having sequence complementary to the primer complement portion of the first strand of a circular vector. This sequence is referred to as the complementary portion of the RCRP. The complementary portion of a RCRP and the cognate primer complement portion can have any desired sequence so long as they arc complementary to each other. In general, the sequence of the RCRP can be chosen such that it is not significantly complementary to any other portion of the circular vector. The complementary portion of a rolling circle replication primer can be any length that supports specific and stable hybridization between the primer and the primer complement portion. Generally this is 10 to 35 nucleotides long, but is preferably 16 to 20 nucleotides long. A separate rolling circle replication primer is not required if the second strand of the circular vector is to serve as the rolling circle replication primer. In this case, the second strand of the circular vector can be referred to as a rolling circle replication primer.
It is preferred that rolling circle replication primers also contain additional sequence at the ~' end of the RCRP that is not complementary to any part of the circular vector. rhhis sequence is referred to as the non-complementary portion of the RCRP. The non-complementary portion of the RCRP, if present, serves to facilitate strand displacement during DNA
replication. The non-complementary portion of a RCIZP may be any length, but is generally 1 to 100 nucleotides long, and preferably 4 to 8 nucleotides long. The rolling circle replication primer may also include modified nucleotides to make it resistant to exonuclease digestion. For example, the primer can have three or four phosphorothioate linkages between nucleotides at the 5' end of the primer. Such nuclease resistant primers allow selective degradation of excess unligated linear vector that might otherwise interfere with hybridization of probes and primers to the amplified nucleic acid. A
rolling circle replication primer can be used as the tertiary strand displacement primer in strand displacement cascade amplification.
G. Strand Displacement Primers Primers used for strand displacement replication are referred to herein as strand displacement primers. One form of strand displacement primer, referred to herein as a secondary strand displacement primer, is an oligonucleotide having sequence matching part of the sequence of the first strand of a circular vector. This sequence is referred to as the matching portion of the strand displacement primer. This matching portion of a secondary strand displacement primer is complementary to sequences in tandem sequence DNA (TS-DNA). The matching portion of a secondary strand displacement primer may be complementary to any sequence in TS-DNA. However, it is preferred that it not be complementary 'hS-DNA
sequence matching either the rolling circle replication primer or a tertiary strand displacement primer, if one is being used. 'this prevents hybridization of the primers to each other. The matching portion of a strand displacement primer may be complementary to all or a portion of the inserted nucleic acid molecule, although this is not preferred. The matching portion of a strand displacement primer can be any length that supports specific and stable hybridization between the primer and its complement. Generally this is 12 to nucleotides long, but is preferably 18 to 25 nucleotides long. It is preferred that the matching portion of the circular vector is near the 3' end of the first strand of the circular vector.

It is preferred that secondary strand displacement primers also contain additional sequence at their 5' end that does not match any part of the first strand of the circular vector. This sequence is referred to as the non-matching portion of the strand displacement primer. The non-matching portion of the strand displacement primer, if present, serves to facilitate strand displacement during DNA replication. The non-matching portion of a strand displacement primer may be any length, but is generally I to 100 nucleotides long, and preferably 4 to 8 nucleotides long.
Another form of strand displacement primer, referred to herein as a tertiary strand displacement primer, is an oligonucleotide having sequence complementary to part of the sequence of the first strand of the circular vector. This sequence is referred to as the complementary portion of the tertiary strand displacement primer. This complementary portion of the tertiary strand displacement primer matches sequences in rfS-DNA. The complementary portion of a tertiary strand displacement primer may be complementary to any sequence in the first strand of the circular vector.
E-Iowever, it is preferred that it not be complementary to a sequence matching the strand displacement primer. This prevents hybridization of the primers to each other. The complementary portion of a tertiary strand displacement primer can be any length that supports specific and stable hybridization between the primer and its complement. Generally this is 12 to 3~
nucleotides long, but is preferably 18 to 25 nucleotides long. It is preferred that tertiary strand displacement primers also contain additional sequence at their 5' end that is not complementary to any part of the first strand of the circular vector. 'This sequence is referred to as the non-complementary portion of the tertiary strand displacement primer. 'hhe non-complementary portion of the tertiary strand displacement primer, if present, serves to facilitate strand displacement during DNA replication. The non-complemeniary portion of a tertiary strand displacement primer may be any length, but is generally l to 100 nucleotides long, and preferably 4 to 8 nucleotides long. A rolling circle replication primer is a preferred form of tertiary strand displacement primer. It is preferred that the complementary portion of the circular vector is near the 5' end of the first strand of the circular vector.
Strand displacement primers may also include modified nucleotides to make them resistant to exonuclease digestion. for example, the primer 5 can have three or four phosphorothioate linkages between nucleotides at the 5' end of the primer. Such nuclease resistant primers allow selective degradation of excess unligated linear vectors that might otherwise interfere with hybridization of probes and primers to the amplified nucleic acid.
Strand displacement primers can be used for strand displacement replication 10 and strand displacement cascade amplification, both described below.
H. Synthesis of Oligonucleotides Linear vectors, rolling circle replication primers, strand displacement primers, and any other oligonucleotides can be synthesized using established oligonucleotide synthesis methods. Methods to produce or synthesize 15 oligonucleotidcs are well known in the art. Such methods can range from standard enzymatic digestion followed by nucleotide fragment isolation (see for example, Sambrook et crl., Molecular Cloning: A Laboratory Manual, 2nd Edition (Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y., 1989) Chapters 5, 6) to purely synthetic methods, for example, by the 20 cyanoethyl phosphoramidite method using a Milligen or Beckman System 1 Plus DNA synthesizer (for example, Model 8700 automated synthesizer of Milligen-Biosearch, Burlington, MA or ABI Model 380B). Synthetic methods useful for making oligonucleotides are also described by Ikuta et al., Ann. Rev. Biochem. 53:323-356 (1984), (phosphotriester and phosphite-25 triester methods), and Narang et al., Methods Enzymol. 65:610-620 ( 1980), (phosphotriester method). Protein nucleic acid molecules can be made using known methods such as those described by Nielsen et al., Bioconjug. Chem.
5:3-7 ( I 994).
Many of~the oligonucleotides described herein are designed to be 30 complementary to ecrtain portions of other oligonucleotides or nucleic acids such that stable hybrids can be formed between them. The stability of these hybrids can be calculated using known methods such as those described in Lesnick and rrcier, Biochemistry 34:10807-10815 (1995), McGraw et al., Biotechniques 8:674-678 ( I 990), and Rychlik et al., Nucleic Acids Res.
18:6409-6412 ( 1990).
I. DNA Iigases Any DNA ligase is suitable for use in the disclosed method.
Preferred ligases are those that preferentially form phosphodiester bonds at nicks in double-stranded DNA. That is, ligases that: fail to ligate the free ends of single-stranded DNA at a significant rate are preferred.
'rhermostable ligases are especially preferred. Many suitable ligases are known, such as T4 DNA ligase (Davis et al., Advanced Bacterial Genetics -.A Manual for Genetic Engineering (Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y., 1980)), l~'. coli DNA ligase (Panasnko et al., J. Biol.
Chem. 253:4590-4592 (1978)), AMPLIGASEO (Kalin et al., Mutat. Res., 283(2):119-123 (1992); Winn-Decn et al., Mol Cell Probes (England) 7(3):179-186 (1993)), Taq DNA ligase (F3arany, Proc. Natl. Acid. Sci. USA
88:189-193 (1991), Thermus thermophilus DNA ligase (Abbott Laboratories), Thermus scotodarctus DNA ligase and Rhodothermzrs marinus DNA ligase (Thorbjarnardottir et al., Gene 151:177-180 (1995)). T4 DNA
ligase is preferred for ligations involving RNA target sequences due to its ability to ligate DNA ends involved in DNA:RNA hybrids (Hsuih et al., Quantitative detection of IIC'V RNA using novel ligation-dependent polymerise chain reaction. American Association for the Study of Liver Diseases (Chicago, IL, November 3-7, 1995)).
J. DNA polymerises DNA polymerises useful in rolling circle replication must perform rolling circle replication of primed single-stranded circles. Such polymerises are referred to herein as rolling circle DNA polymerises. For rolling circle replication, it is preferred that a DNA polymerise be capable of displacing the strand complementary to the template strand, termed strand displacement, and lack a 5' to 3' exonuclease activity. Strand displacement is necessary to result in synthesis of multiple tandem copies of the circular vector. A 5' to 3' exonuclease activity, if present, nuight result in the destruction of the synthesised strand. It is also preferred that DNA
polymerises for use in the disclosed method are hi~;hly processive. The suitability of a DNA polymerase for use in the disclosed method can be readily determined by assessing its ability to carry out rolling circle replication. Preferred rolling circle DNA polymei'ases are bacteriophage X29 DNA polymerase (U.S. Patent Nos. 5,198,543 and 5,001,050 to Blanco et al.), phage M2 DNA polymerase (Matsumoto et al., Gene 84:247 ( 1989)), phage ~PRD1 DNA polymerase (Jung et al., Proc. Natl. Acad. Sci. USA
84:8287 (1987)), VENT~ DNA polymerase (Kong et al., J Biol. Chem.
268:1965-1975 (1993)), Klenow fragment of DNA polymerase I (Jacobsen et al., Eur. J. Biochem. 45:623-627 (1974)), TS DNA polymerase (Chatterjee et 10 al., Gene 97:13-19 (1991)), PRD1 DNA polymerase (Zhu and Ito, l3iochim.
Biophys. Acta. 1219:267-276 (1994)), modified T7 DNA polymerase (Tabor and Richardson, J. Biol. Chem. 262:15330-I5333 (1987); Tabor and Richardson, J. l3iol. Chem. 264:6447-6458 (1989); SequenaseTM (U.S.
Biochemicals)), and T4 DNA polymerase holoenzyme (Kaboord and IS Benkovic, Curr. Biol. 5:149-157 (1995)). X29 DNA polymerase is most preferred. Rolling circle DNA polymerases are also generally useful for strand displacement replication.
Strand displacement can be facilitated through the use of a strand displacement factor, such as helicasc. It is considered that any DNA
20 polymerase that can perform rolling circle replication in the presence of a strand displacement factor is suitable for use in the disclosed method, even if the DNA polymerase does not perform rolling circle replication in the absence of such a factor. Strand displacement factc>rs useful in RCA include BMRFI polymerase accessory subunit (Tsurumi et crl., J. virology 25 67(12):7648-7653 (1993)), adenovirus DNA-binding protein (Zijderveld and van der Vliet, J. Virology 68(2):1158-1 164 (1994)), herpes simplex viral protein ICPB (Boehmer and Lehman, J. I~irology 67(2):711-715 ( 1993);
Skaliter and Lehman, Proc. Natl. Acad. S'ci. LISA 91 (22):10665-10669 (1994)), single-stranded DNA binding proteins (SSB; Rigler and Romano, J.
30 l3iol. Chem. 270:8910-8919 (1995)). and calf thymus hclicase (Siegel et al., J Biol. C:hem. 267:13629-13635 (1992)).
The ability of a polymcrase to can-y out rolling circle replication can be determined by using the polymerase in a rolling circle replication assay such as those described in Fire and Xu, Proc. Natl. Acad. Sci. USA 92:4641-4645 (1995).
It is possible to enhance the specificity of the DNA amplification reactions used in the disclosed method by using a DNA polymerase that is 5 inactive at low temperature, and active only at high temperature. An example of such an enzyme, AmpliTaq Gold, has been described by Moretti et al., Biotechniques 25:716-722 (1998). AmpliTaq Gold is inactive until heated during the PCR before thermal cycling. A similar enzyme could be used in the disclosed method. Temperature activation of DNA polymerase can also be achieved using antibodies specific for the polymerase. For example, antibodies specific for Bst large fragment DNA polymerase could be obtained by immunization of mice. Among such antibodies, one could be chosen on the basis of its ability to bind to and inhibit the enzyme at room temperature. The antibody could also be chosen, using known screening 15 procedures, such that upon heating, the inhibition of the DNA polymerase would cease. Combining the antibody with Bst large fragment DNA
polymerase would generate an enzyme mixture that is activated upon heating.
K. Kits Any combination of the materials useful in the disclosed method can be packaged together as a kit for performing the disclosed method. In particular, linear vectors, rolling circle replication primers, affinity substrates, and strand displacement primers are useful components of~such kits. Enzymes necessary for the disclosed method are also preferred components of such kits.
L. Tandem Sequence DNA
The first strand of the circular vector, when replicated, gives rise to a long DNA molecule containing multiple repeats of sequences complementary to the circular vector. This long DNA molecule is referred to herein as tandem sequences I~NA (TS-DNA). TS-DNA contains sequences complementary to the inserted nucleic acid molecule and the primer complement portion. If the tandem sequence DNA. is itself replicated by strand displacement amplification, the resulting long DNA molecules containing multiple repeats of sequences matching the circular vector are referred to as secondary tandem sequence DNA. If the secondary tandem sequence DNA is in turn replicated by strand displacement amplification, the resulting long DNA molecules containing multiple repeats of sequences 5 complementary to the circular vector are referred to as tertiary tandem sequence DNA.
M. Collected Samples (Library Replica) The usefulness of the disclosed method is increased by producing libraries of clones and saving samples of the clones for later use. Such samples are referred to a collected samples. Collecting samples is analogous to replica plating in cell-based cloning. Samples oJ' amplified nucleic acid can be collected, for example, by transfer with an array of pins (most useful when the nucleic acid is amplified in an array pattern), by transfer into an array, by direct transfer from a spread of amplified nucleic acid on a surface to another surface (this is analogous to colony transfer), and by blotting the amplified nucleic acid unto a membrane (most useful when the nucleic acid is amplified in agarose). Once the samples are collected, they can be further amplified to allow analysis or use of the clones, or to allow another round of replica collection.
II. Method The disclosed method involves inserting nucleic acid molecules of interest into a linear vector to form a circular vector with one continuous strand and one discontinuous strand. rl'he discontinuous strand may include an affinity tag which, by interaction with an affinity substrate, can facilitate separation of the continuous strand from the discontinuous strand. The continuous strand of the circular vector' is amplified by rolling circle replication to form tandem sequence DNA. Rolling circle replication is primed by a rolling circle replication primer complementary to a sequence in 30 the continuous strand. 'Chc tandem sequence DNA itself may be amplified by strand displacement replication, to form second<~ry tandem sequence DNA, using strand displacement primers (complementary to a sequence in the tandem sequence DNA). (he secondary tandem sequence DNA may also be amplified by strand displacement replication, to form tertiary tandem sequence DNA, using strand displacement primers or the rolling circle replication primers (complementary to a sequence :in the secondary tandem sequence DNA). The amplified DNA can be seque:need, probed, subjected 5 to restriction analysis, subcloned, transcribed, subjected to hybridization or denaturation analysis, further amplified, or stored.
Multiple amplification reactions can be carried out in parallel, preferably in arrays or as spreads of diluted vectors on surfaces or embedded in agarose. The resulting "colonies" of amplified DNA represent molecular 10 clones of the progenitor circular vectors with an inserted nucleic acid molecule. Collectively, such colonies form a library of cloned nucleic acid molecules that can be replica plated or arrayed, stored, and screened. These procedures are described in detail below.
When using the disclosed method to produce a library of cDNA
15 molecules, or to analyze mKNA in a sample via cDNA, the cDNA
preparations used for cloning in the disclosed vectors can be prepared using methods that reduce the over-representation of cDNA that corresponds to highly abundant messenger RNA. Libraries made using such methods are called normalized libraries (Bonaldo et al., Genome Res 6:791-806 (1996)).
20 The use of normalized libraries reduces the number of clones that must be screened to find a sequence of interest.
A. Ligation Ligation of nucleic acid molecules into linear vectors can be accomplished using any suitable conditions. Techniques for insertion of 25 nucleic acid molecules into vectors in general are well established and can be used with the disclosed linear vectors. Suitable ligases for the ligation operation are described above. Ligation reactions can involve a single type of linear vector and a single type of nucleic acid molecule to be inserted, a single type of linear vector and multiple different types of nucleic acid 30 molecules to be inserted, multiple types of linear vector and a single type of nucleic acid molecule to be inserted, or multiple types of linear vectors and multiple types of nucleic acid molecules to be inserted. For general cloning and production of nucleic acid libraries it is prefewed that a single type of linear vector and multiple different types of nucleic acid molecules to be inserted be used. For subcloning of specific nucleic acid fragments it is preferred that a single type of linear vector and a single type of nucleic acid molecule to be inserted. Ligation conditions are generally known. Most S ligases require Mgi ~. There are two main types of ligases, those that are ATP-dependent and those that are NAD-dependent. ATP or NAD, depending on the type of ligase, should be present during ligation.
Ligation of compatible ends of nucleic acid molecules and vectors can be facilitated through the use of blunt ends or sticky ends as is known in t:he field of molecular cloning. Both blunt ends and sticky ends can be produced by digestion of the nucleic acid molecules and the linear vectors with appropriate restriction enzymes, by ligation of appropriate linkers to the ends of the nucleic acid molecules and the linear vectors, or both. In the case of linear vectors, appropriate ends can be formed directly by the structure of the linear vector without the need for restriction enzyme digestion of linker ligation. In the case of the nucleic acid molecules to be inserted, appropriate ends can be appended to the ends during preparation of the nucleic acid molecule. For example, appropriate ends can be incorporated into cDNA by using primer having appropriate sequences during cDNA synthesis or by adding a nucleotide tail to the cDNA.
Q. Amplification The circular vectors formed by ligation of linear vectors and nucleic acid molecules of interest serve as substrates for a rolling circle replication.
'This reaction requires the addition of two reagents: (a) a rolling circle replication primer, which is complementary to the primer complement portion of the first strand of the circular vector, and (b) a rolling circle DNA
polymerase. The DNA polymerase catalyzes primer extension and strand displacement in a processive rolling circle polymerization reaction that proceeds as long as desired, generating a molecule of up to 100,000 nucleotides or larger that contains up to approximately 25 tandem copies of a sequence complementary to a 4000 by circular vector. This tandem sequence DNA (TS-DNA) consists of alternating vector sequence and insert sequence. As an alternative, the second strand of the circular vector can serve as the rolling circle replication primer.
During rolling circle replication one may additionally include radioactive, or modified nucleotides such as bromodeoxyuridine triphosphate, in order to label the DNA generated in the reaction.
Alternatively, one may include suitable precursors that provide a binding moiety such as biotinylated nucleotides (Langer et al., Proc. Natl. Acad. Sci.
USA 78:6633 (1981)).
Strand displacement replication is a way to amplify TS-DNA. Strand displacement replication is accomplished by hybridizing strand displacement primers to 1'S-DNA and allowing a DNA polymerase to synthesize DNA
from these primed sites. The product of strand displacement replication is referred to as secondary tandem sequence DNA or TS-DNA-2. Strand displacement replication can be accomplished by performing rolling circle I S replication to produce TS-DNA, and then mixing strand displacement primer with the 'CS-DNA and incubating to replicate the tandem sequence DNA.
'The strand displacement primer is complementary to a part of the circular vector used to generated 'I'S-DNA as described earlier. It is preferred that the strand displacement primer is not complementary to the rolling circle replication primer, or to a tertiary strand displacement primer, if used.
Strand displacement replication can also be carried out simultaneously with rolling circle replication. This is accomplished by mixing strand displacement primer with the circular vector and rolling circle replication primer prior t.o incubating the mixture for rolling circle replication. For simultaneous rolling circle replication and strand displacement replication, it is preferred that the rolling circle DNA
polymerise be used for both replications. 'This allows optimum conditions to be used and results in displacement of other strands being synthesized downstream. Generally. strand displacement replication can be performed by, simultaneous with or following rolling circle replication, mixing a strand displacement primer with the TS-DNA and incubating to replicate the tandem sequence DNA to result in the formation of secondary tandem sequence DNA.

To optimize the efficiency of strand displacement replication, it is preferred that a sufficient concentration of strand displacement primer be used to obtain sufficiently rapid priming of the growing TS-DNA strand to out-compete any remaining unligated linear vectors that might be present for binding to 1"S-DNA. In general, this is accomplished when the strand displacement primer is in very large excess compared to the concentration of single-stranded sites for hybridization of the strand displacement primer on TS-DNA. Optimization of the concentration of strand displacement primer can be aided by analysis of hybridization kinetics using methods such as those described by Young and Anderson, "Quantitative analysis of solution hybridization" in Nucleic Acid Hybridization: A Practical Approach (IRL
Press, 1985) pages 47-71. Alternatively, the efficiency of strand displacement replication can be improved by the removal of unligated linear vectors prior to amplification of the TS-DNA. In strand displacement replication, it is preferred that the concentration of strand displacement primer generally be from 500 nM to 5000 nM, and most preferably from 700 nM to 1000 nM.
As a strand displacement primer is elongated, the DNA polymerise will run into the 5' end of the next hybridized strand displacement molecule and will displace its 5' end. In this fashion a tandem queue of elongating DNA polymerises is formed on the TS-DNA template. As long as the rolling circle reaction continues, new strand displacement primers and new DNA polymerises are added to TS-DNA at the growing end of the rolling circle.
When strand displacement replication is carried out in the presence of a tertiary strand displacement primer, an exponential amplification of TS-DNA sequences takes place. This special and preferred mode of strand displacement replication is referred to as strand displacement cascade amplification (SDCA). In SDCA, a strand displacement primer primes replication of TS-DNA to form TS-DNA-2, as described above. The tertiary strand displacement primer can then hybridize to. and prime replication of, TS-DNA-2 to form TS-DNA-3 (tertiary tandem sequence DNA). Strand displacement of TS-DNA-3 by the adjacent, growing '1'S-DNA-3 strands makes TS-DNA-3 available for hybridization with secondary strand displacement primer. This results in another round of replication resulting in TS-DNA-4 (which is equivalent to TS-DNA-2). TS-DNA-4, in turn, becomes a template for DNA replication primed by tertiary strand displacement primer. The cascade continues this manner until the reaction stops or reagents become limiting. This reaction amplifies DNA at an almost exponential rate, although kinetics are not truly exponential because there are stochastically distributed priming failures, as well as steric hindrance events related to the large size of the DNA network produced during the reaction.
In a preferred mode of SDCA, the rolling circle replication primer serves as the tertiary strand displacement primer, thus eliminating the need for a separate primer. For this mode, the rolling circle replication primer should be used at a concentration sufficiently high to obtain rapid priming on the growing TS-DNA-2 strands. To optimize the efficiency of SDCA, it is preferred that a sufficient concentration of secondary strand displacement primer and tertiary strand displacement primer be used to obtain sufficiently rapid priming of the growing 'rS-DNA strand to out-compete TS-DNA for binding to its complementary TS-DNA, and, in the case of secondary strand displacement primer, to out-compete any remaining unligated linear vector that might be present for binding to TS-DNA. In general, this is accomplished when the secondary strand displacement primer and tertiary strand displacement primer are both in very large excess compared to the concentration of single-stranded sites for hybridization of the strand displacement primers on TS-DNA. For example, it is preferred that the secondary strand displacement primer is in excess compared to the concentration of single-.stranded secondary strand displacement primer complement sites on TS-DNA, 'rS-DNA-3, TS-DNA-S, and so on. In the case of tertiary strand displacement primer, it is preferred that the tertiary strand displacement primer is in excess compared to the concentration of single-stranded tertiary strand displacement primer complement sites on TS-DNA-2, TS-DNA-4, TS-DNA-6, and so on. Such an excess generally results in a primer hybridizing to its complement in TS-DNA before amplified complementary ~I'S-DNA can hybridize. Optimization of primer concentrations can be aided by analysis of hybridization kinetics (Young and Anderson). In a strand displacement cascade amplification, it is preferred that the concentration of both secondary and tertiary strand displacement primers generally be from 500 nM to 5000 nM, and most preferably from 700 nM to 1000 nM.
As in the case of secondary strand displacement primers, if the concentration of DNA polymerase is sufficiently high, the polymerase will initiate DNA synthesis at each available 3' terminus on the hybridized tertiary strand displacement primers, and these elongating TS-DNA-3 molecules will block any hybridization by 1'S-DNA-2. As a tertiary strand displacement primer is elongated to form TS-DNA-3, the DNA polymerase will run into the 5' end of the next hybridized tertiary strand displacement primer molecule and will displace its 5' end. In this fashion a tandem queue of elongating DNA polymcrases is formed on the T'S-DNA-2 template. As long as the reaction continues, new rolling circle replication primers and new DNA polymerases are added to 'I'S-DNA-2 at the growing ends of hS-DNA-2. This hybridization/replication/strand displacement cycle is repeated with hybridization of secondary strand displacement primers on the growing TS-DNA-3 .
Generally, strand displacement cascade amplification can be performed by, simultaneous with, or following, rolling circle replication, mixing a secondary strand displacement primer and a tertiary strand displacement primer with the TS-DNA and incubating to replicate the tandem sequence DNA -- where replication of the tandem sequence DNA
results in the formation of secondary tandem sequence DNA and where replication of the secondary tandem sequence DNA results in formation of tertiary tandem sequence DNA (TS-DNA-3).
Strand displacement replication can also be carried out sequentially.
Following a first round of strand displacement replication, a tertiary strand displacement primer can be mixed with the TS-DNA and 7'S-DNA-2 and incubated to replicate the secondary tandem sequence DNA, where replication of the secondary tandem sequence DNA results in formation of~
tertiary tandem sequence DNA (~I~S-DNA-3). This round of strand displacement replication can be referred to as tertiary strand displacement replication. However, all rounds of strand displacement replication following rolling circle replication can also be referred to collectively as strand displacement replication.
A modified form of strand displacement replication results in amplification of TS-DNA and is referred to as opposite strand amplification (OSA). OSA is the same as strand displacement replication except that a special form of rolling circle replication primer is used that prevents it from hybridizing to TS-DNA-2. This can be accomplished in a number of ways.
For example, the rolling circle replication primer can have an affinity tag coupled to its non-complementary portion allowing the rolling circle replication primer to be removed prior to strand displacement replication.
Alternatively, remaining rolling circle replication primer can be crippled following initiation of rolling circle replication. One preferred form of rolling circle replication primer for use in OSA is designed to form a hairpin that contains a stem of perfectly base-paired nucleotides. The stem can contain 5 to 12 base pairs, most preferably 6 to 9 base pairs. Such a hairpin-forming rolling circle replication primer is a poor primer at lower temperature (less than 40°C) because the hairpin structure prevents it from hybridizing to complementary sequences. The stem should involve a sufficient number of nucleotides in the complementary portion of the rolling circle replication primer to interfere with hybridization of the primer to the circular vector. Generally, it is preferred that a stern involve 5 to 24 nucleotides, and most preferably 6 to 18 nucleotides, of the complementary portion of a rolling circle replication primer. A rolling circle replication primer where half of the stem involves nucleotides in the complementary portion of the rolling circle replication primer and the other half of the stem involves nucleotides in the non-complementary portion of the rolling circle replication primer is most preferred. Such an arrangement eliminates the need for self complementary regions in the circular vector when using a hairpin-forming rolling circle replication primer.
I f an excess of tertiary tandem sequence DNA is desired, the secondary strand displacement primer can be crippled in the same manner as is described above for the rolling circle replication primer (the rolling circle replication primer and tertiary strand displacement primer should not be crippled in this case). The reaction at the higher, permissive temperature should be carried out long enough to produce a reasonable amount of S secondary tandem sequence DNA to serve as a template for tertiary sequence DNA. When the temperature is shifted, the secondary strand displacement primer can no longer prime synthesis and the synthesis of tertiary tandem sequence DNA soon outstrips the amount of secondary tandem sequence DNA. Of course tandem sequence DNA will continue to be produced by rolling circle replication throughout the reaction (since the rolling circle replication primer is not crippled).
When starting the rolling circle replication reaction, secondary strand displacement primer and rolling circle replication primer are added to the reaction mixture, and the solution is incubated briefly at a temperature I S sufficient to disrupt the hairpin structure of the rolling circle replication primer but to still allow hybridization to the primer complement portion of the circular vector (typically greater than SO°C). This incubation permits the rolling circle replication primer to hybridize to the primer complement portion of the circular vector. The solution is then brought to the proper temperature for rolling circle replication, and the rolling circle DNA
polymerase is added. As the rolling circle reaction proceeds, TS-DNA is generated, and as the 'fS-DNA grows in length, the secondary strand displacement primer rapidly initiates DNA synthesis with multiple strand displacement reactions on TS-DNA. These reactions generate TS-DNA-2, 2S which is complementary to the TS-DNA. While TS-DNA-2 contains sequences complementary to the rolling circle replication primer, the primer is not able to hybridize nor prime efficiently at the reaction temperature due to its hairpin structure at this temperature. Thus, there is no further priming by the rolling circle replication primer and the only products generated are TS-DNA and TS-DNA-2. The reaction comes to a halt as rolling circle amplification stops and TS-DNA becomes completely double-stranded. In the course of the reaction, an excess of single-stranded TS-DNA-2 is generated.

Another form of rolling circle replication primer useful in OSA is a chimera of DNA and RNA. In this embodiment, the rolling circle primer has deoxyribonucleotides at its 3' end and ribonucleotides in the remainder of the primer. It is preferred that the rolling circle replication primer have Fve or 5 six deoxyribonucleotides at its 3' end. By making part of the rolling circle replication primer with ribonucleotide, the primer can be selectively degraded by RNAse H when it is hybridized to DNA. Such hybrids form during OSA as TS-DNA-2 is synthesized. The deoxyribonucleotides at the 3' end allow the rolling circle DNA polymerase to initiate rolling circle 10 replication. RNAse H can then be added to the OSA reaction to prevent priming of TS-DNA-2 replication.
Unligated linear vectors may be removed prior to rolling circle replication to eliminate competition between unligated linear vectors and the secondary strand displacement primer for hybridization to ~I~S-DNA.
15 Alternatively, the concentration of the secondary strand displacement primer can be made sufficiently high so that it out-competes unligated linear vector for hybridization to TS-DNA. This allows strand displacement replication to be performed without removal of unligated linear vectors.
C. Separation 20 Once the linear vector and nucleic acid molecule are circularized to form a circular vector with one continuous strand and one discontinuous strand, it is preferred that the continuous strand of the circular vector (a single-stranded, closed circular nucleic acid molecule susceptible to rolling circle replication) is separated from the discontinuous strand of the circular 25 vector. Separation allows rolling circle replication to proceed more efficiently.
It is preferred that the two strands of the circular vector are separated by immobilizing the discontinuous (second) strand of the circular vector and then denaturing and washing away the continuous (i'~rst) strand of the 30 circular vector. This can be accomplished, for example, by including an affinity tag in the second strand of the circular vector which can then be bound to an immobilized affinity target, thus immobilizing the second strand of the circular vector. For this purpose, use of biotin as an affinity tag and streptavidin as an affinity target is preferred.
Complementary oligonucleotides can also be used as binding pairs for separating the first and second strands of the circular vector. An oligonucleotide affinity tag that is a part of or coupled to the second strand of the circular vector is hybridized to an immobilized complementary oligonucleotide. The oligonucleotide affinity tag is preferably an unhybridized tail in an overlap in the second strand of the circular vector (see Figure 3). It is preferred that the oligonucleotide affinity tag be ligated or otherwise coupled to a solid-state substrate or support to keep the second strand immobilized during denaturation. This can be accomplished, for example, by using a circular vector with a staggered tail in the second strand and an immobilized oligonucleotide that can hybridize to the longer strand of the tail such that the end of the immobilized oligonucleotide can be ligated to the shorter strand of the staggered tail. An example of such an arrangement is shown in Figure 3.
The first and second strands of the circular vector can be denatured using any suitable means. Preferred conditions for denaturation include the use of heat, alkaline conditions, chaotropic conditions, and combinations.
These and other means of nucleic acid denaturation are known and can be used in the disclosed method. Washing and collection of the first strand is performed during or after denaturation.
As an example, the first and second strands of circular can be separated using a two-step procedure. First, the circular vector, which contains a biotin residue on the second strand, is bound to beads containing streptavidin, in order to bind the vector via the biotin. The beads are then washed with formamidc at mildly alkaline pH. Under appropriate conditions, the circular vector, which contains an unligated nick site by design, separates into two DNA molecules. Thus, alkaline denaturation releases free single-stranded circles from the beads. 'rhe single-stranded circular molecules (that is, the first strands of the circular vector) are then further purified by gel filtration or ion exchange (Mono-Q 5/5) chromatography in the presence of an alkaline buffer (15 mM Na01-I). This purification step will remove small circles, and small linear vector molecules that contaminate the circular vectors with inserts.
D. Removing Linear Nucleic Acids Unligated linear nucleic acids, including unligated linear vector, can be removed prior to rolling circle replication. In addition to methods described elsewhere herein, the gene 6 exonuclease of phage T7 provides a useful tool for the elimination of linear nucleic acids that might bind to the TS-DNA. This exonuciease digests DNA starting from the 5' end of a double-stranded structure. It has been used successfully for the generation of single-stranded DNA alter PCR amplification (Holloway et al., Nucleic Acids Res. 21:3905-3906 (1993); Nikiforov et al., ~'C.'R Methods and Applications 3:285-291(1994)). This enzyme can be added after ligation, together with the rolling circle DNA polymerase. 'To protect TS-DNA from degradation, the rolling circle replication primer can contain 3 or 4 phosphorothioate linkages at the 5' end, to make this molecule resistant to the exonuclease (Nikiforov et al. (1994)). The exonucleasc will degrade unprotected linear molecules as they become associated with the rolling circle DNA product.
E. Dilution and Division ~hhe disclosed method is particularly useful for creation of a library of cloned nucleic acid molecules. For this purpose it is useful to dilute and divide ligated vectors containing inserts to separate individual circular vectors and allow production of clonal "colonies" of nucleic acid amplified from a single circular vector. It is preferred that a solution containing circular vectors is diluted sufficiently such that amplification reactions contain, on average, a single circular vector. This can be accomplished, for example, by making a range of dilutions, dividing the diluted vector solutions among reactions, and performing amplif ic:ation. fhe dilution is considered optimal when about 33% of the reactions produce amplified nucleic acid. This follows from well known distribution statistics.
The division of the diluted circular vectors c;an be accomplished in several ways. For example, the diluted solution can be spotted as an array on a surface. The diluted solution can also be spread on a surface with the circular vectors becoming separated. 'hhe diluted solution can also be mixed with agarose and spread on a surface. A preferred way to divide the diluted circular vectors is to use a microarray of very small liquid droplets on a glass surface. The micro-droplet arrays contain the diluted circular vectors, and 5 enable the generation of DNA clones in very small compartments, without the need for physical barriers such as tubes or wells.
One useful way to combine division of the ligation reaction with amplification is to spread the ligation reaction (diluted or not as appropriate) on a surface and then to spot amplification reaction components (that is, buffers. reagents, polymcrase) on the surface in an array. Techniques for spotting are described, for example, in U.S. Patent No. 5,807,522 to Brown et al. The method here differs, however, in that it is the reaction components, rather than individual samples, that are spotted. All of the spotted material is identical. The amplification reagents will then cause 15 amplification of whatever nucleic acid is present at the location of each spot of reagents. The nucleic acid on the surface comes from the spread of the ligation reaction.
As an example of spreading, the ligated DNA can be diluted serially to obtain concentrations in the range of several million molecules per milliliter. Approximately 22 ~I of this DNA solution (containing approximately 30,000 to 120,000 DNA molecules) is placed on a cover slip, and covered with a polylysine-coated microscope slide. 'fhe DNA is allowed to bind the polylysine-covered surface for 30 minutes at 37°C. The slide then dipped in 0.01 % Tween-20, and dried at room temperature. Using an arraying instrument, an array is constructed consisting of 6000 individual micro-droplets of a solution containing the four dideoxynucleoside triphosphates, a suitable buffer, and enzymes for amplification. The diameter of the droplets can be approximately 0.150 millimeters. Droplets preferably are dispensed on the surface of the slide in a controlled humidity 30 atmosphere. in order to maintain a constant droplet volume during amplification. Alternatively. the diluted circular vectors may be placed on the glass surface using the arrayer instrument to dispense small volumes that on the average contain a single ligatcd DNA molecule. It can be calculated t 1 that when 33% of the droplets grow molecular colonies. and the rest do not grow anything. most colonies are likely to be of clonal origin. Adjusting the initial inoculum densit~~, it should be possible to obtain up to 1500 clonal colonies per 6000-droplet array.
As an example of agarose spreading. dilutions of the vectors can be mixed with a buffer containing melted agarose at 60°C and overlaying the solution on a petri dish to form a thin agarose layer (0.2 to 2% agarose).
such that the concentration of the vectors is in the range of 500 to 5.000 per plate.
The embedded vectors can then be amplified using the disclosed method. At appropriate dilutions, DNA molecular colonies clonally derived from single vectors will form in the thin film of agarose. The initial density of seed DNA molecules should be such that the DNA molecular colonies do not overlap. One useful technique of agarose spreading is describe in U.S.
Patent No. 5.616.478 to Chctverin et al.
F. Sample Collection (Replica Elating) The usefulness of the disclosed method is increased by producing libraries of clones and saving samples of the clones for later use. Such samples are referred to as collected samples. Collecting samples is analogous to replica plating in cell-based cloning. Samples of amplified nucleic acid can be collected, for example. by transfer with an array of pins (most useful when the nucleic acid is amplified in an array pattern). by transfer into an array, by direct transfer from a spread of amplified nucleic acid on a surface to another surface (this is analogous to colony transfer).
and by blotting the amplified nucleic acid unto a membrane (most useful when 25 the nucleic acid is amplified in agarose). Once the samples are collected, they can be further amplified to allow analysis or use of the clones, or to allow another round of replica collection.
Where droplet arrays are used, the molecular colonies (that is, the droplets following amplification) can be replicated by contacting the array with a multi-pin replicator that will bind only a fraction of the volume of the micro-droplet. Colony replicas may be stored by blotting the replicator on a membrane such as nitrocellulose of NA-45 (S&S). A preferred way to store replicas is to contact the replieator with a polylysine-coated class slide.

which will permit hybridization or primer extension sequencing with fluorescent probes. Replicas of reaction droplets on a surface can also be made by contacting a second surface with the droplets. In general, it is preferred that the second surface contact only the droplets and not the first surface. Replicas of spreads of nucleic acids on a surface likewise can be made by contacting a second surface with the spread.
Replicas of amplified nucleic acid bound covalently to glass can be made using any suitable coupling procedure. For example, in order to facilitate a subsequent step of covalent binding to a glass surface, each of the two strand displacement replication primers used for the strand displacement cascade amplification of circular vectors may be synthesized with a primary amino group at the 5' end. At the end of the amplification reaction, the glass slide will contain thousands of liquid droplets harboring DNA clones, and all the DNA molecules will contain 5'-terminal reactive amino groups. At this I 5 point the glass slide can be contacted with another glass slide, leaving an air gap of less than 1 mm (defined by the thickness of a plastic spacer), in such a manner that the glass slide on top will contact all of the liquid droplets without excessive compression. The lower face of this slide (the upper slide) can be derivatized using the methods described by Guo et crl., Nucleic Acids Research 22:5456-5465 (1994), Guo et al., Nature Biotechnology 15:331-335 (1997)), Guo et al., Nucleic Acids Res 22:5456-5465 (1994), or Beier and Hoheisel, Nucleic Acids Res 27:1970-1977 (1999), to make the glass surface chemically reactive. The two-slide sandwich is incubated for 1 to 2 hours at 37°C (or as appropriate for the derivative chemistry involved) 25 in order to obtain covalent coupling of the amplified DNA contained in each droplet to the lower face of the upper slide.
Where agarose has been used to form molecular colonies, analysis of the colonies can be facilitated by blotting the amplified nucleic acid unto a membrane or other blotting surface. Many such blotting techniques are 30 known and can be used with the disclosed method. roe example, the agarose flm can be placed in a vacuum-blotting device that contacts both the bottom and the top of the agarose film. The nucleic acid on the agarose film can then be vacuum-blotted to two membranes simultaneously, one placed on top 4>

of the agarose and the other below the agarose, to generate two replicas of the molecular colonies on the surface of the membrane.
G. Detecting Amplified Nucleic Acid Molecules The amplified nucleic acid can be used for any purpose for which nucleic acids can be used. For example, the nucleic acid can be sequenced, probed, subjected to restriction analysis, subcloned, transcribed, subjected to hybridization or denaturation analysis, further amplified, or stored.
Diagnostic methods, such as sequencing and probing for specific sequences, are preferred. For these purposes, the amplified nucleic acid can be analyzed using standard molecular biology procedures, such restriction enzyme digestion, cloning in a plasmid vector, PCR amplification, which are well known.
Libraries of cloned nucleic acids formed by the disclosed method can be screened using any of the methods used for screening conventional libraries. For example, cDNA libraries made using the disclosed method can be analyzed using conventional screens. Libraries can also be used for zn ,sitar transcription to generate RNA colonies, which can then be analyzed (in situ or in replicas) by appropriate screens, such as aptamer screens or ribozyme activity screens. Libraries can also be screened by in situ translation on array replicas (see, for example, Saris et al., Nucleic Acids Res.
10:4831-4843 (1982)). Libraries can also be screened by in situ coupled transcription-translation systems, and subsequent catalytic activity assays for the analysis of mutagenized enzymes.
The disclosed method can also be used for serial analysis of gene expression (SAGE), making it more efficient by streamlining the cloning and sequencing into a single process stream. This method involves amplification of eDNA inserted into linear vectors as described herein prior to the SAGE
analysis. This means of amplification is useful since PCR amplification of t:he eDNA prior to cloning, which can skew the abundance of cDNA
sequences due to differential amplification, is avoided. 'I~he disclosed method insures that sequence tag frequencies in the clone population (which are measured in SAGE) reflect the original Ircqucncies of the cDNAs.

WO 00/15779 P~T/US99/21291 'fhe method of Welford et al., Nucleic Acids Res. 26(12):3059-3065 ( 1998), in which thousands of colonies produced using laborious traditional procedures were analyzed in an array, can be modified to make use of the disclosed method and thereby become more streamlined and efficient. The method could even be automated. Detection of differences between nucleic acid samples or probe sets can be accomplished by adapting the technique described by George et al., Nucleic Acids Research 27:1517-1523 (1999), to the disclosed method. George et al. describes combination of suppression subtractive hybridization (SSH) and cDNA microarrays for rapid identification of differentially expressed genes. In this method, a set of cDNA clones, including inserrts amplified by PCR, is arrayed using robotic printing. The cDNA arrays can then be hybridized with fluorescent labeled probes prepared from RNA obtained from a cell line or tissue of interest.
H. Sequencing Amplified Nucleic Acid Molecules The amplified nucleic acid can be sequenced using any suitable procedure. Many such procedures are known. One preferred form of sequencing for use with amplified sequences produced with the disclosed method is nanosequencin g or single-nucleotide extension sequencing.
Nanosequencing methods are described below and by Jalanko et al., Clinical Chemistry 38:39-43 ( 1992); Nikiforov et al., Nucleic Acids Research 22:4167-4175 (1994); and Kobayashi et al., Molecular and Cellular Probes 9:175-182 (1995).
Two forms of primer extension sequencing that can be used with the disclosed method are described in PC1' Application WO 97/20948. One is single nucleotide primer extension sequencing involving interrogation of a single nucleotide in an amplified target sequence by incorporation of a specific and identifiable nucleotide based on the identity of the interrogated nucleotide. The other is degenerate probe primer extension sequencing involving sequential addition of degenerate probes to an interrogation primer hybridized to amplified target sequences.
Nanosequencing operations can be performed in batch. For example, if the slide contains 3000 clots. all 3000 dots are sequenced in a single batch operation. 'l~his can be accomplished by washing the slide with 1% ammonia after imaging of the first primer extension reaction. This alkaline solution denatures the labeled primer, but the cloned DNA remains on the slide because it is bound covalently. The subsequent primer extension reactions are performed, imaged, and washed with ammonia at each step until all five primers have been extended and the fluorescence incorporated by the primer has been imaged at each step.
The next step consists of determining a very short stretch of nucleotide sequence in the amplified DNA in each replica on the nitrocellulose. This entails sequencing just nine bases in a clone (which are 10 referred to as the sequence tag of the clone). In this example, this is accomplished by two separate sets of sequencing reactions taking place in each replica-membrane. One set of sequencing reactions will determine the first five bases in the upper strand of the clone. The other set of sequencing reactions will determine the first five bases in the lower strand of the clone 15 (always reading 5' to 3') as shown by the underlined X's below.
»»>
5'-NNNNNNNNNNNNNNXXXXXXXXXNNNNNNNNNNNNNNNNN-3' 3'-NNNNNNNNNNNNNNXXXXXXXXXNNNNNNNNNNNNNNNNN-5' ccccc 20 Each stretch of five bases is interrogated by using a mixture of specific primers for each base to be sequenced, using a single addition of a dideoxynucleotide triphosphate (ddNTP). The 3' end of the first primer (Primer 1) is positioned,just before the first base to be sequenced. ~hhe sequence of the primer is defined by complementarity to vector sequences 25 flanking the insert (or non-variable sequences flanking the region to be sequenced). The 3' end of the second primer (Primer 2) is positioned just before the second base to be sequenced. The sequence of the Primer 2 is defined by complementarity to the Ranking sequences, but the last base at the 3' end is degenerate. A example design for Primer I, Primer 2, and 30 subsequent primers is shown below. The letter N indicates interrogation bases in a clone. The letter D indicates a degenerate base position in the primer. The question mark ('?) indicates the nucleotide to be added to the primer.

Cloned sequence (5EQ ID N0:1):
TAAGTCTAGTTGACAGGATGCATGNNNNNNNNNtcagacagttgttgactgatggctg ATTCAGATCAACTGTCCTACGTACNNNNNNNNNagtctgtcaacaactgactaccgac S
Primer 1 (complexity = 1) (SEQ ID NOa:2 and 1) TCTAGTTGACAGGATGCATG?
ATTCAGATCAACTGTCCTACGTAC_NNNNNNNNNagtctgtcaacaactgactaccgac l~
Primer 2 (complexity = 9) (SEQ ID NOs:3 and 1) CTAGTTGACAGGATGCATGD?
ATTCAGATCAACTGTCCTACGTACNNNNNNNNNagtctgtcaacaactgactaccgac Primer 3 (complexity = 16) (SEQ ID NOs:9 and 1) TAGTTGACAGGATGCATGDD?
20 ATTCAGATCAACTGTCCTACG"'ACNNNNNNNNNagtctgtcaacaactgactaccgac Primer 4 (complexity = 69) (SEQ ID NOs:S and .7.) AGTTGACAGGATGCATGDDD?
ZS ATTCAGATCAACTGTCCTACGTACNNNNNNNNNagtctgtc;aacaactgactaccgac Primer 5 (complexity = 256) (SEQ ID NOs:6 and 1) GTTGACAGGATGCATGDDDD?
3~ ATTCAGA'I'CAACTGTCCTACGTACNNNNNNNNNagtctgtc:aacaactgactaccgac While primer 1 is not degenerate, primer 2 contains one degenerate position, primer 3 contains two degenerate positions, primer 4 contains three degenerate positions, and primer S contains four degenerate positions.
35 Although primers 4 and 5 may prime at incorrect positions, the low complexity of the amplified DNA in a DNA colony produced by the disclosed method tends to ensure correct priming reads, on the average.
Primer extension is carried out for S minutes at 38°C in a primer extension solution containing Primer 1 as the only primer. 'hhe primer 40 extension mixture preferably contains a thermostable DNA polymerise such as Taq polymerise, and a mixture of four fluorescent dideoxy-oligodeoxyribonuclcotides, each labeled with a different dye, as in standard fluorescent sequencing (Perkin Elmer-Applied Biosystems,Inc.). Because only dideoxynucleotides are present in each colony replica, and because the 45 colony contains millions of~eopies of the nucleic acid sequence of interest, the added fluorescent label will be easily detectable for each reaction. After primer extension, the slide is washed to remove excess fluorescent ddNTPs, and imaged in a suitable fluorescence-imaging instrument capable of discriminating the four colors of the four different fluorescent dideoxy-oligodeoxyribonucleotides. Each DNA "colony" will light-up in a color corresponding to the base present at the interrogated position in each clone.
The procedure outlined above is then repeated another four times using subsequent primer sets (Primer 2, followed by Primer 3, and so on) in order to obtain the sequence at the next four positions. Signals are identified by coordinates of the clone, and the bases are ordered. The same procedure is carried out with the membrane replica, except using primers designed for sequencing the five bases in the sequence tag on the complementary strand.
The use of the fifth primer (Primer 5) may be optimized, if required by performing a pre-hybridization and washing prior to primer extension. By making multiple replicas of a molecular clone array or spread, the entire primer extension sequencing procedure can be carried out in parallel by using ten replicas, five of which are used for primer extension in one direction, and five for primer extension in the opposite direction.
The order of the bases in a sequenced segment (that is, the sequence tag) can be used to identify each of the clones. The number of possible sequence segments containing nine bases (that is, the number of different sequence tags) is 262,144. It is thus desirable to use five different linear vectors, each containing a different restriction enzyme site or sticky ends.
The use of five different linear vectors will increase the total number of possible sequence tags to 1,310,720. With this number of different sequence tags, it may be possible to identify uniquely up to 50,000 different mRNAs.
Thus, as many as 50,000 expressed sequence tags (EST) may be distinguishable on the basis of their unique sequence using the disclosed method.
When the method is used for the sequencing of larger inserts, the situation is as follows:
»>;.>
'-NNNNNNNNNNNNNNXXXx:_XXXXXX..//..XXXXXXXXXXNNNNNNNNNNNPJNNNNN-3' ''-NNNNNNNNN~!NNNNXXXx:XXXXXX..//..XXXXXXXXXXNNNNNNNNNNNNNNNNN-5' ««<

The short five-base sequences on each end, together with the sequences of flanking restriction enzyme sites, are sufficient to serve the function of unique tagging of each cDNA clone using the procedures described above.
Replica-binding of amplified nucleic acid molecules to glass slides, as described elsewhere herein, enables sequencing using a single slide. This is accomplished by washing the slide with 1% ammonia after imaging of the first primer extension reaction. This alkaline solution denatures the labeled primer, but the cloned nucleic acid remains on the slide because it is bound covalently. rfhe subsequent primer extension reactions are performed, 10 imaged, and washed with ammonia at each step until all five primers have been extended and the fluorescence incorporated by the primer has been imaged at each step.
An alternative method to read the output of nanosequencing reactions is to use mass spectroscopy instead of fluorescence. The use of mass 15 spectroscopy for sequence identification of primers that have been extended by only one base has been described by ~Iaff and Smirnov, (~enome Research 7: 378-388 (1997).
The sequencing scheme shown above can permit the sequencing of a population of cDNA molecules derived from a single type of mRNA
20 molecules. This can be accomplished, for example, as follows. First, a specific mRNA is amplified from any biological source using RT-PCR to obtain full-length amplification products. The ampliCed PCR product may have been derived from a mixture of wild-type sequence transcripts and also a small proportion (1/100, for example) of mutant transcripts that contain a 25 single point mutation at a specific locus. Any other DNA fragment can also be used.
The fCR product is then nicked with DNAse I to generate a random population of DNA fragments. The DNA is nick-translated with Kienow DNA polymerasc to generate a population of DNA fragments in the range of 30 120 to 200 nucleotides. This population of subsequcnces generated from the population of cDNA is then cloned into linear vectors to Form circular vectors and the strands of the circular vectors separated as descrihed herein.
Vectors with inserts in tine size range of 120 to 200 base pairs can be isolated (preferably before strand separation) by gel electrophoresis, or, preferably, by chromatography in Mono-Q 5/5/ (Pharmacia-LKB). The circular vectors are then amplified as described herein and the sequence of the ends of the inserts (which are thus the sequence tags) is determined as described above.
5 The sequence tags obtained from each clone consist of two pentamer (or even two hexamer) sequences. These sequence tags, which are known to be separated by segments of 120 to 200 bases, arc catalogued and assembled into a contiguous sequence using techniques developed for hybridization sequencing. Starting with a cDNA product of a single type of mRNA, the 10 cDNA can be entirely sequenced by assembling a catalog of sequenced clones. The sequence obtained from each clone is a pair of non-adjacent pentamers or hexamers. When a large number of molecular clones are analyzed, the method can reveal the presence of point mutations, even if they are present as I /100th of the cDNA population.
15 Starting with a complex mixture of cDNAs, small DNA segments (sequence tags) present in clones that originated from individual cDNA
molecules in the cDNA population can be sequenced in situ using a similar procedure. 'The method can be scaled up by increasing the density of molecular colonies, and the number of colony replicas.
20 I. Illustrations of the Method The disclosed method is further illustrated lay the following examples.
Illustration I: Cloning using a linker A DNA sample is amplified by PCR using Standard procedures, 2s except that both oligonucleotide primers are designed to contain unique restriction enzyme sites, such that after amplification the PCR product may be cleaved, generating different sticky ends on each side of the linear DNA
product. One of the PCR primers additionally contains a spacer sequence.
The digested PCR product is then placed in a ligaticm mixture containing 30 linkers designed to circularize the amplified DNA. 'hhe linkers represent the linear vector. The linkers are designed with a chemically modified terminus in one of the oligonucleotides, such that after ligation the resulting circular DNA molecules (that is circular vectors) will contain a single nick (or several nicks, if more than one linker is incorporated by ligation) in one of the strands. The modifier group may be a biotin, and it may be located either at the 5' or the 3' end of one of the linker oligonucleotides.
One of the PCR primers can also contain additional non-priming sequence designed to constitute the small spacer or backbone of the circular vector to allow amplification by rolling circle replication. The spacer sequence preferably contains a site for a rare-cutter restriction enzyme, which can be used to regenerate circles from linear DNA produced by amplification.
Optionally, the continuous strand of the circular vector can be separated from the discontinuous strand and unligated vector pieces using a two-step procedure. First, the circular vector, which contains a biotin residue, is bound to beads containing streptavidin, in order to bind the vector via the biotin present in one of the DNA strands that comprise the circular vector. The beads are then washed with formamide at mildly alkaline pH.
Under appropriate conditions, the circularized DN,!1, which contains an unligated nick site by design, separates into two DNA molecules. Thus, mild alkaline-formamide denaturation releases free single-stranded circles from the beads. The single-stranded circular molecules are then further purified by gel filtration or ion exchange (Mono-Q 5/5} chromatography in the presence of an alkaline buffer ( 15 mM NaOH). This purification step will remove small linear molecules that contaminate the circular vector (which contain inserted nucleic acid molecules). The purpose of the purification procedure is useful for selecting certain DNA sire classes, because this is 25 desirable in certain applications. This separation is optional and the procedure can be performed with out strand separation or purification.
Dilutions of the DNA are then mixed with a buffer containing two primers (at approximately 1 Molar concentration) designed for strand displacement cascade amplification (that is, a secondary strand displacement primer and a rolling circle replication primer/tertiary strand displacement primer) and melted agarose at 60°C. The solution is overlayed on a petri dish to form a thin agarose layer ( 1.0% to 2% agarose), such that the concentration of circular DNA molecules is in the range of 500 to 5,000 per plate. Then enzymes and dNTPs required to initiate rolling circle replication are then added. The agarose film is incubated for 0.5 to 3 hours at 38°C (if the enzyme used is exo-Klenow) or at 60°C (if the enzyme used is exo (-) Bst or exo (-) Bca; Walker et al. (1992); Walker, IBC International 5 Conference, December, 1996). At appropriate dilutions, molecular "colonies" clonally derived from single circular vectors bearing DNA that originated in the PCR-amplified product, will form in the thin film of agarose. The initial density of seed DNA molecules should be such that the molecular colonies do not overlap.
In a preferred embodiment, one of the primers used in the amplification reaction is capable of forming a secondary structure. By lowering the temperature from 60°(: to 50°C, this special primer forms a hairpin structure that interferes with priming while the other primer continues to function normally. As the amplification reaction is continued 15 for another 45 minutes, a large proportion of the DNA product becomes single-stranded DNA generated by strand displacement driven by the single functional primer. In this way, during the latter phase of the amplification reaction a large proportion of the DNA contained in each colony becomes single-stranded.
20 After amplification, the agarose film is placed in a vacuum-blotting device with a membrane; that contacts the bottom of the agarose film. Pant of the DNA on the agarose film is vacuum-blotted onto the membrane. This generates a replica of the DNA colonies on the surface of the membrane.
Blotting is carried out for a brief period of time, so that approximately half of 25 the amplified DNA remains in the agarose.
In a preferred embodiment, the DNA from the colonies may be blotted to a CAM membrane, a special membrane that permits reversible binding of DNA. CAM is cellulose acetate membrane containing cystamine (2,2'-dithio-bis[cthylamine]). The membrane contains primary amino 30 groups, positively charged below pl-1 9.5, that can be easily removed under mild reductive conditions. (AM has been used to reversibly capture DNA
fragments separated by electrophoresis. C.'AM has been successfully used with DNA fragments ranging from 0.5 to 320 Kbp. C:AMs with different group densities can be synthesized (up to 1.65 umole/sq cm); CAM with 1 mole amino/sq cm has a binding capacity of at least 10 ~tg DNA/sq cm.
The standard elution conditions for DNA fragments up to 10 Kbp are: 2 hours at room temperature in 25 mM EDTA, 0.2 M NaCI, and 25 mM 2-5 mercaptoethanol. Larger fragments require higher concentrations of the reducing agent. The chemistry involved in the preparation of CAM is well established {see Sundberg and Porath, .I. Chromatog. 90:87-98 ( 1974); Uy and Wolf, (1977)).
CAM is prepared in two steps: { 1 ) Oxirane groups are introduced by reacting cellulose acetate membranes (0.45 pm) with variable concentrations (0 to 30 % v:v; depending upon the final group density required) of 1,4-butanediol diglycidyl ether in 0.1 M NaOI-1 containing 2 mg/ml sodium borohydride; the reaction is allowed to proceed for 16 hours at RT, with mild agitation. (2) Cystamine is then coupled to the oxirane-containing membrane 15 by reacting with 0.1 M cystamine in 0.1 M sodium tetraborate buffer, pH
9.5, for 16 hours at 37°C. Newly synthesized CAM is fully stable for at least 120 days at 4°C. The content of both oxirane and amino groups can be easily determined by standard reactions.
The thin agarosc gel containing amplified DNA molecular colonies may be stained with a sensitive dye such as SIBR-GREEN II (Molecular Probes) in order to localize the position of each colony. The coordinates of the colony position then serves to locate the position of the replicas on the membrane.
In order to recover DNA from a CAM membrane and obtain single-stranded DNA that can be sequenced by standard methods, the procedure is as follows: Molecular colonies are generated as described above to generate single-stranded DNA (embodiment using one specifically structured primer that is inactivated by lowering the temperature), blotted to a CAM
membrane, then a small droplet of DN,1 elution buffer (25 mM EDTA, 0.2 30 M NaCI, and 25 mM 2-mcrcaptoethanol) is placed on top of the desired colony, releasing in a few minutes a large proportion of the DNA of that colony replica. The small droplet is then recovered and mixed wish four 5 -s volumes of a buffer containing a sequencing primer and a suitable sequencing mixture for standard Sanger dideoxy sequencing.
In order to recover DNA from a membrane and regenerate replicatable DNA circles that can be amplified in solution by rolling circle 5 replication, or grown again as molecular colonies, the procedure is as follows: a small droplet of elution buffer (25 mM EDTA, 0.2 M NaCI, and 25 mM 2-mercaptoethanol) is placed on top of the desired colony, releasing in a few minutes a large proportion of the DNA of that colony. The small droplet is then recovered and mixed with four volumes of a buffer containing a restriction enzyme that will cleave the amplified DNA at the rare-cutter site that was designed into the spacer sequence of one of the original PCR
primers. After inactivating the restriction enzyme, the DNA is treated very briefly with highly diluted alkaline phosphatase, in order to cause partial dephosphorylation of the termini of the cleaved DNA. After phosphatase inactivation, the DNA is diluted and ligatcd in the presence of T4 DNA
ligase, thus regenerating closed circular molecules. A fraction of the re-circuiarized molecules will contain a single nick, resulting from dephosphorylated ligation junctions, and these molecules will be capable of initiating rolling circle replication and strand displacement cascade amplification reactions.
Illustration 2: Cloning using a Y-vector 1. Vector design: This linear vector has a 3' protruding T residue at each end, so as to permit ligation with PC_'R products that contain a 3'-terminal A (generated during PCR) at each end. The panhandle or tail of the Y is formed by two oligonucleotides that together constitute the second strand of the linear vector. The longer oligonucleotide contains an oligo-dA
sequence of 16 bases at the 3' terminus. The oligo-dA sequence serves as an affinity tag (where the aFfinity target will be oligo-dT). 'rhe shorter oligonucleotide contains a 5' phosphate. Sequences of an example of a 30 functional Y-vector are shown below.
K.~8 (SEQ !D N0:7) P-CATGAGGACTAGCAGATGGA I'GC(~GCCGCAGCTCG
T'(~ rAA-I~ACGACTC.ACTATAGGGT-3' A.60 (SEQ ID N0:8) P-CCCTATAGTGAGTCGTATTACACGAGCTGCTAGCAT

B.42 (SEQ ID N0:9) S P-GGCTAATGATGCTAGGCCGCATCCATCTGCTAGTCCTCATGT-3' 2. The Y-vector is assembled by incubation of oligonucleotide K.58, A.60 and B.42, for 5 minutes at 40°C, and then ligated with a mixture of PCR-amplicons using T'4 DNA ligase at 16°C for 16 hours, to generate 10 circular vectors with inserts. Oligonucleotide K.58 is in the first strand (that is, continuous strand) of the circular vector. Oligonucleotides A.60 and B.42 are in the second strand (that is, the discontinuous strand) of the circular vector. Insert sequences are in both strands of the circular vector.
3. After Iigation, the vectors with ligated inserts are incubated at room 15 temperature with oligo-d~I"-cellulose (Life Sciences, Inc.) in the presence of DNA ligase. The oligo-dT-cellulose is an affinity substrate where the cellulose is the solid-state substrate and the oligo-dT is the affinity target.
The Y-vector is ligated to the solid matrix via the panhandle sequence (the 5' end of B.42 is covalently bound to the 3' end of the oligo-dT on the oligo-dT-20 cellulose). 'hhe solid matrix is then washed with 20 mM Tris pl-I 8, 0.1 M
NaCI, to remove unligated vectors.
4. The matrix is washed with 0.5 ml of 50 mM NaOH, releasing singlc-stranded circular DNA (the first strand of the circular vector) from the cellulose matrix. 'rhe now immobilized second strand of the circular vector 25 remains attached to the cellulose matrix.
5. The circular vector is diluted serially to obtain concentrations in the range of several million circular vector molecules per milliliter.
Approximately 22 pl of this DNA solution (containing approximately 30,000 to 120,000 DNA molecules) is placed on a cover slip, and covered with a 30 polylysine-coated microscope slide. The DNA is allowed to bind to the polylysine-covered surface for 30 minutes at 37°C. The slide then dipped in 0.01% Tween-21>, and dried at room temperature.

6. Using an arraying instrument, an array is constructed consisting of 6000 individual micro-droplets of a solution containing two suitable primers designed for the constant sequence domains of the Y-vector, compatible buffer, and polymerase (Large fragment Bst, or exo-Vent DNA polymerase) capable of supporting rolling circle replication.
Primer 1 (23) (SEQ ID NO:10) GCATCCATCTGCTAGTCCTCATG
Primer 2 (22) (SEQ ID NO:11 ) CGCAGCTCGTGTAATACGAC'TC
Primer 1 serves as the rolling circle replication primer and a tertiary strand displacement primer. Primer 2 serves as a secondary strand displacement primer. The use of these primers will result in strand displacement cascade amplification. The diameter or the droplets should be approximately 0.150 to 0.200 millimeters. Droplets are dispensed on the I 5 surface of the slide in a controlled humidity atmosphere, in order to maintain a constant droplet volume for a period of 90 minutes. Alternatively, the diluted circular DNA molecules may be placed on the glass surface using the arrayer instrument to dispense small volumes of liquid that on the average contain a single ligatcd DNA molecule.

7. The array is incubated for 90 minutes at constant temperature (62°C) to amplify any DNA molecules in contact with (or within) the droplets.
When 33% of the droplets grow DNA colonies, and the rest do not grow anything, most colonies are likely to be of clonal origin. Adjusting the initial inoculum density, it should be possible to obtain up to 1500 clonal colonies per 6000-droplet array.
Optionally, the amplified nucleic acid can be replica plated in order to save a copy of the clones or to perform additional operations on the clones. In order to facilitate replica plating (via covalent binding to a glass surface), each of the two primers used for the strand displacement cascade 30 amplification of the circular vectors may be synthesized with a primary amino group at the 5' end. ~1t the end of the SDCA reaction, a glass slide is placed over the glass slide with the reaction droplets leaving an air gap of less than 1 mm (defined by the thickness of a plastic spacer) in such a manner that the glass on top will contact all of the liquid droplets without excessive compression. Prior to use, the lower face of this slide (the upper slide) is derivatized using the methods described by Guo et al., Nucleic Acids Research 22;5456-5465 ( 1994), and Guo et al., Nature Biotechnology 5 15:331-335 (1997)), to make the glass surface chemically reactive with amino groups. The two-slide sandwich is incubated for 1 to 2 hours at 37°C
in order to obtain covalent coupling of the amplified DNA contained in each droplet to the lower face of the upper slide.

8. DNA colonies may be identified by staining with the dye Sybr-10 Green-I (Molecular Probes, Inc.). Alternatively, replica slides may be made as indicated above, and used for any desired microarray hybridization experiment. 'fhe DNA in each colony may also be isolated and identified or analyzed by DNA sequencing.
s7

Claims

I claim:

1. A method of isolating and amplifying a nucleic acid molecule, the method comprising (a) ligating a nucleic acid molecule into a linear vector to form a circular vector comprising the vector and the nucleic acid molecule, wherein the linear vector is a double-stranded linear nucleic acid comprising two nucleic acid strands, wherein the second strand of the circular vector is discontinuous, and wherein the first strand in the circular vector is a closed circular strand, (b) amplifying the first strand by rolling circle replication to form tandem sequence DNA, wherein the amplification results in amplification of the nucleic acid molecule in the first strand.

2. The method of claim 1 wherein the second strand of the linear vector contains at least one nick, wherein the nick cannot be ligated.

3. The method of claim 1 wherein either the 5' or the 3' end of the second strand of the linear vector cannot be ligated.

4. The method of claim 1 wherein the second strand of the linear vector contains at least one gap or overlap.

5. The method of claim 1 wherein the method further comprises, following ligation and prior to amplification, separating the first strand from the second strand.

6. The method of claim 5 wherein the second strand includes an affinity tag.

7. The method of claim 6 wherein the first strand is separated from the second strand by binding the affinity tag to a substrate, denaturing the first and second strands prior to, simultaneous with, or following binding, and separating the first strand from the substrate.

8. The method of claim 5 wherein the second strand of the linear vector contains at least one overlap, part of the overlapping portions of the second strand are complementary, and the 3' end of the overlap extends beyond the part of the overlapping portions that are complementary, wherein the first strand is separated from the second strand by ligating one end of the second strand to a nucleic acid molecule coupled to a substrate, denaturing the first and second strands following ligation of the second strand, and separating the first strand from the substrate.

9. The method of claim 1 wherein step (a) comprises ligating a plurality of nucleic acid molecules into a plurality of linear vectors in a single reaction to form a plurality of circular vectors, each circular vector containing at least one nick, gap, or overlap in the second strand, wherein step (b) comprises amplifying the first strand of the plurality of circular vectors, and wherein the method further comprises, prior to amplification, dividing the ligation reaction to produce a plurality of separate amplification reactions.

10. The method of claim 9 further comprising making a replica of the amplification reactions.

11. The method of claim 10 wherein the replica of the amplification reactions is made by contacting the amplification reactions with a surface to which nucleic acids can bind.

12. The method of claim 10 wherein the replica of the amplification reactions is made by transferring part of each amplification reaction to form a replica amplification reaction.

13. The method of claim 9 wherein the ligation reaction is divided by spreading the ligation reaction onto a surface to form a spread, and wherein the separate amplification reactions are the locations of circular vectors on the surface after spreading.

14. The method of claim 13 further comprising making a replica of the amplification reactions.

15. The method of claim 14 wherein the replica is made by contacting the spread with a second surface to which nucleic acids can bind.

16. The method of claim 9 wherein any number or all of the amplification reactions are ordered as an array of reaction droplets or in an array of reaction vessels.

17. The method of claim 16 wherein, following amplification, all or part of the contents of any number or all the individual reaction droplets or reaction vessels are transferred by one to one mapping to a new set of reaction droplets or reaction vessels.

18. The method of claim 17 further comprising, following amplification, determining the presence of amplified nucleic acid in the amplification reactions, and transferring all or a part of the contents of the amplification reactions containing amplified nucleic acid reaction to a new set of reaction droplets or reaction vessels.

19. The method of claim 16 further comprising making a replica of the amplification reactions.

20. The method of claim 19 wherein the replica of the amplification reactions is made by contacting the amplification reactions with a surface to which nucleic acids can bind.

21. The method of claim 19 wherein the replica of the amplification reactions is made by contacting the amplification reactions with a surface treated with an affinity target capable of binding an affinity tag, wherein the amplified nucleic comprises affinity tags incorporated during amplification, wherein a portion of each amplification reaction is transferred to the surface.

22. The method of claim 21 wherein the affinity tag is biotin and the affinity target is streptavidin.

23. The method of claim 21 wherein the affinity tag is a reactive moiety and the affinity target is a corresponding reactive moiety, where a chemical reaction between the affinity tag and the affinity target results in the amplified nucleic acid being covalently coupled to the surface.

24. The method of claim 23 wherein the affinity target is phenylene diisothiocyanate, disuccinimidylcarbonate, disuccinimidyloxolate or dimethylsuberimidate and the affinity tag is a reactive amine.

25. The method of claim 19 wherein the replica of the amplification reactions is made by transferring part of each amplification reaction to form a replica amplification reaction.

26. The method of claim 16 wherein, following amplification, all or part of the contents of any number or all of the reaction droplets or reaction vessels are transferred and combined to create one or more sets of pooled reactions.

27. The method of claim 16 wherein the amplification reactions are arranged on the surface of a substrate.

28. The method of claim 27 wherein the substrate comprises acrylamide, cellulose, nitrocellulose, polystyrene, polyethylene vinyl acetate, polypropylene, polymethacrylate, polyethylene, polyethylene oxide, glass, polysilicates, polycarbonates, teflon, fluorocarbons, nylon, silicon rubber, polyanhydrides, polyglycolic acid, polylactic acid, polyorthoesters, polypropylfumerate, collagen, glycosaminoglycans, polyamino acids, chemical resistant metals, or corrosion resistant metals.

29. The method of claim 9 wherein the method further comprises, prior to dividing the ligation reaction, diluting the ligation reaction such that, on average, each amplification reaction contains a single circular vector.

30. The method of claim 9 wherein the method further comprises, following amplification, collecting a sample of each amplification reaction.

31. The method of claim 9 wherein the method further comprises detecting or sequencing the nucleic acid molecules in the amplification reactions or in the collected samples.

32. The method of claim 1 wherein rolling circle replication is primed by the second strand.

33. The method of claim 1 wherein rolling circle replication is primed by a rolling circle replication primer.

34. The method of claim 33 wherein the tandem sequence DNA is amplified by strand displacement replication to form secondary tandem sequence DNA.

35. The method of claim 34 wherein the secondary tandem sequence DNA is amplified by strand displacement replication to form tertiary tandem sequence DNA.

36. The method of claim 34 wherein strand displacement replication of the tandem sequence is primed by a strand displacement primer.

37. The method of claim 9 further comprising detecting one or more amplified nucleic acid molecules in one or more of the amplification reactions.

38. The method of claim 37 wherein the nucleic acid molecules are derived from cDNA generated by suppression subtractive hybridization.

39. The method of claim 37 wherein the plurality of nucleic acid molecules are all derived from the same source.

40. The method of claim 37 further comprising, following amplification, creating a replica of the amplification reactions.
contacting the amplification reactions with a first set of labeled nucleic acid probes and the replica amplification reactions with a second set of labeled nucleic acid probes, and comparing the pattern of hybridization of the first set of probes to the pattern of hybridization of the second set of probes, wherein differences in the patterns of hybridization indicate differences in the probe sets.

41. The method of claim 40 further comprising selecting for isolation or further analysis amplification reactions that hybridize to the first set of probes but not to the second set of probes, amplification reactions that hybridize to the second set of probes but not to the first set of probes, amplification reactions that hybridize to the both sets of probes, or amplification reactions that do not hybridize to either set of probes.

42. An in vitro method of cloning nucleic acid molecules, the method comprising (a) dividing a nucleic acid sample to produce a plurality of separate amplification reactions, (b) amplifying nucleic acid molecules in the amplification reactions, (c) making a replica of the amplification reactions, (d) testing nucleic acid molecules in either the amplification reactions or the replica amplification reactions to identify nucleic acid molecules of interest, and (e) retrieving the identified nucleic acid molecules of interest from the corresponding amplification reactions or replica amplification reactions that were not tested.

43. The method of claim 42 wherein the nucleic acid sample is divided by spreading the sample onto a surface to form a spread, and wherein the separate amplification reactions are the locations of circular vectors on the surface after spreading.

44. The method of claim 43 wherein the replica of the amplification reactions is made by contacting the spread with a second surface to which nucleic acids can bind.

45. The method of claim 42 wherein the method further comprises, prior to dividing the nucleic acid sample, diluting the nucleic acid sample such that, on average, each amplification reaction contains a single nucleic acid molecule.

46. A method of isolating and amplifying nucleic acid molecules, the method comprising (a) ligating a plurality of nucleic acid molecule into a plurality of linear vectors in a single reaction to form a plurality of circular vectors, each circular vector comprising a vector and a nucleic acid molecule, wherein the linear vectors are double-stranded linear nucleic acid comprising two nucleic acid strands, wherein the circular vectors each contain at least one nick, gap, or overlap in the second strand, and wherein the first strand in each circular vector is a closed circular strand, (b) separating the first strands from the second strands, (c) diluting and dividing the first strands to produce a plurality of separate amplification reactions that, on average, each contain a single circular vector, (d) amplifying the first strands of the plurality of circular vectors by rolling circle replication to form tandem sequence DNA, wherein the amplification results in amplification of the nucleic acid molecules in the first strands.

47. The method of claim 46 wherein the tandem sequence DNA is amplified by strand displacement replication to form secondary tandem sequence DNA.

48. The method of claim 47 wherein the secondary tandem sequence DNA is amplified by strand displacement replication to form tertiary tandem sequence DNA.

49. A method of isolating and amplifying a nucleic acid molecule, the method comprising (a) ligating a nucleic acid molecule into a linear vector to form a circular vector comprising the vector and the nucleic acid molecule, wherein the linear vector is a double-stranded linear nucleic acid comprising two nucleic acid strands, wherein the circular vector contains at least one nick in the second strand and wherein the first strand in the circular vector is a closed circular strand, (b) amplifying the first strand, wherein the amplification results in amplification of the nucleic acid molecule in the first strand.

50. A kit for isolating and amplifying nucleic acid molecules, the kit comprising (a) a linear vector wherein the linear vector is a double-stranded linear nucleic acid comprising two nucleic acid strands, and wherein (1) the linear vector contains at least one nick, wherein the nick cannot be ligated, (2) either the 5' or the 3' end of the second strand of the linear vector cannot be ligated, (3) the second strand of the linear vector contains at least one gap, (4) the second strand of the linear vector contains at least one overlap, or (5) any combination of (1), (2), (3) or (4);
(b) a rolling circle replication primer, wherein the rolling circle replication primer is complementary to a portion of the first strand of the linear vector; and (c) a strand displacement primer, wherein the strand displacement primer matches a portion of the first strand of the linear vector.

51. A linear vector wherein the linear vector is a double-stranded linear nucleic acid comprising two nucleic acid strands, wherein the second strand of the linear vector contains an affinity tag, and wherein (1) the linear vector contains at least one nick, wherein the nick cannot be ligated, (2) either the 5' or the 3' end of the second strand of the linear vector cannot be ligated, (3) the second strand of the linear vector contains at least one gap, (4) the second strand of the linear vector contains at least one overlap, or (5) any combination of (1), (2), (3) or (4).

52. A linear vector wherein the linear vector is a double-stranded linear nucleic acid comprising two nucleic acid strands, and wherein the second strand of the linear vector contains at least one overlap, part of the overlapping portions of the second strand are complementary, and the 3' end of the overlap extends beyond the part of the overlapping portions that are complementary.