OF NUCLEIC ACIDS AND OTHER LINEAR MACROMOLECULES

IN GEL MEDIA WITH RESTRICTIVE PORE DIAMETERS

TECHNICAL FIELD

This invention relates to methodology and techniques for the high resolution electrophoretic separation of nucleic acid fragments (DNA, RNA, etc.) in gel media with restrictive pore diameters.

BACKGROUND

Electrophoretic analysis is a powerful and widely used technique in the fields of biochemistry and molecular biology. Nucleic acid analyses are carried out on sample mixtures which range in size from oligonucleotides several nucleotides in length to very large DNA fragments millions of base pairs in length. Agarose and polyacrylamide are the two main types of supporting mediums which are use to produce porous gels for the electrophoretic separation of nucleic acids.

The short to intermediate sized nucleic acid fragments [10 to 1000 base pairs (bp) ] are separated in homogeneous polyacrylamide gels (20% to 2.5% in polyacrylamide concentration) arranged in vertical slab formats. [The widely accepted nomenclature is to express polyacrylamide concentration as %T/%C, where %T is the total monomer and crosslin er (grams per 100 ml) , %C is the percentage of T which is due to crosslinker] . The higher concentrations produces gels with the smallest pores sizes and the lower concentration produces the larger pores. The intermediate to high molecular weight nucleic acid fragments (1000 to 10 ⁶ bp) are separated in homogeneous agarose gels (3% to 0.25%) arranged in a

horizontal slab formats. Agarose gels have pore diameters much larger than the polyacrylamide gels.

In general, both the agarose and polyacrylamide slab gels used for nucleic acid separations range in size from so-called minigels which are approximately 5 cm x 5 cm to the more standard sized gels which are 20 cm x 20 cm formats. These formats provide running distances of 5 to 20 centimeters in which the separation can occur. In the case of DNA sequencing, where resolution of DNA fragments varying by one nucleotide iε required, homogeneous polyacrylamide gels from 20 cm to 40 cm are commonly used, and in some cases gels as long as 100 cm have been used. In addition to length, the higher resolution needed for DNA sequencing is achieved by using lower percentage (%T/%C) and thinner polyacrylamide gels.

There are three general processes for separation of nucleic acids and proteins in porous gel media. The gel sieving process separates nucleic acids in homogenous agarose gels and separates nucleic acids and proteins in homogenous polyacrylamide gels. The pulse field process separates very large DNA fragments in homogeneous agarose gels. The pore limiting process separates nucleic acids and proteins in gradient polyacrylamide gels. A clear understanding of these three physical processes is important and relevant to this invention.

The physical mechanisms describing the mobility of DNA in homogeneous polyacrylamide and agarose gels is usually described (mathematically) by the formalisms and relationships developed by Ogston et al, Trans. Faraday Soc.. 54:1754-1757 (1958), and Ferguson, Metabolism, 13:985-1002 (1964). In general, it is taught that separation in homogeneous gels is achieved by the "sieving process or mechanism" where

the size or effective radius (R _π ,) of the DNA molecules being separated are of the order and/or smaller than the average pore size (a) of the gel. Rodbard et al, Proc. Natl. Acad. Sci. USA. 65:970-977 (1970). Thus, for the optimal electrophoretic separation of a given range of DNA molecules, the concentration of polyacrylamide (%T/%C) or agarose is chosen to provide a pore size so that R,- < a. The limit of resolution for very large DNA molecules (>50,000 bp) is reached when the size or radius of gyration (R _g ) of the DNA molecules exceeds the maximum pore size obtainable with an agarose gel. For purposes of this invention the effective radius R„, and R _g can be considered equivalent. At that point, the DNA molecules are no longer resolved by the sieving process, and they begin to migrate end-on through the gel at close to the same rate. This phenomena is called "reptation", and a number of theoretical descriptions have been published. See for example Lerman et al, Biopolymers. 21:995-997 (1982) ; Lumpkin et al, Biopolymers,

21:2315-2316 (1982); Stellwagen et al, Biochemistry. 22:6180-6185 (1983); Slater et al, Phys. Rev. Lett.. 55:1579-1582 (1985); Slater et al, Biopolymers. 24:2181-2184 (1985); Slater et al, Biopolymers, 27:509-524 (1988); Slater et al, Biopolymers 28:1781- 1791 (1989) .

It should be kept in mind that although the length of the these DNA molecules is greater than the pore size, the pore diameter of a low concentration agarose gel (0.25 %) is still very much larger ( ^" 200 nm) than the diameter of double-stranded DNA ( ^" 2 nm) .

A special process or technique called "pulsed field" electrophoresis iε required to overcome the loss of resolution for large DNA molecules in agarose gels. Schwartz et al, Cell, 37:67-75 (1984). In this

method, a pulsed alternating orthogonal electric field is applied to the gel, which re-orients the DNA molecules, and resolution of very large DNA fragments is then achieved. Loss of resolution does not only occur for very large DNA molecules on agarose gels, but is also seen when short to intermediate size DNA fragments begin to exceed the pore size of a given homogeneous polyacrylamide gel. Significantly reduced mobilities and concurrent loss of resolution is seen for those DNA fragments where R _g > a (exceeding the pore size) relative to those fragments where R _g < a. For example, mobilities of DNA fragments in the 300-bp to 2000-bp range become significantly reduced as polyacrylamide concentrations increase from 3.5%T to 10.5%T. Holmes et al, Electrophoresis, 12:253-263 (1991) .

Nucleic acids and proteins have also been separated in non-homogeneous or gradient gels by the "pore-limit process". Gradient gels usually have a polyacrylamide concentration that increaseε from 2%T- 3%T at the top of the gel to 30%T-40%T at the bottom of the gel. As moleculeε sieve through the gel the pores become smaller, until finally they reach a pore size or pore limit, which slows the velocity of the molecule to almost zero.

Early work by Jeppesen Anal. Biochem.. 58:195-207 (9174) , involving the εeparation of DNA fragmentε (114-bp to 21,266-bp) on gradient gels (2.5%T to 7.5%T) demonstrates the effect very clearly. Under normal electrophoresis conditions, the mobility of the DNA fragments approach zero velocity as they reach the so-called "pore limit" of the gel. The position in the gel where a given size DNA fragment reaches its pore limit, correlates with the gel pore diameter, that is where R _g = a. The term "pore limit" is not

co pletely accurate, in that increased running time shows some very small continued mobility of the bands. However, in these normal gel formats there is no detectable change in the relative position of the bands, that iε, there iε no further resolution of the fragments.

For homogeneous gels, the problem of decreasing mobility of DNA fragments iε readily overcome by simply going to a lower polyacrylamide gel (%T/%C) or agarose gel concentration. Exceptions are the very large DNA fragments which greatly exceed the maximum pore size of agarose gel, and require the pulsed field technique. Because of the marked reduction in mobility and apparent loss of resolution, the separation of DNA fragments in more highly concentrated homogeneous polyacrylamide gels, where R _g > a, has not been described. This is particularly true for those higher concentrations of polyacrylamide gels (>20%T) where the pore sizes begin to approach the diameter of double-stranded DNA. This invention is primarily concerned with novel methodology for achieving high resolution separations of nucleic acids fragments which greatly exceed the gel pore size (R _g » a) . A few early attempts were made to investigate the potential for microelectrophoresis. Edstrom, Biochem. Biophyε. Acta, 22:378 (1956), firεt deεcribed a microtechnique for the electrophoretic εeparation of purine and pyrimidine baεeε along a εilk thread. Matioli et al, Science, 150:1824 (1965), have εeparated hemoglobin variantε on polyacrylamide fiberε. In thiε work, 20% acrylamide gelε were used to separate hemoglobin variantε from εingle cellε. Hemoglobin moleculeε (MW ^" 64,000) with molecular radii of 2.66 n are not pore εize limited in 20%

polyacrylamide gelε, thus the separation occurs by the "normal gel sieving process". See Andrews, in "Electrophoresis: Theory, Techniques, and Biochemical and Clinical Applications", Oxford University Press, New York, chapter 2, pp. 5-74 (1986) . As will be shown, the present invention is concerned with the microelectrophoretic separation of molecules whose effective radius, or radius of gyration, are significantly greater than gel pore size. Other separation syεtemε have been deεcribed that were referred to aε mini-gel electrophoreεis, that emphaεize separation of proteins. Grossbach, in "Electrophoreεiε and Iεoelectrofocuεing in Polyacrylamide Gelε", edε., Allen et al, Walter de Gruyter, New York, p.207 (1974), deεcribeε separation of proteins in 50 urn to 100 urn diameter capillaries tubes which were 2.5 centimeters long, using standard gels and bufferε. Neuhoff et al, Biochem J. , 117:623 (1970) and Biεpink et al, in "Electrofocusing and Isotachophoreεiε", eds., Radola et al, Walter de

Gruyter, New York, pl25 (1977) , deεcribe analytical PAGE on proteins carried out using 5 centimeter capillary tubes. "Micro versionε" of polyacrylamide slab gels have been prepared on 75 mm x 25 mm microscope slides [Maurer et al, Anal. Biochem.. 46: 19 (1972) ] and on 82 mm x 102 mm glass microscope slides [Matsudaira et al, Anal. Biochem. , 87:386 (1978) ].

In all of the above work, gel lengths or running distances are at least five to forty times longer than described by the present invention; the above described syεtems are more accurately described as mini-gels or scaled down version of large gels. In the above work, εtandard gel for ulationε were used. That iε, gel formulationε in the above εyεtemε do not

significantly deviate from what iε used for the corresponding large scale separation. More importantly, in the above εyεtemε the molecules (proteinε) are being εeparated by the "normal sieving procesε" which occurε for molecules which have a molecular radius that iε ε aller than the gel pore size of the εeparation medium.

Gradient gel electrophoreεiε is a technique in which a gel matrix having an increasing concentration of polyacrylamide (3% to 40%) along the separation axis is used to separate macromolecules in a wide range of sizes. In gradient gel electrophoresiε the rate of migration of the componentε through a gel gradient varieε inverεely with time. After a sufficient electrophoresis time a stable pattern develops in which the different components continue to move slowly but their relative positions remain constant. That is, as the components reach the gel pore size that is close to their own size (molecular radius) , their terminal velocity approaches zero. See, for example, Andrews, in "Electrophoresiε: Theory, Techniqueε, and Biochemical and Clinical Applicationε", Oxford Univerεity Preεε, New York, chapter 4, pp. 93-116 (1986). A great deal of theoretical work and application of techniqueε for determining molecular weightε and molecular radii of proteins has been described by Rodbard et al, Anal. Biochem. , 40:95-134 (1971); Manwell, Biochem. i, 165:487-495 (1977); and Campbell et al, Anal. Biochem. , 129:31-36 (1983).

DNA separation using gradient gel electrophoresis haε been deεcribed by Jeppesen, Anal. Biochem. , 58:195-207 (1974). DNA fragments of molecular weightε from about 7 x 10 ⁴ daltonε (114 bp) to about 14 x 10° daltonε (21,226 bp) were εeparated on linear gradient

polyacrylamide gels having concentrations from 3.5% to 7.5% or from 2.5% to 7.5% with a crosslinker (C) concentration ranging from 2.5% to 5%. The gels described by Jeppesen were 14 cm long x 14 cm wide x 0.3 cm in thickneεs. Electrophoresis was carried out at 10 volts/cm from 16 to 20 hours, until the DNA fragments reached terminal velocities that approached zero. According to Jeppesen, maximum separation is achieved as the fragments approach zero velocity (gel pore limit) , and further increase in running time resultε in little change in band poεition. Table 1 below shows the approximate acrylamide concentration (%T) or gel pore limit at which point the separated DNA fragments reached terminal velocities:

TABLE 1

Fragment %Gel(%T) Terminal Size (bp) ¹ Velocity Reached

1,000 5.2 ^" 646 5.9 ^" 492 6.5 ^" 354 6.7 ^" 323 7.0 ^" 185 7.2

¹ Lambda DNA/R1 fragments are the double-stranded (ds) DNA fragments produced by an Eco RI restriction endonuclease digeεtion of lambda DNA having εizeε described by Phillipsin et al, J. Mol. Biol. , 123:371 (1978) . SV40 DNA/R frag entε are the dε DNA fragmentε produced by an endonucleaεe R (Hind III) digeεt of SV40 DNA and having εizes estimated from the given molecular weights. Fragments H, I, J, and K continue to migrate very slowly during continued electrophoresiε.

The Jeppeεen work demonstrates two important points: (1) that in large gel formats the further εeparation of DNA fragmentε larger than their pore limit εize iε not obεerved; and (2) that acrylamide gel concentrations (%T/%C) only up to certain identified levels are useful to separate a given range of DNA fragments by the "normal gel sieving procesε".

Another study involving separation of single- stranded RNA's in miniature gradient gels was described by Neuhoff, in "Micromethods in Molecular Biology", Springer-Verlag, New York, pp. 56-72, 1973. Gradient gels of from 8 mm to 20 mm in length were used to separate ribosomal RNA fragments from 28S ( ^" 5000 nucleotideε or nt) to 4S (85 nt) in size. Separations were carried out for 30 to 60 minutes at 40 to 60 volts/cm. Reεultε clearly confirm the Jeppeεen εtudy, with the εingle-εtranded RNA fragments migrating to gel pore limits that would be expected for their εize. Moεt importantly, the reεults on these εmaller gradient gels do not suggeεt that

separations of large fragmentε can be carried out in the more concentrated partε of the gel.

Nucleic acid fragment analyεis using denaturing po _l yacrylamide gels was described at least by Maniatis et al, Biochemistry, 14:387 (1975). Maniatis (1975) described the relative electrophoretic mobilities of RNA and DNA molecules having chain lengths of 10-150 nucleotides (nt) when εeparated in 12% polyacrylamide/3.3% croεεlinker (N,N'-methylene biεacrylamide) (12%T/3.3%C) gelε containing 7 M urea (denaturing gel) , uεing gel dimenεionε of 20 cm x 20 cm x 0.15 cm and running conditionε of IX TBE buffer for εeveral hours at a constant voltage of 10 volts per centimeter (10 volts/cm) in the vertical direction. Sanger et al, Nature, 265:687 (1977), showed the relative mobility of bacteriophage φX174 DNA/Hind II fragments with chain lengths of 1,049, 770, 609, 495, 393, 335/340/345 (unresolved triplet) , 297/291 (unresolved doublet), 163, and 79 nt in 5% polyacrylamide gel containing 98% formamide. Sanger's gel dimenεions were 20 cm x 20 cm and the gels were run in 0.02 M phosphate buffer. Maniatis et al, in "Methods in Enzymology", vol. 65, part 1, eds. Groεsman et al, Academic Presε, New York, p 299 (1980) , recommendε the above general conditions for separation of short to intermediate size DNA and RNA fragments under denaturing conditionε.

Sambrook et al, in "Molecular Cloning: A Laboratory Manual," 2nd edition, Cold Spring Harbor, New York, pp. 6.2 to 6.63 (1989), and the references contained within, recommend the following gel concentrations and conditions for separating nucleic acid fragments on agarose gels having fragment sizeε in the range of 100 to 60,000 nt at the following percentageε:

TABLE 2

AGAROSE GEL ELECTROPHORESIS

Range of Sizes for

Following the above recommendations, agarose gel electrophoreεis of DNA was generally carried out in the horizontal direction using slab gels ranging from 14 to 20 centimeters in length. It was generally recommended that the gels be run at no more than 5 volts/cm. Depending upon the degree of resolution required and the voltage utilized, running times for agarose gels varied from one to εixteen hourε. However, agaroεe gelε have very poor resolution below 100 nt, and only intermediate resolution at higher chain lengths (>100 nt) . Therefore, agarose gels are more frequently uεed for Southern analyεiε or reεtriction fragment analyεiε, rather than for DNA εequencing applicationε which requireε high resolution of shorter chain lengths.

Very large linear dε-DNA moleculeε were found to migrate through agaroεe gelε at the same rate. The limit of resolution was reached when the radius of gyration of the linear DNA duplex exceedε the pore εize of the gel. At that point the DNA can no longer

be sieved by the gel according to size but must now migrate end-on through the narrow pores. This process of end-on migration is known as "reptation". Sambrook et al, in "Molecular Cloning: A Laboratory Manual" 2nd edition, Cold Spring Harbor, New York, pp. 6.2 to 6.63 (1989) . One εolution to the problem of εeparating large DNA molecules is a technique called pulsed field electrophoresiε developed at least by Schwartz et al., Cell, 37:67 (1984). In this method, pulsed, alternating, orthogonal electric fields are applied to agarose gels. The large DNA molecules become trapped in their "reptation tubes" and can make no further progresε through the gel until they have reoriented along the new axis of the electric field. The larger DNA molecules require a longer reorientation time; the smaller moleculeε with reorientation timeε less than the pulse begin to separate according to size. Pulse field electrophoresiε involves large agarose gel formats, complex electrode arrangements, and very long running times to achieve separations. Pulse field electrophoresiε provideε important background information for the present invention because it shows; that very large DNA molecules which are larger than gel pore size are not separated effectively using regular large scale gel formats and procedures, and that one solution to the problem is the complicated and extremely long process described above.

Sambrook et al, in "Molecular Cloning: A Laboratory Manual", 2nd edition, Cold Spring Harbor, New York, pp. 6.2 to 6.63 (1989), and the references contained within, recommend the following gel concentrations and conditions for separating nucleic acid fragments on non-denaturing polyacrylamide gelε in the following concentration ranges where the DNA fragments have sizes in the range of 6 to 2000 nt in

length:

TABLE 3

NON-DENATURING ACRYLAMIDE GEL ELECTROPHORESIS

Range of Sizes

Acrylamide (%T/%C) for DNA Moleculeε (nt)

3.5%T/3.3%C 100-2000

5.0%T/3.3%C 80-500

8.0%T/3.3%C 60-400

12.0%T/3.3%C 40-200

15.0%T/3.3%C 25-150

20.0%T/3.3%C 6-100

Following the above recommendationε, gels were typically run in vertical format and over lengths from 10 to 100 centimeters depending on the resolution required. Gels were run in IX TBE buffer at voltage gradients between 1 volts/cm to 8 volts/cm. Higher voltages were not recommended due to the problems associated with overheating. Runs generally took from one to twelve hours depending upon resolution required. Strand-separating polyacrylamide gels were recommended for nucleic acid fragmentε below 1000 nt in length, in particular for εequencing by the Maxam- Gilbert procedure [Maxam et al, Proc. Natl. Acad. Sci. U. S. A.. 74:560 (1977)] and for hybridization to low abundance RNA'ε (Maniatiε et al, in "Molecular

Cloning: A Laboratory Manual", Cold Spring Harbor, New York, pp. 179-185, 1982) . For DNA fragmentε greater than 200 nt in length, 5%T/2%C gelε were recommended; while 8%T/3%C gelε were recommended for fragmentε less than 200 nt in length. These gels were typically run

in IX TBE and at 8 V/cm.

Sambrook et al, in "Molecular Cloning: A Laboratory Manual", 2nd edition, Cold Spring Harbor, New York, pp. 6.2 to 6.63 (1989), and the referenceε contained within, recommend the following gel concentrationε and conditionε for separating nucleic acid fragments on denaturing polyacrylamide gels containing 7 M urea, where the duplex DNA fragments are in the 10 to > 200 nt range:

TABLE 4

DENATURING ACRYLAMIDE GEL ELECTROPHORESIS

Range of Sizes for

Acrylamide (%T/%C) DNA Molecules (nt) 4%T/5%C >200

5%T/5%C 80-200

8%T/5%C 40-100 12%T/5%C 10-50

Following the above recommendations, denaturing gels were typically run at 20 volts/cm in IX TBE, and it was typically recommended that the DNA fragments be allowed to migrate the full length of the gel in order to obtain maximum separation. It should be pointed out that the agarose and polyacrylamide gel formulations discussed above have been frequently used in so-called mini-gel formats, which are generally about five to ten centimeterε in length. Mini-gelε repreεent a two to four fold size reduction in the linear dimension, and are basically only scaled down versions of the systems described above. As will be discuεεed below, the high reεolution necessary for DNA

εequencing iε not achieved when accepted large-εcale procedures are scaled down to mini-gel format. More importantly, as will be demonstrated by the present discloεure, microelectrophoresis producing the reεolution described herein can not be obtained by simply scaling down large scale procedures.

In the case of DNA sequencing, where separation of DNA or RNA fragments differing by a εingle nucleotide iε required, electrophoreεiε techniques with speed and high resolution are required. Maxam et al developed the chemical cleavage sequencing method .Proc. Natl. Acad. Sci. U. S. A.. 74:560 (1977)] and used the following polyacrylamide compoεitionε and procedures for sequencing gels: for DNA εequenceε from 1 to 30 nt a 20%T/5%C (8.3 M Urea) gel waε used; for sequenceε from 25 to 250 nt a 8%T/5%C (8.3 M Urea) gel waε uεed; and for εequenceε greater than 250 nt a 6%T/5%C gel waε uεed. The gelε were 20 cm x 40 cm in width and length, reεpectively, and gel thickneεεeε of 1.5 mm, 0.5 mm, and 0.3 mm were used. The thinner gels of 0.3 mm were typically preferred due to their capacity to withεtand higher voltageε and faεter run timeε. Sequencing gelε were run in IX TBE buffer at voltageε of 25 volts/cm for 1.5 mm gelε, and 50 volts/cm for the 0.3 mm gels. See, for example, Maxam et al, in "Methods in Enzymology", vol. 65, part 1, eds., Academic Press, New York, p. 499 (1980). Depending on the separation required, resolution of the DNA fragments in the sequencing reactions takes many hours to complete.

More recently Barker described in "Nucleic Acid Sequencing: A Practical Approach", edε. et al, IRL Preεε, New York, Chapter 5, 117 (1989) , the use of one meter long gels for DNA εequencing by the Maxam and Gilbert method. Early workerε carrying out the primed

εynthesis DNA sequencing method [Sanger et al, Proc. Natl. Acad. Sci. USA.. 74:5463 (1977)] used the following polyacrylamide compositionε and procedureε for sequencing gels: for DNA fragments from 30 to 250 nt an 8%T/5%C (7 M urea) gel was used, and for fragments >250 nt a 6%T/5%C (7 M urea) gel was used. See, for example, Smith in "Methods in Enzymology", vol. 65, part 1, eds, Groεsman et al, Academic Press, New York, p. 560 (1980) ; and Sanger et al, FEBS Lett. , 87:107 (1978). The gels were 20 cm x 40 cm, and 0.35 mm in thickness. Gels were run at 30 to 40 volts/cm and take many hours to complete. In general, as DNA sequencing developed, the trend in gel formulationε and size of polyacrylamide slab gels progresseε toward thinner gelε and εlightly leεs concentrated gels, but there were no significant reductions in the length of these sequencing gels. See, for example, Andrews in "Electrophoresis: Theory, Techniques, and Biochemical and Clinical Applicationε", Oxford University Press, New York, chapter 6, pp. 148-177 (1986) ; and "Nucleic Acid Sequencing: A Practical Approach", eds. Howe et al, IRL Press, New York, (1989) , and references contained within.

Automated DNA sequence analyzerε baεed on fluorescent detection have become available in the past few years. Workerε at Applied Bioεystems, Inc. (ABI) and the California Institute of Technology have developed an automated DNA sequencer which uses four fluorescent dye-labelled primers. Connel et al. BioTechniσueε, 5:342 (1987). Each of the dye-labelled primerε iε paired with one of the four dideoxynucleoεide triphoεphate chain terminatorε, and used in the Sanger sequencing method to introduce fluorescent labels into DNA fragments produced by primer extension. More recently, fluorescent

dideoxynucleotide nucleotide derivativeε have become available for uεe on the ABI εequencing machines, thus eliminating the need for fluorescent primers. The fluorescent fragments produced in the εeparate A, C, G, and T reactionε are combined and can be co- electrophoresed in the same lane and distinguished during electrophoresis by the color of their fluorescence. The system has an argon-ion laser which excites each of the fluorescent fragmentε aε they paεs through a small area near the bottom of the separation gel. The fluorescent signal from the fragments is focused by a collection lens through a four wavelength selectable filter onto a photomultiplier tube (PMT) . The digitalized εignal from the PMT iε tranεferred directly into a computer for εubεequent processing and display.

After computer proceεεing, the εequence information iε preεented graphically as a linear array of colored peakε with the actual baεe εequence (A, T, G, & C) given above each peak. The system useε

6%T/5%C (7 M Urea) polyacrylamide gelε, 0.4 mm or less in thicknesε. An 8% gel iε recommended for better reεolution cloεer to the primer; while a 4% gel can be used for better downstream resolution (>400 bp) . A running distance of 25 cm is required to achieve the appropriate resolution of the fluorescent DNA fragments. Generally, electrophoresiε iε carried out at 30 to 40 voltε/c , with sequences up to 500 bases being determined in a 8 to 12 hour run. Heiner et al, in "Nucleic Acid Sequencing: A Practical Approach" eds. Howe et al, IRL Presε, New York, Chapter 8, pp.221-235 (1989).

Another automated DNA εequencer using fluorescent detection haε been developed at the E. I. du Pont Company as described by Prober et al, Science. 238:336

(1987) . This system useε an 8%T/5%C (7M Urea) gel in a 20 cm wide by 40 cm long by 0.3 mm thick format. Thuε, the εtate of the art for the high reεolution separation of DNA fragments on polyacrylamide slab gels for sequencing purpoεeε: (1) uses gel concentrations of about 6%T/5%C, (2) requires gel lengths of at least 25 to 40 cm, (3) is limited to voltages of 40 volts/cm, and (4) requireε many hourε of running time. A newer technique, which may be considered second generation DNA sequencing technology, is capillary gel electrophoresiε. Morriε et al, U.S. Patent No. 4,909,919; Luckey et al, Nucl. Acids Res. , 18:4417 (1990); Drosεman, Anal. Che . , 62: 900 (1990); and Guttman et al, Anal. Chem. , 62:137 (1990). Capillary electrophoreεiε involves the use of very fine glass capillary tubes 50 to 100 urn in diameter and 40 cm to 100 cm in length. Capillary gel electrophoresiε haε an advantage in that much greater electric fields may be applied, because of the reduced Joule heating in the small diameter capillary. This resultε in as much as a 14 fold faster separation speed over conventional slab gel methodologies. Thus, whereas a DNA sequencing separation of 300-400 bases run at 30 to 40 volts/cm on a 40 cm slab gel takes 7-8 hours to complete, the same separation on a 40 cm capillary gel run at over 400 volts/cm takes only 30-40 minutes to complete. The lower concentration gel [3.2%T/2.7%C (7M Urea) ] used in the capillary columns is one factor in the improved separation speed.

Thus while capillary gel electrophoresiε represents a significant improvement in separation speed, the gel lengths necesεary for achieving the εeparation are 40 to 70 cm long utilizing very low polyacrylamide concentrationε and εeparation timeε of

30 minutes or more.

BRIEF SUMMARY OF THE INVENTION

It has now been discovered that electrophoretic separation of macromolecules can be accomplished in relatively εhort gel εupport (matrix) diεtanceε, referred to herein aε εeparation patterns, by using matrices having εubstantially higher gel concentrationε than uεed previously before in a format that allows for the application of subεtantially higher electric fieldε in the range of 5 to 500 voltε per millimeter (mm) of gel εupport.

Thuε in one embodiment the invention contemplates an electrophoretic method to effect differential net migration based on molecular size of electrically charged linear macromolecules through a gel support in a single dimension. The method comprises subjecting electrically charged macromolecules applied to a gel support to an electric field oriented along a.single axis within the gel for a time period sufficient to effect migration in the direction of the oriented field and form a separation pattern in order of the respective molecular weights of the linear macromolecules in a distance of about 0.1 to about 20 mm. The gel support has a restrictive pore diameter relative to the effective radius of the macromolecule. The electric field is applied in an amount of about 5 to about 500 volts per mm of gel support along the axis length, and said gel support having a width perpendicular to said axis of about 0.05 to 2.0 mm. In one embodiment, the gel support is in a capillary tube format having an inner diameter of about 0.1 to 1.0 millimeters.

In another embodiment, the present method describeε electrophoretic εeparationε where the

applied electric field is about 15 to 50 volts per mm of gel support length.

Depending on the size of the linear macromolecules to be separated, the invention describes preferred gel supports of different ranges of acrylamide and croεεlinking biε-acrylamide.

Preferred gel εupport lengthε are alεo described as being lesε than 25 millimeterε (mm) , preferably between 0.1 to 20 mm and more preferably about 2 to 15 mm. Within these gel support lengths, and by virtue of the separation pattern resolution and speed of resolution afforded by the present invention, preferred separation patterns are deεcribed having lengthε of 0.5 mm to 20.0 mm when practicing the methodε of the preεent invention.

The preεent invention provides for the separation of large size ranges of single-stranded and double- stranded nucleic acid fragments in pore εtructureε (2 to 8 nanometerε) that are closer to the diameter of these molecules (2 nanometers) rather than their respective effective radii, or radii of gyration. This methodology provides three important and useful advantages over other electrophoretic separation processes: (1) A significant reduction in the linear dimension necessary to carry out electrophoretic separations of nucleic acid fragments in both homogeneous and non-homogeneous gel media. Separations normally requiring many centimeters of running distance, are resolved in millimeters or less. (2) High resolution of DNA fragments, with extremely narrow banding patterns. Separations of DNA fragments with band widths of 10 micron (or lesε) can be achieved. (3) A εignificant reduction in εeparation time. Separationε can be achieved in a matter of minuteε. This methodology is presently the only

electrophoretic technique which can achieve all three of theεe important advantages simultaneouεly.

BRIEF DESCRIPTION OF THE DRAWINGS FIGURE 1 in panel A illustrates a capillary tube microgel of this invention. Panel B showε a microelectrophoresis chamber as described herein.

FIGURE 2 illustrates the complete separation of the 0X174 DNA/Haelll restriction fragment ladder on a 6 mm, 12%T/6%C microgel (40X Image) aε deεcribed in

Example 1. The electrophoresis was carried out at 33 volts/mm for 2 minutes. The fragments are separated in a distance of 3.3 mm measured from the top of gel to lower the band. FIGURE 3 illustrates the upper band separation of the φX174 DNA/Haelll restriction fragment ladder on a 6 mm, 12%T/6%C microgel (100X Image) as described in Example 1. The electrophoresiε waε carried out at 33 voltε/mm for 2 minutes. The scale bar = 120 urn. FIGURE 4 illustrateε the intermediate band εeparation of the 0X174 DNA/Haelll reεtriction fragment ladder on a 5 mm, 12%T/6%C microgel (100X Image) as described in Example 1. The electrophoresis waε carried out at 40 volts/mm for 1 minute and 15 secondε. The fragmentε are separated in a distance of 2.03 mm measured from the top of gel to the 194 bp band. Separation of the 281 and 271 bp fragments is visible under certain detection conditions. The scale bar = 240 u . FIGURE 5 illustrateε the upper band εeparation of the 0X174 DNA/Haelll reεtriction fragment ladder on a 5 mm, 18%T/6%C microgel (100X Image) as deεcribed in Example 2. The electrophoreεiε waε carried out at 40 voltε/mm for 1 minute, 50 εecondε. The fragmentε are εeparated in a diεtance of 0.94 mm meaεured from the

top of gel to the 310 bp band. The scale bar = 240 urn.

FIGURE 6 illustrateε the εeparation of the 0X174 DNA/Haelll restriction fragment ladder on a 11 mm, 22%T/6%C microgel (4OX Image) aε described in Example 5. The electrophoresis was carried out at 30 volts/mm for 5 minutes. The fragments are separated in less than 1.5 mm. The white scale bar = 550 urn. The large vertical arrow to the right is 2.0 mm and extends from the top to the bottom of the gel. The eleven dashes at the left indicate the position of the following nucleic acid fragmentε, from top to bottom: 1353 bp, 1078 bp, 872 bp, 603 bp, 310 bp, 281 bp, 271 bp, 234 bp, 194 bp, 118 bp and 72 bp. FIGURE 7 illuεtrateε the εeparation of fluorescent 18-mer, 21-mer, 23-mer, and 32-mer oligonucleotides on an 8 mm, 35%T/12%C microgel (100X Image) as described in Example 8. Electrophoresiε waε carried out at 24 voltε/mm for 6 minuteε. Separation occurs in less than 2 mm. The scale bar = 240 urn. FIGURE 8 illuεtrateε the εeparation of fluoreεcent 18-mer, 21-mer, and 22-mer, oligonucleotides separated on a 10 mm 35%T/12%C microgel (100X Image) as described in Example 9. Electrophoresis was carried out at 25 volts/mm for 7 minutes. Separation occurs in lesε than 2.5 mm. The εcale bar = 240 urn.

FIGURE 9 illuεtrates the separation of the single-stranded (εε) fluoreεcent 34-mer (Panel A) and the εame 34-mer double-εtranded (dε) oligonucleotide (Panel B) , on an 8 mm, 35%T/12%C microgel (40X Image) , aε described in Example 10. In Panel A, the upper arrow points at the top of the gel, and the lower arrow points to the sε 34-mer fragment. In Panel B, the arrow pointε to the dε 34-mer fragment and the top

of the gel. Electrophoreεiε was carried out at 25 volts/mm for 8 minutes. The single-stranded 34-mer has migrated approximately 500 um into the gel, while the double-stranded 34-mer remains on the surface of the gel. The scale bar = 550 um.

DETAILED DESCRIPTION OF THE INVENTION

A. Definitions Gradient gel: A non-homogeneous gel where the polyacrylamide concentration (%T) increases from the top of the gel to bottom of the gel.

Homogeneous gel: A gel composed totally of one polyacrylamide concentration. Normal sieving process: A physical separation process occurring in polyacrylamide and agarose gels, where molecules being separated are of the order and/or smaller then pore size of the gel media (R _g < a) . Pore limit proceεε: A phyεical εeparation proceεε involving gradient gelε, where moleculeε reach terminal velocitieε that approach zero aε they reach the poεition in the gel where the pore size is equivalent to their effective size (radius of gyration) .

Pulεe field proceεε: A phyεical εeparation proceεε uεed to reεolve high molecular weight DNA involving application of pulεing and alternating orthogonal electric fieldε.

B. Methodε for Separating Macromoleculeε

The preεent methods generally describe micro- formatted electrophoretic separation of macromolecules, such as nucleic acids fragments, in relatively concentrated and crosslinked gels. These

gel compositions provide highly restrictive pores through which DNA molecules migrate by the force of high voltage gradients. The gel support compositions used to obtain the resolution for a particular range of double-stranded DNA fragmentε or single-stranded polynucleotideε/oligonucleotides are of much higher concentration than normally recommended or uεed by thoεe practicing the art. Results presented herein show separations of large size ranges of DNA fragments carried out in gel support media having pore sizes significantly smaller than the effective radius (or radius of gyration) of any of the macromolecules being εeparated (R _g » a) . Stated differently, the differential net migration occurs in a gel support having restrictive pore diameters relative to the effective radius of macromolecules being electrophoresed.

In the extreme, reεultε show separations occurring in gel supportε having gel poreε with average diameterε of double-εtranded DNA

(approximately 2 nanometerε) or single-stranded DNA (lesε than two nanometerε) .

Pore sizes of a gel support can be confirmed with protein standardε of known εizeε aε deεcribed herein thereby eliminating any poεεibility of trivial or artifactual cauεeε, εuch aε partial polymerization at the gel εurfaces (see Table 9) . The proceεs described in this invention is clearly different than the "normal gel sieving procesε", the "gradient gel or pore limit proceεε", or the "pulse field process", the extensive physical baεiε of which were discussed in the Background Section.

The process of the present invention is different from normal electrophoresis because effective separationε of DNA fragments are being carried out at

gel concentrationε where terminal velocities would have been reached. Terminal velocities are reached when the macromolecules being resolved no longer separate according to the normal gel sieving procesε. This difference is de onεtrated by comparing the gradient gel results for the SV40 DNA/Endo R fragments from Jeppeεen'ε work (Table 1 in the Background Section) with the microgel results for the 0X174 DNA/Hae III fragment in Examples 1, 2, 5, and 6. Both the SV40 DNA/Endo R fragments (114 to 1000 bp) and the 0X174 DNA/Hae III fragments (72 to 1,353 bp) are in approximately the same εize range. Jeppesen*ε gradient gel results show the largest fragment reaching its terminal velocity in approximately a 5.2% gel concentration, and the smallest fragment slowing considerably in velocity at approximately a 7.5% gel concentration.

In contrast, effective separations are observed by the present methods in homogeneous gel supports having restrictive pore diameters and gel concentrations that would otherwise produce terminal velocities and little or no resolution of the DNA fragments. According to the present invention, an improved electrophoretic separation is accomplished when higher voltages are applied than were previously utilized for macromolecules within a gel having matrix pore εizeε relative to macromolecular radii that under conventional voltages would result in "terminal velocity" type electrophoretic conditions. Pore diameters under these conditions are referred to as "restrictive pore diameters".

The øX174/Hae III electrophoretic separation resultε in Examples 1, 2, 5, and 6 show the complete separation of the same size range of DNA fragments in 12%T/6%C, 18%T/6%C, 22%T/6%C, and 26%T/6%C homogeneous

gelε, in running times of 2 to 3 minutes, and in total running distances of only 1 mm to 3 mm. For the gel concentrations used in these examples, protein studies show pore sizes of; ^" 8.3 nm for the 12%T/6%C gel, 5.5 nm for the 18%T/6%C gel, "4.5 nm for the 22%T/6%C gel, and "4.0 nm for the 26%T/6%C gel. Table 12 showε that the 0X174 DNA/Hae III fragments range in size (R _g ) from 83 nm for the 1,353 bp fragment, to 9 nm for the 72 bp fragment. Thus, with the possible exception of the 72 bp fragment in the 12%T/6%C gel case, all fragments are larger than the gel pore size. Actually, for the higher gel concentrations (18%T/6%C, 22%T/6%C, and 26%T/6%C) , the fragments are considerably larger (R _g » a) ; up to 20 times. Thuε, the εeparation process is clearly not the normal sieving proceεε. The methodology deεcribed herein provideε advantages of significant reduction in the linear dimension necessary for the separation, improved speed for the separation, and high resolution with extremely narrow banding patterns. Table 5 illustrateε advantages of this new methodology when compared with state of the art capillary electrophoresiε εystems for the separation of DNA fragments.

The resultε εhow the practical advantageε of the preεent methods in terms of combining speed of the separation, resolution, and significant reduction in running diεtance. Theεe three advantages are not present in the other separation techniques described previously.

TABLE 5

COMPARISON OF DNA RESTRICTION FRAGMENT ANALYSIS MICROELECTROPHORESIS (MEP) VERSUS STATE OF THE ART CAPILLARY ELECTROPHORESIS SYSTEMS

¹ Beckman PACE 2000 Capillary Electrophoresis System separating 0X174 DNA/Hae III fragments (72-1353 bp) described by Ulfelder et al in the Beckman Application Bulletin DS-810.

² BioRad HPE100 Capillary Electrophoresiε Syεtem separating 10 DNA fragments (88-1746 bp) described in the Biorad US/EG Bulletin 1679 and 1575.

³ Total running distance, required to achieve the full separation of all fragments.

Other important reεultε that have been εhown for thiε technology include: (1) the ability to achieve εingle baεe reεolution (εee Example 9) , indicating potential for the technology for DNA εequencing applications; (2) demonstration that single- and double-stranded oligonucleotides of the same length

can be differentiated by gels in the 30%T to 35%T range (see Example 10) providing unique separation applications; and (3) the ability of non-homogeneous gelε to further compreεs banding patterns, and reduce running distanceε (see Example 7) .

The process described by this invention now provides for the rapid separation of DNA molecules having effective radii that are relatively larger than the gel pore size in the gel support, i.e., the pores have a restrictive pore diameter. The micro-scale format, the higher concentration gel compositionε, and running conditionε (voltageε, etc.) used in the present methodology combine synergistically and lead to the resolution of the relatively large (in relation to pore size) linear DNA fragments as they migrate through the narrow pore structures. The following synergiεtic effectε contribute to the processes of the invention. The microgel format and high gel concentration inherently reduces Joule heating, and allow very high voltage gradientε to be uεed to drive the separationε in the relatively reεtricted gel pores. The higher voltage gradientε lead to faster separation times. The highly concentrated gels produce very narrow band widths and have greatly reduced diffusion. The combination of effects leads to the highly efficient separation procesε being obεerved.

The overall benefits of the new technique are: (1) the linear dimension neceεεary to carry out separations is significantly reduced to the range of millimeters; (2) the highly reεolved extremely narrow banding patternε are produced for a wide range of DNA fragmentε; (3) the separation times are extremely fast relative to the "normal sieving process", most being complete in minutes; and (4) the complete high

reεolution εeparation pattern can now be quickly detected and imaged with high reεolution and high εenεitivity uεing an epifluorescent microscope and electronic imaging systemε. The preεent invention iε particularly uεeful for nucleic acid fragment analyεiε including εingle or double stranded DNA molecules and fragments, RNA's, labelled (fluorescent, radioiεotope, etc.) nucleic acid fragmentε, modified and un-modified εynthetic oligonucleotides, and synthetic nucleic acid analogs. This methodology is also useful for separating other synthetic and natural macromolecules, including denatured proteins and polypeptides.

Thus, the present invention describeε a method of εeparating macromoleculeε within a gel εupport that compriεeε subjecting the macromolecules to an electric field oriented in a direction along a εingle axiε within and through the gel εupport.

The macromoleculeε are electrically charged macromolecules such aε nucleic acid moleculeε (RNA, DNA) or proteinaceouε molecules, such as detergent- protein complexes or other forms of protein having net charges proportional to their molecular weight, such that said charged macromolecules exhibit differential net migration within the gel support to an extent dependent on the molecular size of the macromolecules when subjected to the electric field.

The applied electric field is subjected to the macromoleculeε for a time period and in an amount εufficient to effect net migration of the electrically charged macromoleculeε in the direction of the electric field and within the gel support and to form a separation pattern comprised of a distribution and detectable separation of the macromolecules within the gel εupport. The εeparation pattern iε oriented along

the axis of the applied electric field and contains the charges macromoleculeε ordered according to their reεpective molecular weightε. By virtue of the reεolution attainable in the preεent method, due principally to the εmall pore εize provided by the gel support and the voltage settings used to form the electric field, the method forms a separation pattern (migration distance) in less than 2.0 centimeters (cm), typically in about 0.1 to 20 millimeters (mm), and preferably in about 0.5 to 5 mm.

The gel support, in one preferred embodiment, is about 3 to 40 percent acrylamide, as described further herein below. Other matrix materials are also contemplated for providing a matrix having the disclosed pore sizes.

The method includes subjecting the macromolecules to an electric field of about 5 to about 500 volts per mm of gel support along the axis length. Thus for gel supportε, of about .1 to 20 mm, voltageε of about 5 to 500 voltε per mm, or 50 to 5000 volts per centimeter are utilized.

The high voltages utilized in the present method can be tolerated without adversely affecting the gel support due to the thickness of at least one dimension of the gel εupport relative to the axiε of migration. Thus, the method utilizes a gel support having a narrow width in the direction perpendicular to the axiε length of about 0.05 to 2.0 mm, preferably about 0.05 to 1.5 mm, and more preferably about 0.05 to 1.0 mm in thickness, that facilitates minimum heating by the resistance produced by the gel support to the applied electric field. Particularly preferred are cylindrical gelε that have a diameter equal to the above thicknesses. Thus, the present invention deεcribes a method

for electrophoretic separation of a sample of electrically charged macromolecules that have been applied to a surface of a gel εupport and that are then εubjected to an electric field aε described herein. After the electric field has effected the desired net migration, a procesε referred to aε electrophoreεis, the separation pattern formed in the gel support is then detected by some evaluation means that allows for determining the relative location of each separated macromolecule specieε within the εample.

In a related embodiment the preεent method can be uεed for the isolation of a specieε of macromoleculeε from a sample by first electrophoresing the macromoleculeε and then isolating a discreet band

(specieε of separated molecules) from the gel support, thereby isolating the specieε of macromoleculeε.

Other permutationε of the preεent electrophoretic method will be apparent to one skilled in the art based on the present invention and the specific disclosureε made herein.

1. Sample Preparation

Sample εize (volume) and concentrationε uεed in thiε new methodology are dependent, along with other factorε, on (1) the dimenεion and geometry of the gel surface or microwell to which it is applied, and (2) to the inherent senεitivity of the detection εyste being used to analyze the banding patterns. With regards to sample volume, and asεuming εample concentration iε appropriate, the ratio of the εample volume in nanoliterε (nl) to gel surface area in square micrometers (u ² ) εhould be kept as low as posεible. For gel εurface areaε between 10 ⁶ um ² to 10 ⁴ um ² , the sample volume to surface area ratio is 1 to

10,000, preferably 1 to 100,000 and more preferably 1 to 1,000,000. Thuε for a one millimeter diameter microgel in capillary format having a εurface area of 785,400 um ² , a εample volume of 100 nl can be used, but a 10 nl volume is preferred, and a 1 nl volume is more preferred. In practice the choice of sample volume used might depend on other factors. For example, if only a low resolution separation is required or a longer running time is acceptable, a larger sample volume can be used because the resulting band broadening can be afforded. However, gross overloading should be avoided (ratios lower than 1 to 1000) . For very small sample volumes (<1 nanoliter) , appropriate micro-dispensing equipment and techniques are needed.

In regards to sample concentration, if one assumeε no limitation in ability to detect the resolved bands, then under a given set of conditions there is: (1) an upper level sample concentration which is not acceptable because of loss of band resolution, (2) a sample concentration range in which bands are adequately reεolved, and (3) a general improvement in reεolution aε sample concentration decreaseε. Preferred sample concentrations for use in the present invention are based on two sets of results presented herein using fluorescently labelled oligonucleotides or ethidium bromide stained ds-DNA fragments, which were each detected using the epifluorescent microscope and imaging system as described in the Detection and Example Section.

In one embodiment using fluorescent labels, fluorescent oligonucleotideε were labelled with fluoreεcein or other equivalent fluorophoreε (Texas Red, Rhodamine, etc.) at one label per molecule.

Synthetic oligonucleotides with a single fluorescein moiety per molecule covalently labeled at the 5'- terminal position can be commercially obtained from Synthetic Genetics, Inc. Mixtures of fluorescently labeled oligonucleotideε can thuε be prepared having concentrations ranging from 1 x 10 ^"13 to 5 x 10 ^'15 mole of individual oligonucleotide for sample volumeε in the range of 1 nl to 100 nl. Higher εample concentrations will lead to band broadening, especially in the narrower (<0.5 mm I.D.) or thinner (< 0.5 mm thicknesε) microgel formatε. Lower amountε of sample can be used if more sensitive detection is available, and/or higher quality or special optical arrangementε which reduce fluorescent background are used.

In another embodiment using ethidium bromide (EtBr) stained ds-DNA fragments, the relative fluorescence of EtBr-stained DNA fragmentε increases with their size. This iε becauεe the longer fragmentε intercalate increasing amounts of ethidium bromide. Thus, at an equivalent concentration, a 1000 bp fragment might have 10 timeε stronger fluorescent signal than a 100 bp fragment. For ethidium bromide stained ds-DNA fragments in the 100 to 1000 bp range, in sample volumeε of 1 nl to 100 nl, concentration rangeε of from 5 x 10 ^"14 to 5 x 10 ^"16 mole of each fragment are recommended. For fragmentε >1000 bp, the relative amount of fragment in moleε can be reduced proportionately to the increaεe in fragment εize. Sampleε at appropriate concentrationε are generally made up in a loading buffer which contains 0.5X to IX TBE buffer (IX TBE iε 0.089M Tris, 0.089M Borate, 0.2 mM EDTA, pH 8.0) and 10% to 30% (V/V) glycerol (a small amount of bromphenol blue and xylene cyanol dyes may be added to the loading buffer if

desired) . In caseε were ethidium bromide εtaining iε required the dye is typically added at a concentration of about 500 nanograms (ng) per milliliter (ml) . Other loading buffer formulationε common in the practice of gel electrophoreεis are also acceptable.

2. Gel Support Media Compositionε Gel compositions are an important part of the present invention in that they provide the appropriate pore size for εeparating a particular εize range of DNA fragmentε in a microgel formatε. In all cases, the gel concentrations are higher than would normally be expected, providing pore size where R _g » a for a given size range of DNA fragments. As described in the Background Section, in contrasting

"normal gel sieving" methods use gel concentrations to provide pore sizes at least equal to and larger than the radius of gyration of the molecules being separated. While homogeneous acrylamide gels are used as exemplary of the present invention, agarose, starch, natural and synthetic hydrogels, and many other materials are potentially applicable as media for producing a porus gel matrix. The following liεtε some of the potential gel producing media: polyacrylic acid; polymethylacryl acid; polymethylacrylamide; poly-N-iεopropyl acrylamide; polyacrylylglycinamide; polymethacrylylglycinamide; polyvinyl alcohol; polyvinyl pyrrolidone; polyethylene glycol; polypropylene glycol; hydroxyethyl celluloεe; hydroxypropyl cellulose; hydroxyethyl starch; modified dextrans, modified carrageenans, modified alginates, modified pectins, modified agars, modified xanthan gums, and modified furcellarans. It is also within the scope of this invention to include forms of solid

materialε (with nanometer or molecular εize channelε or poreε) which could be produced from glaεs, silicon, plastics, or other materials. These materials would not be considered gelε, but would εtill be εuitable aε a εeparation matrix. The main criteria for εelection of an appropriate media or material, would be the ability of the material or media itself (or with appropriate crosεlinkers, or composites with other polymers) to form pore structures in the 2 nm to 20 nm size range, or more preferably in the 1 nm to 10 nm size range. Other criteria would include mechanical and chemical stability, compatibility with the biological compounds (DNA, RNA, proteins) being separated, and little or no interference of the material with the subεequent detection of the compounds being separated.

Methods for determining gel matrix pore sizeε and molecular radii are generally well known. See for example the teachings of Jeppesen, Anal. Biochem. , 58: 195-207 (1974); Manwell, Biochem. J.. 165:487-495

(1977); and Campbell et al, Anal. Biochem. , 129:31-36 (1983) . The teachings of theεe and all other cited references are hereby incorporated by reference. Jeppesen teaches that the gel percent at which terminal velocity is approached defines the average pore size of the gel matrix in terms of the molecular radiuε of the macromolecule whose migration εtops (i.e., approaches terminal velocity). Thus Table 1 provides teachingε for correlating acrylamide gel percentε to the maximum nucleic acid molecule able to be effectively εeparated in the gel εupport by the normal sieving process.

With regard to homogeneous acrylamide gels the following table gives the broad concentration range of %C [percent crosslinker (bisacryla ide) relative to

total acrylamide monomers] that is acceptable for gels used in the present invention which range from 3%T to 40%T (total percent acrylamide and crosslinker) concentration.

TABLE 6

BROAD CROSSLINKER RANGE %T Total Acrylamide / %C Crosεlinker (Range) 40%T/(5-14%C) to 35%T/(5-13%C)

35%T/(5-13%C) to 30%T/(5-12%C) 30%T/(5-12%C) to 25%T/(4-ll%C) 25%T/(4-ll%C) to 20%T/(4-10%C) 20%T/(4-10%C) to 15%T/(3-8%C) 15%T/(3-8%C) to 10%T/(2.5-7%C)

10%T/(2.5-7%C) to 5%T/(2.5-6%C) 5%T/(2.5-6%C) to 3%T/(2.5-5%C)

With regard to homogeneous acrylamide gels the following table gives the broad concentration range of %C [percent crosslinker (bisacrylamide) relative to total acrylamide monomers] that are "optimal" for gels used in the present invention which range from 3%T to 40%T (total percent acrylamide and crosslinker) concentration.

TABLE 7 OPTIMAL GEL RANGE %T Total Acrylamide / %C Crosslinker (Optimal Range) 40%T/(9-12%C) to 30%T/(8-ll%C) 35%T/(9-12%C) to 25%T/(6-10%C)

30%T/(8-ll%C) to 20%T/(5-9%C) 25%T/(6-10%C) to 15%T(4-8%C) 20%T/(5-9%C) to 10%T/(3-6%C) 15%T(4-8%C) to 5%T/(2.5-5%C) 10%T/(3-6%C) to 3%T/(2.5-4%C)

In the case of polyacrylamide gels, these can be prepared in similar manner and technique common to the art of electrophoresiε. In general, fresh mixtureε of the appropriate percentage acrylamide/bisacrylamide (%T/%C) solutionε are made up in 1/2X to IX TBE buffer. The higheεt purity reagentε available εhould be uεed to prepare polyacrylamide gel εolutionε. As an example, 100 mis of a 26%T/6%C (non-denaturing) polyacrylamide gel solution would be made up aε followε: 23.4 g of acrylamide and 1.6 g of N,N'- methylenebiεacrylamide (biεacrylamide or crosslinker) are weighed out and dissolved in ^" 75 ml of distilled water. Now 10 ml of 10X TBE (108 g Tris-base, 55 g

Boric acid, and 9.3 g sodium EDTA in 1000 ml distilled water) is added and the εolution iε brought to a volume of 100 ml.

Other croεεlinking agentε are acceptable for producing polyacrylamide gelε with appropriate pore εizeε. However, becauεe of differences in reactivity (see Gelfi et al, Electrophoresiε. 2:213, 1981), each formulation εhould be checked with protein εtandardε to corroborate the actual pore εize as described herein to provide a restrictive pore size. Example 3

demonεtrateε the uεe of N,N'-dihydroxyethylene-biε acrylamide (DHEBA) aε a croεεlinking agent. Relative to BIS, the croεslinking reagent DHEBA is reported to produce a slightly smaller pore structure at high acrylamide concentrations and a more open structure at low gel concentrations (Connell et al, Anal. Biochem., 76:63, 1976). Resultε from Example 3 for the 18%T/6%C gel indicated that the pore size might be slightly smaller than when using the normal Bis crosslinker reagent. Other potential crosslinking agents include, but are not limited to: N,N'-bisacrylylcystamine (BAC) ; piperazine diacrylamide (PDA) ; diacrylylethylenedipiperidine (EDIP) ; diallyltartardiamide (DATD) ; triallyl citric triamide (TACT) ; ethylene diacrylate (ED) ; n-poly- ethylene glycol diacrylate (PEGDA) ; acetone- bisacrylamide (ABIS) ; dimethylethylenebisacrylamide (DMEBIS) ; and l,6-heptadiene-4-ol (HEP).

The reεtrictive pore diameter of a gel support of this invention is a diameter that is equal to or leεs than the effective radius of the macromolecule being electrophoresed. Insofar as a gel support is a sieve¬ like matrix comprised of multiple pores, it is understood that the restrictive pore diameter of a particular gel support can be expressed as the average pore diameter, that is determined empirically and represents the average of the multiple pores in the gel support. The determination of a macromolecule's effective radiuε and average pore diameter for a gel εupport can readily be determined as described herein. For nucleic acids, a useful restrictive pore diameter is in the range of 1 to 25 nanometers, as shown in Tables 10 and 11. Depending on the particular specieε of nucleic acid to be resolved, it is preferable to select a pore diameter that provides

higher reεolution for that species, as described further herein in the Description and the Examples. While homogeneous acrylamide gels are used as exemplary of this invention, non-homogeneous gels can alεo be uεed. However, completely unlike normal gradient gel εeparationε which depend on the pore limit effect for reεolution, εeparationε in non- homogeneous gels as taught in thiε invention occur in the moεt concentrated portions of the gel. For example, Jeppeεen'ε work deεcribed previouεly εhows that all DNA fragments resolved in the portion of the gradient gel where the concentration was < 10%T. In contrast, Example 7 of this invention shows separation for a similar range of DNA fragments occurs in the 18%T to 26%T region of the non-homogeneous gel. More importantly, these non-homogeneous gels produce rapid separations, with further significant reduction in running distance (< 1 mm for the 0X174 DNA/Hae III ladder) , and very narrow band widths < 10 um. In addition to the above advantages, the non-homogeneous microgels taught in thiε invention further improve reεolution by keeping all bands (lower as well as upper) relatively compreεεed during the εeparation proceεε. A non-homogeneous gel, in one embodiment is a gel εupport having a gradient of gel concentrationε, and therefore a gradient of decreaεing pore diameterε.

For the εeparation of double-εtranded DNA fragmentε on non-homogeneouε gels, microgels with concentrations of 12%T-30%T (providing pore εizeε which range from 8 nm to 3 nm) to 26%T-30%T (providing pore εizeε which range from 4 nm to 3 nm) can be uεed. However, the uεe of a εpecifir concentration range for the non-homogenouε gel can improve both the reεolution and εpeed of the εeparation for a given range of DNA

fragments. For example, if Lambda DNA/Hind III fragments, which range from 23,130 bp to 564 bp were to be separated, a 12%T-30%T non-homogeneous gel can be used. For the separation of the 0X174 DNA/Hae III fragments, which range from 1,353 bp to 72 bp, an

18%T-30%T non-homogeneous gel can be used. For the separation of the pBR322 DNA/Hae III fragments, which range from 587 bp to 8 bp, a 20%T-30%T non-homogeneous gel iε preferred. It should be pointed that the restriction ladder DNA fragments diεcuεsed above, are meant only as examples of DNA size ranges, and not as limits to the use of this invention. In general, for double-stranded DNA fragment separationε, non- homogeneous gels which range from 12%T-30%T to 26%T- 30%T can be used, depending upon (1) the range of fragments to be separated, (2) the degree of resolution required (1 bp, 10 bp, 100 bp, etc.), and (3) the required speed for the separation. For the separation of double-stranded DNA fragments, the actual length of the non-homogeneous gel can be from 1 mm to 10 mm long, or more preferably from 1 mm to 5 mm long.

For the separation of single-stranded DNA fragments, RNA, polynucleotides, and oligonucleotideε on non-homogeneouε gels, microgels with concentrationε of 12%T-40%T (providing pore εizeε which range from 8 nm to <2 nm) to 30%T-40%T (providing pore sizes which range from 3 nm to <2 nm) can be used. However, the use of a specific or narrower concentration range for the non-homogenous gel can improve both the reεolution and εpeed of the εeparation for a given εize range of single-stranded DNA fragments. This is particularly important for DNA sequencing applications where single base resolution is required for fragments that can range from 20 nucleotides (nt) to 1000 nt. For

example, if fragmentε which range from 1000 nt to 20 nt were to be separated, a 16%T-40%T non-homogeneous gel can be used. For the εeparation of εingle- εtranded fragmentε, which range from 500 nt to 20 nt, a 20%T-40%T non-homogeneous gel can be used. For the separation of fragments which range from 250 nt to 2 nt, a 26%T-40%T non-homogeneous gel is preferred. In general, for single-εtranded DNA fragment εeparationε, non-homogeneous gels which range from 12%T-40%T to 30%T-40%T can be used, depending upon (1) the range of fragments to be separated, (2) the degree of resolution required (1 bp, 10 bp, 100 bp, etc.), and (3) the required speed for the separation. For the separation of single-εtranded DNA fragmentε, the actual length of the non-homogeneous gel can be from 1 mm to 20 mm long, or more preferably from 2 mm to 15 mm long. Assuming a 10 um band resolution, the complete separation of a continuum of fragmentε differing by one nucleotide, which ranged from 20 nt to 1000 nt, would require a 10 mm long microgel.

With regard to the preparation of non-homogeneous microgels, they can prepared aε deεcribed in Example 7, or in manners generally known to the art. Regarding microformatε (capillary or microεlab) , diameters (I.D.) or thicknesses for microgels, percentages of crosslinker (%C) in polyacrylamide formulations (%T/%C) , pore size determinations, etc. ; the same general teachings apply to non-homogeneous gels as to homogeneouε gels. Both denaturing and non-denaturing gels can be uεed in thiε methodology. Denaturing gelε containing 6 to 7M Urea are preferred, particularly for εingle- εtranded DNA and RNA εeparationε. Becauεe microgel formatε (capillary tubeε or microεlabε) are uεed, only εmall amountε of gel solutions (1 to 2 ml) need to be

prepared for polymerization and pouring of micro-gels. Ammonium persulfate and N,N,N'N*-tetramethyl- ethylenediamine (TEMED) are uεed to catalyze the polymerization proceεε. The preparation of 10 mm (26%T/6%C) microgelε in 1 mm (I.D.) capillary tube format, will be uεed aε an example of how to prepare a microgel. Several 100 ul, 1 mm (I.D.) x 11.5 cm long capillary tubeε (Drummond) are cut into 5.75 cm lengthε uεing a capillary cleaving tool (Supelco # 2-3740). A volume of 1.5 ml of 26%T/6%C polyacrylamide/biεacrylamide solution (room temperature) is placed in a small teεt tube. The solution is de-gassed by placing it under a vacuum for about 30 seconds. After de-gasεing, 5 ul of a fresh ammonium persulfate solution (l g/1.5 ml distilled water) is added and gently mixed for several seconds. Next, 3 ul of TEMED is added and mixed. This level of ammonium persulfate and TEMED initiates polymerization in about 2 to 3 minutes. About 250 ul of the solution is then quickly transferred to a 1.5 ml Eppendorf tube (conical bottom) . Six of the 5 cm capillary tubes are placed vertically in the Eppendorf tube containing the 26%T/6%C polyacrylamide solution. The tubes are kept in vertical position while the solution fills them by capillary action to a height of about 2 to 3 cm. Within 2 to 3 minutes the gel begins to polymerize. It is important that the tubes be kept vertical until polymerization iε complete, and then allowed to εet for at leaεt 30 minutes before using them. After polymerization is complete, the gel tops can be flushed with running buffer (1/2X to IX TBE) and then cut into appropriate lengths. The capillary cleaving tool is used, and the tubes are first cut carefully 5 to 10 mm above the gel top for the upper buffer chamber,and then at the bottom to produce an

actual gel length of.10 mm. The tubes should be kept submerged in running buffer until they are ready to be used for microelectrophoresis.

3. Microgel Formats and Geometry

The microgel formats are a synergiεtic part of the methodology of thiε invention which allowε DNA fragmentε and other macromolecules to be separated within pore structures that are smaller than their molecular radius or radius of gyration. Microgelε can be prepared in either capillary tube, microεlab (εlab) , or microchanneled formatε. The sizes, lengths, and diameters and thicknesεeε are to be uεed with the sample volumes discussed above and the voltage gradients and other conditions discusεed below. It iε important to point out that ultra-micro separations on gels <0.05 mm in diameter or thickness could be carried out, however, special devices and techniques may be required to fabricate the gels and carry out the ultra-micro sample applications. Thuε, thiε methodology is not inherently limited to the microgel scaleε being deεcribed in more detail below.

For εingle εample applicationε, glaεs capillary tubes from 5 to 27 mm long and from 0.05 mm to 2 mm inner diameter (I.D.), and preferably 0.1 to 1.0 mm

I.D., are appropriate. The "actual gel length" within the tube will be shorter, since from 2 to 10 mm of the upper area is used for εample application and aε an upper buffer chamber; or to hold a piston type mold for preparing flat gel εurfaceε or εample micro-wells (see Figure 1) . The "actual gel length " for microgels can range from 0.1 to 25 mm, however gels from 1 to 20 mm are more ideal. Regarding microgel lengthε, the following referε to homogeneouε acrylamide gelε, non-homogeneouε gel lengthε were

described in the previous section. If longer and thicker diameter tubes (> 1 mm I.D.) are used, then Joule heating dissipation is slower, and the gel and separation process may begin to be adversely effected by overheating. Wider I.D. tubes or channels can be used, but will require lower voltages and slower running times, and therefore are less preferred. Also, thicker gels may lead to loss of detection sensitivity, by increasing background (autofluorescence, etc.).

Narrower tubes, microcapillaries, or slab gels (< 0.1 mm) can be used, but special ultra-micro sample application devices and techniques may be required. The choice of actual gel length (separation matrix) is dependent upon the size range of DNA fragments and the level of resolution that is required. In general, for the εeparation of a large range of double-stranded DNA fragments from 100 to 20,000 bp, that require only low to intermediate resolution, a 0.1 mm to 1 mm I.D. gel having an actual gel length of 2 to 10 mm is appropriate, with a 5 mm length being optimal. For separation of a range of single-εtranded DNA fragments (e.g., 10 to 500 bp) requiring high resolution (i.e., separation of fragments that differ by one nucleotide) a 1 mm to 0.1 mm I.D. gel having an actual gel length of 5 to 20 mm is appropriate, with a 15 mm length being optimal. For a narrower range of fragments a gel length of 10 mm can be used to produce high resolution. It should be pointed out that in many caεeε, the full gel length iε not required for the complete separation of the DNA fragmentε (see Examples) ; and the gel length used is determined more by the ease in preparing and manipulating a slightly longer microgel. Thus, it is convenient to refer to the length of the separation pattern rather than the

actual gel length. The separation pattern length is the distance from the top of the gel to the smallest macromolecule migrating in the matrix and forming the pattern of separated macromolecules in the gel matrix after electrophoresiε.

In further regard to gel lengthε, it εhould be pointed out that conditionε do exiεt under which longer gels could be used. If relatively smaller diameter or thinner gels are used, and/or lower voltages are applied, and much longer separation times are acceptable, then gelε longer then 25 mm could be used. However, now the advantage of rapid separation times and compact banding patterns suitable for imaging detection would be lost. For multiple sample applications, micro-slabs or icrochanneled arrangements can be used. These can be fabricated from thin glasε εheets (microscope cover slipε) , uεing plaεtic or glaεε εpacer materialε, or εpecially fabricated microchannelled glaεε, or a parallel array of capillary tubes. Preferred are slab gel supportε having a generally rectangular shape and a width as described herein. Particularly preferred are slab gels having multiple channels for εimultaneouε separation of different macromolecular compositions.

The lengths and thickness used in these multiple sample formats would correspond to those in the above description of capillary formats; lengths in the range of 0.1 to 25 mm, and thicknesses of from 0.05 to 2 mm can be used. Thinner gel formats (< 0.05 mm) can also be used, but devices and techniques for fabrication and ultra-micro sample application are required to accommodate the system. The overall width of these formats would be determined by the number of sample lanes; with 10 to 12 lanes being a typical number of

lanes. The width of the sample lanes would be determined by the gel thickness and sample volume to be applied. In general, for these types of formats with gel thicknesses of 0.1 to 1 mm, lane widths from 0.1 mm to 1 mm are contemplated. It is desirable to have the multiple sample arrangement (e.g., 10 lanes or more) so that the whole array of samples can be simultaneously detected and imaged in the field provided by 4X to 10X microscope objectives (4OX to 100X magnification) . The advantage in this iε that the complete εeparation of 10 or more εampleε could be detected and imaged εimultaneouεly. The fabrication of a εimple multiple well microslab gel is discuεεed in Example 4. Larger numberε of laneε or arrayε or lanes can also be utilized in practicing the invention.

An important aspect of microgel electrophoresis involves the relative flatnesε of both the top of the gel where the sample of macromolecules firεt enterε the gel matrix during the beginning of electrophoresis, and the bottom of the gel. This flatness is important for several reasons. The first reason involves the nature of the micro-separation process which produces highly compact bands with very narrow band widths. Thus, any distortions, irregularities, and unevenness on the surface of the top of the gel where the macromolecules enter the gel leadε to loεs of band resolution. The second reason involves the aspect ratio of the microgels width to length. Because these are extremely short gels, slight differences in lengths between the top and bottom of the gel will have different voltage gradients. This will cause the sample at thoεe locationε to run at εlightly different velocitieε leading to loεs of band resolution. For example, a

gel poured in a 1 mm I.D. capillary tube frequently has a meniscus, which produces a gel with a concave top. In a 5 mm gel the center may be 0.2 mm lower than the outer edges. This leads to a 4% difference in the voltage gradient between the center of the gel and the edges, which can cause some band broadening (bulleting) . For capillary formats this can be overcome by uεing a flat bottom piεton mold (plastic, teflon, and the like) inserted into the top of the capillary tube during the polymerization process to produce a flat even surface on the top of the gel. Another aεpect of microgel εeparation that leads to loss of band resolution involves the "edge effect". Distortion of the bands iε obεerved where the glaεs and gel meet. Thus, an appropriate mold that produces a sample micro-well in the top of the gel having a flat well bottom is contemplated for reducing the edge effect problem. Particularly preferred is a mold that forms a square indentation (well) within the top surface of the gel and having a flat well bottom surface and positioned as to keep the εample from contacting the walls holding the gel.

4. Sample Loading Sample loading for the methods of thiε invention involves the deposition of nanoliter quantities of sample on to a small area at the top of the gel or into a micro-well. Because of the nature of the micro-separation procesε care εhould be taken during the εample application proceεε. Samples can be applied uεing a microliter or εub-microliter εyringe (Hamilton # 701, # 7001N, and #7000.5N). The syringe should have a needle εize of 25 gauge (0.5 mm) or εmaller and a blunt tip (90° point) . While delivery by hand can be carried out, a micromanipulator

re¬ arrangement iε preferred. In thiε arrangement the syringe and microgel are secured so the syringe needle can be carefully inserted, by a mechanical drive mechanism, into the upper part of microgel chamber above the top of the gel. A magnifying lens can be used to view the procesε enabling it to be carried out with more preciεion.

Before the sample is applied, the gel top is preferably first flushed with running buffer (1/2X or IX TBE) and the upper microgel chamber (above the top of the gel) is left filled with running buffer. A syringe with a larger capacity (50 to 100 ul) can be used for this operation. Contact or abrasion to the top of the gel is preferably avoided. The microgel is preferably pre-electrophoresed prior to loading the sample, typically by subjecting the gel to 20 volts/mm for 2 minutes.

The sample is then applied to the top surface or micro-well of the gel. To that end, the appropriate syringe is filled with sample to the appropriate preselected volume and the tip of the needle is slowly inserted into the upper chamber of the microgel and brought within a distance of 0.5 mm or less above the center of the top of the gel or micro-well. The sample iε then slowly discharged from the syringe onto the top of the gel, with care taken not to agitate, dilute, or splash the sample against the walls of the microgel. The needle iε removed εlowly, and the micro-gel placed in the micro-electrophoreεiε unit (εee Figure 1) . The microgel iε immerεed completely into the lower buffer chamber εo that εome cooling can be provided to the gel. Care εhould be taken not to agitate the εample. Electrophoreεiε iε carried out thereafter, and preferably immediately after εample loading without delay. The microgel iε

electrophoreεed in the vertical position. Preferably the sample application is carried out with the microgel in the microelectrophoresiε chamber. Thiε requires a special arrangement of the sample application device (micromanipulator) with the microelectrophoresiε unit. Thiε arrangement has advantages in that the microgel need not be moved and electrophoreεis could be started even faster. The appropriate voltage is applied for the time necesεary to achieve the deεired εeparation aε deεcribed below. The microgel iε then removed from the microelectrophoreεiε chamber and detection and imaging of the banding patternε iε carried out aε deεcribed below.

5. Electrophoreεiε Voltage Conditionε The microgel formatε used in the present methodology allows much higher voltage gradients to be applied, than could normally be used in conventional slab gel formats. This iε becauεe the εhort and thin microgelε with relatively high gel concentrationε have reduced Joule heating, and have the capacity to diεεipate heat faεter than larger εcale gel formatε. Thus, higher voltage gradients can be used to drive the novel separation procesε of thiε invention leading to very faεt εeparation of DNA fragmentε into highly compact microεcopic banding patternε.

The microelectrophoreεiε chamber uεed for micro¬ gels in the capillary tube format is a relatively simple structure (Figure 1) . Running buffers used for microelectrophoresiε ranged from 1/2X to IX TBE; other bufferε common to the art of electrophoreεiε would be acceptable. Voltageε gradientε applied to the micro¬ gels are given in unitε of voltε/mm, which iε more applicable to microgelε than voltε/cm. The voltε/mm

units are determined by dividing the voltage reading on the power supply (for example, a BioRad Model 160/1.6 or Model 500/200 power supply) by the "actual gel length", within the micro-gel capillary tube or microslab gel. For example, if a microgel has an "actual gel length" of 5 mm, and 200 volts is the reading on the power supply meter, then an electric field in the form of a voltage gradient of 40 volts/mm waε applied. By way of compariεon 40 voltε/mm would be equivalent to 400 volts/cm in a large scale gel format.

Actual gel length, as opposed to separation distance, is the length of the gel media itself, which length provides resistance to current and so must enter into calculationε of voltage.

A wide range of voltage gradientε (applied electric fields) of 5 to 500 volts/mm can be applied to microgels depending upon the length, diameter or thickness, and polyacrylamide concentrations %T/%C (shown in Table 6) . In general, very low voltages (< 5 volts/mm) can be applied in practicing the preεent invention. However, the separation times will be very long. Alternatively, very high voltage gradients can be used, but may begin to cause band distortion. However, high voltages (such as voltages >500 volts/mm) could be used if they are switched on and off (0.1 sec. to 10 sec.) to allow time for heat dissipation to occur, and/or if relatively thinner gels are used. Pulse field modeε to applying electric fields are also contemplated, and can be delivered in a variety of means, varying the duration of the on or off phase of the field, repetitive cycles of on and off fields (pulsed) , and varying the field strength. In a related embodiment, it may be advantageouε, particularly with non-homogeneouε gelε, to vary the

field εtrength over the courεe of the electrophoresis process. For example, the separation process can be begun by applying a lower voltage field and then increasing the field to a higher value as the separation procesε proceedε. Alternatively, the electric field can be applied over the electrophoresis period in varying field directions, as to effect a net migration along a single axis within the gel εupport. The more acceptable and optimal voltages are those that will produce separation of a given range of fragmentε, with the appropriate level of reεolution in a time period of from 10 εecondε to 10 minuteε. Table 8 below εhows the range of acceptable voltages that can be used for both homogeneous or non homogeneous polyacrylamide gels, in the concentration ranges indicated.

TABLE 8

VOLTAGE GRADIENTS

Gel Concentration (%T/%C) Voltage (volts/mm)

40%T/(5-14%C) to 30%T/(5-12%C) 10 to 500

30%T/(5-12%C) to 20%T/(4-10%C) 7 to 300 20%T/(4-10%C) to 10%T/(2.5-7%C) 5 to 200

10%T/(2.5-7%C) to 3%T/(2.5-5%C) 5 to 100

6. Detection of Separated Macromolecules Detection and imaging of the fluorescent banding patterns produced by microelectrophoresiε of fluoreεcent DNA fragmentε can be carried out uεing an epifluoreεcent microεcope combined with a high reεolution imaging system. In general, fluorescent analysiε haε advantageε over other methodε

(radioisotopes, chromogenic stains, etc.) for detecting DNA fragment bands in electrophoresis gels. First, it is more sensitive than most chromogenic stains, which usually require involved time consuming washing procedures. Second, it is nearly as sensitive as radioisotope ( ³² P) , which has the disadvantage of requiring special handling and disposal. Therefore, the detection of fluorescent DNA fragments would be a preferred method. This new technology involves a unique synergistic combination of high resolution electronic imaging and epifluorescence detection meanε aε applied to microelectrophoresiε. For a deεcription of the state of the art of imaging, see Wick et al, BioTechniques, 7:262-268 (1989). High resolution electronic imaging technology [microchannel plates (MCP) , intensified Charged Coupled Devices (CCD) , cooled CCD, and Silicon Intensified Tubes (SIT) ] provideε the ability to easily carry out high resolution and sensitive detection and analysis of complex banding patterns in microgels. With these systems the complete electrophoretic separation pattern can be observed and recorded at any time during the process. Also, a large number of samples run in parallel micro-format arrangements could all be simultaneously analyzed with such syεtemε. A further advantage iε that theεe systems are relatively fixed with few moving parts. In addition to high resolution, some of the systems have the extremely high senεitivity which could be used to detect very low concentrations of fluorescent labelled DNA fragments (MCP systemε can carry out photon counting imaging) . Theεe systemε also have very faεt image acquiεition times (5-10 seconds) , and high contrast ability (250 gray levels) . All the above properties give electronic imaging

diεtinctive advantages for use with the preεent invention. Electronic imaging makes it posεible to know the relative poεition of all bandε within the gel at any given time, and to better resolve the center band fluorescence in diffuse or overlapped bands.

However, detection and analysiε of fluorescent banding patterns in microgels could also be carried out by laser or other scanning fluorescent detection syεtemε. A fixed fluoreεcent detection εystem that would simply monitor a position near the end of the gel and measure the fluorescent intensity as the bandε past through could also be used. These other formats are acceptable, especially if lesε coεtly detection εystem is desired. The actual fluorescent detection and imaging of the microgel banding patterns were carried out uεing an AO Model 2070 Vertical Fluoreεcence epifluoreεcent microεcope combined with a Hamamatεu C2400-97 Intenεified CCD Camera and a Hamamatsu Argus-10 Image Processor. This particular system has a sensitivity of ^" 10 photon/cm ² .sec. Real time and processed images from the Argus-10 syεtem waε viewed on a Sony RGB monitor, and recorded uεing a Mitεubiεhi HS-328UR Video Cassette Recorder. Photographs were later made from the video tape recordings using a Sharp GZ-P21 Color Video Printer. The excitation source on the epifluorescent microscope was a 50 Watt Hg εhort arc lamp.

For the detection of fluoreεcein labelled oligonucleotideε, the epifluoreεcent microεcope was filtered for excitation at ^" 490 nm and emission at ^" 520 nm. For detection of ethidium bromide fluorescence in ds-DNA fragments, the epifluorescent microscope was filtered for excitation at ^" 450 nm (in some caseε UV or < 400 nm excitation was used) and

emission at "610 nm. Most observations and imaging were carried at either 40X magnification (4X objective and 10X eyepieces, ^" 5 mm field of view), or 100X magnification (10X objective and 10X eyepieces, ~ 1.2 mm field of view) .

After completing the electrophoresiε the micro¬ gel is laid on a microscope slide, and set on the microscope stage. Fluorescence bands can be visually observed at the higher concentrations of fluorescent fragments. With the imaging system the banding patterns can be observed in real time, or enhanced by collecting the image for several seconds. In any case, the image acquisition process takes less than one minute. The processed images can be presented in color, in positive white on black, or negative black on white form.

Other detection systems can be used based on refractive index detectors, conductance, light absorption and the like. Exemplary are the detection systems described in U.S. Patent No. 4,909,919, which teachings are hereby incorporated by reference.

7. Gel Pore Size Determination with Proteins A significant amount of early experimental work haε been carried out regarding the determination of molecular εize (Stokeε radius, R _s ) of proteins in both homogenous [Ferguson, Metabolism 13:985-1002 (1964); Rodbard et al, Anal. Biochem. , 40:95-134 (1971); and Chrambach et al, Science, 170:440-451

(1971) ] and gradient polyacrylamide gels [Margolis et al, Anal. Biochem. , 25:347-362 (1968); Slater, Anal. Chem. , 41:1039-1041 (1969); Kopperschlager et al, FEBS Letterε. 5:221-224 (1969); Rodbard et al, Anal. Biochem. , 40:135-157 (1971); Manwell, Biochem. J. ,

165:487-495 (1977); Campbell et al, Anal. Biochem.. 129:31-36 (1983); and Margolis et al, J__ Chromatography. 106:204-209 (1975).

Electrophoresis of a large number of different proteins on gradient polyacrylamide gels shows a linear relationship between the hypothetical limiting pore size (the reciprocal of a limiting gel concentration, G _L ) and the cube root of the molecular weight. Manwell, Biochem. J.. 165:487-495 (1977). Manwell (1977) describes how the pore sizeε of polyacrylamide gelε at variouε %T/%C concentrationε can be eεtimated. For example, the limiting pore size (G _L ) for a 30%T gel is ^" 3.3 nm, 20%T gel is ^" 5.0 nm, and for a 10%T gel is "10 nm. These values are consistent with values determined by other workers using gradient gels. Campbell et al, Anal. Biochem. , 129:31-36 (1983); and Margolis et al, J. Chromatography. 106:204-209 (1975) .

Microgels were checked with low molecular weight proteins to obtain some corroboration of pore size based on the extensive past εtudieε uεing proteinε, and to verify that the obεerved reεultε were not due to artifacts, such as the top few millimeters of the microgel not being fully polymerized. The two proteins used in the study were Cytochrome C (MW 12,400, Rs = ^" 1.5 nm) , and alpha-Lactalbu in (MW 16,000, Rε = ^" 1.9 nm) . Theεe two proteinε were picked becauεe they are near the diameter of double- stranded DNA ("2 nm) , and Cytochrome C is colored and easily detected in the microgel. These proteinε were applied to microgelε, and run under the εame basic conditions (20 to 30 volts/mm for 4 to 5 inuteε; polarity reversed for Cytochrome C runs) that were used for DNA fragments and oligonucleotides. The

results are shown in Table 9 below.

TABLE 9

CYTOCHROME C OR ALPHA-LACTALBUMIN MIGRATION

INTO MICROGELS

%T/%C Gel Migration Distance (mm)

35%T/12%C 0.0 30%T/10%C 0.5 - 1

26%T/8%C 1 - 2

22%T/6%C 3 - 4

18%T/6%C 7 - 9

The migration distances shown in Table 9 were very carefully determined by microscopic examination of the colored band (cytochrome C) or the stained band (alpha-Lactalbu in) . There is basically no penetration of these proteins into the 35%T/12%C gel and very little entry (0.5 to 1 mm) for the 30%T/10%C gel, under the same conditions that allowed single- stranded fluorescent oligonucleotides in the 20-mer to 40-mer range to be effectively separated (εee Section 3.3). The obεerved 35%T to 30%T limiting pore size for these proteins is consistent with the classical studies. Manwell, Biochem. J. , 165:487-495 (1977) ; Campbell et al, Anal. Biochem.. 129:31-36 (1983) ; and Margolis et al, . Chromatography. 106:204-209 (1975) . The mobility of the proteins increaεeε for the

26%T/8%C and 22%T/6%C gels, but they migrate only about the same distance that was needed to resolve the full 0X174 DNA/Hae III ladder. In the 18%T/6%C gel the protein mobilities increase significantly, and they migrate beyond total separation distance required

to resolve the 0X174 DNA/Hae II ladder.

C. Gels For Separating DNA Fragmentε By the present invention a set of conditions has been diεcovered by which DNA fragmentε can be quickly separated in microgel formats in polyacrylamide gels at higher concentrations (%T/%C) than is normally expected or used. The highly concentrated polyacrylamide gels, for example, provide pores, such that R _g >> a, for the given range of DNA fragments being separated. Polyacrylamide gels have been primarily used in this invention for two reasons: one, they provide the most widely available media for producing a gel matrix with appropriate pore εizeε; and two, a great deal of prior experimental work verifieε the nature and εize of the polyacrylamide poreε. However, the basic separation process and methodology is not limited to polyacrylamide gels alone but can be applied to other gel support medium as described herein.

Table 10 below gives the ranges for homogeneous polyacrylamide gel concentrations (%T) and the corresponding ranges of double-stranded DNA fragments that can be optimally resolved within those gel ranges by this technique. Although some of the low gel concentrations shown in Table 10 have previously been described in the literature for separating macromolecules, these gel concentrations have not previously been utilized in the manner described herein where moleculeε have molecular radii larger than the pore εize of the gel εupport (matrix) . Table 10 alεo giveε the length and radiuε of gyration for the DNA fragmentε.

TABLE 10

OPTIMAL POLYACRYLAMIDE GEL RANGE AND PORE SIZE FOR DOUBLE-STRANDED DNA FRAGMENT SEPARATIONS

assumes the optimal crosslinker (%C) from Table 7,

Table 11 below gives the ranges for homogeneous polyacrylamide gel concentrations (%T) and the corresponding ranges of single-stranded DNA fragments that can be optimally resolved within those gel ranges by this technique.

Although some of the low gel concentrations shown in Table 11 have previously been described in the literature for separating macromoleculeε, eg, 6% DNA sequencing gels, these gel concentrations have not previously been utilized in the manner described herein where molecules have molecular radii larger than the pore size of the gel support (matrix) . Table 11 also gives the length and radiuε of gyration for the DNA fragmentε.

TABLE 11

OPTIMAL POLYACRYLAMIDE GEL RANGE AND PORE SIZE FOR SINGLE-STRANDED DNA FRAGMENT SEPARATIONS

¹ Aεsumes the optimal crosslinker (%C) from Table 7.

Suitable %T/%C ranges for non-homogeneous polyacrylamide gels, and their corresponding DNA fragment size ranges were given in the previous section.

The optimal gel ranges for separating other linear macromolecules can be determined from knowledge of their molecular radius. Also, the approximate pore limit sizes for the various gel compositions can be determined more accurately using standard slab gel procedures and protein molecular weight standardε for which the molecular radii are known.

For the gel compositions, microgel formats, voltage gradients, and other conditions described above this methodology provideε relatively faεt εeparation (1 to 10 inuteε) for DNA fragmentε that range from 10 to 10,000 bp. The methodology provideε relatively high resolution separations for DNA fragments in the range of 10 to 500 bp and intermediate for DNA fragments in the 500 to 5000 bp

range.

Thus, in another embodiment, the present invention contemplates a gel matrix (support) for use in the present microelectrophoresis methodε. A microelectrophoretic gel εupport medium is therefore contemplated according to the present teachings and based on the discoveries reported herein. The medium has a matrix comprising about 3 to about 40 percent acrylamide and the matrix is about 0.1 to 20.0 millimeters in length along a first axis oriented in the direction of migration of the macromolecules being εeparated.

To accommodate cooling during electrophoresis to counteract the Joule heating that arises during the electrophoretic separation, the medium is about 0.05 to 2.0 millimeters in width perpendicular to said first axis, preferably about 0.1 to 1.0 millimeters in width.

In a preferred embodiment, the gel support is in a cylindrical format, such as in a capillary tube as described herein, preferably a tube having an inner diameter of about 0.1 to about 1.5 millimeters in width, and particularly about 1.0 mm in width.

In an embodiment particularly suited for DNA sequencing of macromolecules of about 2 to about 500 nucleotides in length, a gel matrix in a capillary tube is particularly preferred having an inner diameter of 0.1 to 0.5 mm, an acrylamide concentration of about 15 to about 35 percent acrylamide, and a crosεlinker concentration of about 4 to 12 percent biεacrylamid .

The capillary tube can be glass, plastic or any other containment material so long as the ability to detect the migration of the macromolecules in the electrophoretic medium is not obscured.

In a related embodiment, a gel support medium can be expressed in terms of the average pore diameter within the gel, as diεcuεsed earlier. Thus, a gel support medium of thiε invention can have poreε that provide an average pore diameter of from 1 to 25 nanometerε (nm) . Actual pore diameterε can varied depending on the particular molecule to be separated. Where the separation is to resolve single-stranded nucleic acids from double-stranded nucleic acids, pore diameters of 1 to 8 nm, preferably 1 to 4, and more preferably 1 to 2 nm, are contemplated.

EXAMPLES

The following examples are given for illustrative purposes only and do not in any way limit the scope of the invention.

Examples 1 to 6 demonstrate the separation of the 0X174 DNA/Hae III ladder in homogeneous polyacrylamide microgels which range in concentration from 12%T/6%C to 26%T/6%C. These examples alεo de onεtrate the multiple εample εlab format, and the uεe of other croεεlinking reagents. Example 7 demonstrateε the εeparation of the 0X174 DNA/Hae III fragmentε in a non-homogeneous 18%-26%T/6% microgel. The 0X174 DNA/Hae III ladder contains eleven ds-DNA fragments ranging in size from 1,353 to 72 bp. Table 12 gives the fragment size (bp) , approximate molecular weight (MW) , length of the fragment (L) , and the radius of gyration (R _g ) determined by formula [Benoitet al, J _^ . Phvs. Chem.. 57:958-963, (1953)] given below:

R _g = {pL/3 [1 - p/L + p/L exp(-L/p)]) ^/2 (1)

where p = the perεiεtence length of DNA ( ^" 50 nm) , and L = the length of the fragment (bp X 0.34 nm) .

TABLE 12

SIZE AND RADIUS OF GYRATION (R _g )FOR DOUBLE-STRANDED FRAGMENTS IN THE 0X174 DNA/HAE III LADDER

1. SEPARATION OF (6X174 DNA/Hae III FRAGMENTS ON 12%T/6%C MICROGELS

For these separationε, homogeneous non-denaturing 12%T/6%C microgels (5 mm to 6 mm actual gel length) in 15 mm X 1.04 mm (I.D.) capillary tubes were used. The microgels were first pre-run at 33 volts/mm for 3 minutes. The running buffer was 1/2X TBE containing ethidium bromide at a concentration of 200 ng/ml. After pre-running, the microgel was flushed with running buffer. Then, approximately 50 nl of a sample containing 0X174 DNA/Hae III restriction fragments at a concentration of 5.7 x 10 ^"14 mole/ul in sample loading buffer (0.5X TBE, ethidium bromide 200 ng/ml, and 20% glycerol) was added to the top of the microgel. The 50 nl sample applied to the microgel

contained "2.8 x 10 ^'15 mole of each DNA fragment. The microgel was immediately electrophoresed (22° C) at 33 voltε/mm for 2 minuteε. Both the xylene cyanol and bromphenol blue dye markers migrate off the gel during the run. The microgel was analyzed and imaged using the epifluorescent microscope imaging syεtem described in the detection section. Epifluorescent detection for ethidium bromide fluorescence was carried out at an excitation of 450 nm and an emisεion of 610 nm. The gel was examined and imaged at 4OX and 100X magnifications.

Figure 2 shows the 4OX magnified image of the complete 0X174 DNA/Hae III separation. Table 13 gives the εeparation diεtances between the top of the gel and each band, and the approximate band width for each fragment shown in Figure 2.

TABLE 13

0X174 DNA/Hae III SEPARATION ON 6 mm 12%T/6%C MICROGEL

Figure 3 is a 10OX image showing the separation of the upper bands. Figure 4 is another separation showing the intermediate bands of the 0X174 DNA/Haelll restriction fragment ladder on a 5 mm, 12%T/6%C microgel (100X Image) . This separation was carried out at 40 volts/mm for 1 minute and 15 εecondε. The fragmentε were separated in a distance of 2.03 mm measured from the top of gel to the 194 bp band. Separation of the 281 and 271 bp fragments is observed.

The results from this first example show that the 0X174 DNA/Hae III fragments were reεolved in a distance of ^" 3.3 mm. The band widths observed ranged from 36 um to 96 um. Except for the εhorter fragmentε (9, 10, and 11) , all of the other fragments exceed the pore size of a 12%T/6%C gel.

2. SEPARATION OF <6X174 DNA/Hae III FRAGMENTS ON 18%T/6%C MICROGEL

This example demonstrateε the separation of the 0X174 DNA/Hae III fragments on a 18%T/6%C microgel. For this separation, a homogeneous non-denaturing 18%T/6%C microgels (5 mm actual gel length) in 15 mm X 1.04 mm (I.D.) capillary tube was used. The microgel was first pre-run (electrophoresed) at 30 volts/mm for 3 minutes. After pre-running, the microgel was flushed with running buffer. Approximately 20 nl of a 0X174 DNA/Hae III sample at a concentration of 5.7 x 10 ^"14 mole/ul in sample loading buffer was added to the top of the microgel. The 20 nl sample applied to the microgel contained 1.2 x 10 ^"15 mole of each DNA fragment. The microgel waε immediately electrophoreεed ( ^" 22° C) , at 30 volts/mm for 1 minute and 50 secondε. Both the xylene cyanol and bromphenol

blue dye markers migrated off the gel during the run. The microgel was analyzed and imaged using the epifluorescent microεcope imaging εystem described in the detection section. The gels were examined and imaged at 4OX and 100X magnification. Figure 5 shows the 100X magnified image of the upper bands and some of the intermediate bands. The Table 14 below giveε the separation distances between top of the gel and each band, and the approximate width of each band. The results in Table 14 show that the 0X174

DNA/Hae III fragments were resolved in a distance of only "1.7 mm, in lesε than 2 minuteε. The band widths ranged from 18 um to 60 um in obεerved width. The size (R _g ) of all the 0X174 DNA/ Hae III fragments clearly exceeds the pore size of an 18%T/6%C gel, but they were nonethelesε reεolved in a very εhort time and diεtance.

TABLE 14

0X174 DNA/Hae III SEPARATION ON 5mm 18%T/6%C MICROGEL

3. SEPARATION OF ΦX174 DNA/Hae III FRAGMENTS ON 18%T/6%C MICROGEL USING DHEBA CROSSLINKER

Thiε example demonεtrates that separation of the 0X174 DNA/Hae III fragments could be carried out on microgels which were produced using a crosslinking reagent other than Bis (N,N ^f -methylene-bis-acrylamide. In thiε case N,N'-dihydroxyethylene-bis-acrylamide (DHEBA) was used as the crosslinking reagent to form an 18%T/6%C microgel. For this separation, a homogeneous non-denaturing 18%T/6%DHEBA microgel (6 mm actual gel length) in a 15 mm X 1.04 mm (I.D.) capillary tube was used. The microgel was firεt pre- run (electrophoreεed) at 30 volts/mm for 2 minuteε. After pre-running, the microgel waε fluεhed with running buffer. Approximately 20 nl of a 0X174 DNA/Hae III εample at a concentration of 5.7 x 10 ^"14 mole/ul in sample loading buffer was added to the top of the microgel. The microgel was immediately electrophoresed ( ^" 22° C) , at 30 volts/mm for 2 minutes and 30 seconds. Both the xylene cyanol and bromphenol blue dye markers migrated off the gel during the run. The gels were examined and at 4OX and 10OX magnification. The resultε showed that complete separation of all the fragments had occurred in less than a 2 mm running distance. Nearly the same results are obtained using the DHEBA crosslinker, as the when using the BIS crosslinker. However, the pores are slightly smaller for DHEBA, as the total running distance is shorter.

4. SEPARATION OF ΦX174 DNA/Hae III FRAGMENTS ON 18%T/6%C MULTIPLE LANE SLAB FORMATTED MICROGEL

This example demonstrates separation of the 0X174

DNA/Hae III fragments on a multiple lane, slab formatted microgel. An 18%T/6%C slab microgel (approximately 1 cm X 1 cm X 0.75 mm) was formed between two microscope coverslipε (22 mm X 22 mm) . The εpacerε and a 3 well comb for the microgel were cut to εize from normal 0.75 mm plaεtic spacer and comb material (BioRad) . The 3 wells formed in the microgel were approximately 1 mm wide and 2 mm deep, with about a 2 mm spacing between the wells. The microgel was first pre-run at 30 volts/mm for 2 minutes. After pre-running, the microgel was flushed with running buffer. Approximately 20 nl of a 0X174 DNA/Hae III sample at a concentration of 5.7 x 10 ^"14 mole/ul in εample loading buffer was added to each of the three wells in the slab microgel. The microgel waε immediately electrophoresed ("22° C) , at 30 volts/mm for 2 minutes. Both the xylene cyanol and bromphenol blue dye markers migrated off the gel during the run. Each of the lanes on the εlab microgel gel was examined at 40X and 100X magnification. The reεults showed complete separation of all the fragments had occurred in about a 2 mm running distance. Esεentially, the εame reεultε were obtained on this multiple lane slab formatted 18%T/6%C microgel as on the previouεly deεcribed capillary formatted microgel.

5. SEPARATION OF d>X174 DNA/Hae III FRAGMENTS ON 22%T/6%C MICROGEL

This example demonstrates separation of the 0X174 DNA/Hae III fragments on a 22%T/6%C microgel. For this separation, a non-denaturing microgel (11 mm actual gel length) in a 15 mm long (0.77 mm I.D.) glass capillary was used. The microgel was first pre- run (electrophoresed) at 30 volts/mm for 2 minutes. After the pre-running, the microgel was flushed with running buffer. Then, approximately 20 nl of a 0X174 DNA/Hae III sample at a concentration of 5.7 x 10 ^"14 mole/ul in sample loading buffer was added to the top of the microgel. The 20 nl sample applied to the microgel contained 1.2 x 10 ^"15 mole of each DNA fragment. The microgel was immediately electrophoresed ( ^" 22° C) , at 33 volts/mm for approximately 5 minuteε. The microgel waε examined and imaged at 4OX and 100X magnificationε. Figure 6 εhowε the results. The doublet bands containing the 281 and 271 bp fragments that appear blurred into a single band in the present photograph can be resolved into two diεcreet bandε when viewed with video diεplay apparatuε. The results indicate that resolution of all 0X174 DNA/Hae III fragments can be achieved on a 22%T/6%C microgel. These are important reεultε becauεe they εhow that relatively large DNA fragmentε can be εeparated on very highly concentrated and croεεlinked gelε where R _g » a.

6. SEPARATION OF 6X174 DNA/Hae III FRAGMENTS ON 26%T/6%C MICROGEL

Thiε example demonεtrated the εeparation of the 0X174 DNA/Hae III fragments on a 26%T/6%C micro-gel.

For this separation, a non-denaturing 26%T/6%C micro¬ gel (5 mm actual gel length) in a 15 mm long X 1.04 mm (I.D.) glass capillary was used. The microgel was first pre-run (electrophoreεed) at 40 voltε/mm for 2 minuteε. The running buffer was 1/2X TBE containing Ethidium Bromide at a concentration of (100 ng/ml) . After the pre-running, the microgel was flushed with running buffer. Then, approximately 20 nl (0.2 ul) of a 0X174 DNA/Hae III εample at a concentration of 5.7 x 10 ¹⁴ mole/ul in εample loading buffer waε added to the top of the microgel. The 20 nl εample applied to the micro-gel contained 1.2 x 10 ¹⁵ mole of each DNA fragment. The microgel waε immediately electrophoreεed ("22° C) , at 40 voltε/mm for 2 minuteε 55 εecondε. The microgel waε then analyzed and imaged uεing the epifluoreεcent microscope imaging system described in the Detection section. The gels were examined and imaged at 4OX and 100X magnifications. Unfortunately, photographic prints of the video image lack the resolution to show the finely separated bands that can be observed on the VCR recordingε. The VCR results showed that except for the upper two bands (1,353 bp and 1,075 bp) the 0X174 DNA/Hae III fragments were resolved in a running distance of lesε than 1.5 mm. Theεe are important reεultε becauεe they εhow that relatively large DNA fragments can be rapidly separated on very highly concentrated and crosεlinked gelε where R _g >> a.

7. SEPARATION OF C&X174 DNA/Hae III FRAGMENTS ON

A NON-HOMOGENEOUS 18%-26%T/6%C MICROGEL

Thiε example demonεtrated the εeparation of the 0X174 DNA/Hae III fragmentε on a 18%-26%T/6%C micro- gel. For thiε εeparation, a non-denaturing 18%-

26%T/6%C microgel (10 mm actual gel length) in a 15 mm long X 1.04 mm (I.D.) glass capillary was used. The non-homogeneous microgel was prepared by layering 5 mm of a 18%T/6%C polyacrylamide solution over 5 mm of a 26%T/6%C polyacrylamide solution just prior to polymerization (about 30% less TEMED was used in order to extend the polymerization time long enough to layer the solutions) . This procedure produceε a 1 mm to 2 mm gradient (18%T to 26%T) at the microgelε mid εection. The microgel waε first pre-run at 30 volts/mm for 3 minutes. The running buffer was 1/2X TBE containing Ethidium Bromide at a concentration of (100 ng/ml) . After the pre-running, the microgel was flushed with running buffer. Then, approximately 20 nl (0.2 ul) of a 0X174 DNA/Hae III sample at a concentration of 5.7 x 10 ¹⁴ mole/ul in sample loading buffer was added to the top of the micro-gel. The 20 nl sample applied to the micro-gel contained 1.2 x 10 ¹⁵ mole of each DNA fragment. The microgel was immediately electrophoresed ( ^' 22° C) , at 30 volts/mm for 4 minuteε.

The microgel waε then analyzed and imaged uεing the epifluoreεcent microεcope imaging εyεtem deεcribed in the Detection εection. The gelε were examined and imaged at 40X and 100X magnificationε. Unfortunately, video printε lack the reεolution to εhow the finely εeparated bandε that appear on the VCR recordings. The VCR results showed that the 0X174 DNA/Hae III fragments in a banding pattern of lesε than 1 mm, at the gels mid section. These are important results because they show that the non-homogeneous gel can further compresε the banding pattern, and keep lower fragmentε band widthε from broadening relative to the upper fragmentε. The band widths for some of the upper fragments appear to be 10 um or lesε.

8. SEPARATION OF FLUORESCENT OLIGONUCLEOTIDES WITH TWO BASE RESOLUTION

The present example demonstrates the separation of a sample mixture containing 18-mer, 21-mer, 23-mer, 32-mer and 41-mer εingle-stranded oligonucleotides. Each of the oligonucleotides was labelled at its 5'- terminal position with fluorescein. The εample was run on a denaturing (6M Urea) 35%T/12%C microgel (8 mm actual gel length) in a 15 mm X 0.77 mm (I.D.) glasε capillary tube. The microgel waε first pre-run (electrophoresed) at 30 volts/mm for 2 minuteε. The running buffer used was 0.5X TBE. After the pre- running, the microgel was fluεhed with running buffer. Approximately 50 nl of the fluoreεcent oligonucleotide mixture at a concentration of 5.0 x 10 ^*13 mole/ul in εample loading buffer was added to the top of the microgel. The 50 nl sample applied to the microgel contains 2.5 x 10 ^"14 mole of each fluorescent oligonucleotide. The microgel was immediately electrophoresed, at room temperature ( ^" 22 C) at 30 volts/mm for 6 minutes.

The microgel waε then analyzed and imaged uεing the epifluoreεcent microscope imaging syεtem deεcribed in the detection section. Epifluorescent detection for fluorescein fluorescence was carried out with excitation at 450 nm and emisεion at 520 nm. The gelε were examined and imaged at 40X and 100X magnificationε. The reεultε are εhown in Figure 7 (100X image) . The reεults show the complete resolution of all the oligos, including the 21-mer and 23-mer (the 41-mer is out of the observable field for the 10X objective) . The complete separation, with two base resolution was achieved in a total running diεtance of leεε than 2 mm from the top of gel to the

18-mer. The bands are approximately 50 um in width.

9. SEPARATION OF FLUORESCENT OLIGONUCLEOTIDES SHOWING ONE BASE RESOLUTION

This example shows the separation of a sample mixture containing 18-mer, 21-mer, 22-mer, and 32-mer oligonucleotides (single-εtranded) run on a denaturing 35%T/12%C microgel (10 mm actual gel length) in a 15 mm X 0.77 mm (I.D.) glaεs capillary tube. Each of the oligonucleotides was labelled at its 5'-terminal position with fluorescein. The microgel was first pre-run (electrophoresed) at 25 volts/mm for 2 minute. The running buffer used was 0.5X TBE. After pre- running, the micro-gel waε fluεhed with running buffer. Approximately 50 nl of the fluorescent oligonucleotide mixture at a concentration of 5.0 x 10 ^"13 mole/ul in sample loading buffer was added to the top of the microgel. The 50 nl sample applied to the microgel contains 2.5 x 10 ^"14 mole of each fluorescent oligonucleotide. The microgel was immediately electrophoresed, at room temperature ("22 C) at 30 volts/mm for 7 minuteε. The gelε were examined and imaged at 4OX and 10OX magnificationε. The results are shown Figure 8 (100X image) . The resultε show the complete resolution of all the oligonucleotides, including the 21-mer and 22-mer (the 32-mer is out of the observable field for the 10X objective) . The complete separation, with one base reεolution, waε achieved in a total running diεtance of only 2.5 mm (top of gel to the 18-mer) . The bandε are approximately 50 um in width. Similar experiments were carried out uεing 30%T/10%C microgelε, with the εame reεolution obtained in slightly longer running distance. These resultε clearly εhow the potential of

thiε technology for DNA εequencing applications.

Examples 8 and 9 show the separation of fluorescent oligonucleotides which ranged in size from an 18-mer to a 41-mer ("mer" refers to the number of nucleotideε in the sequence) . The separationε were carried out in homogeneouε polyacrylamide microgelε at high gel concentration (35%T/12%C) .

10. SELECTIVE SEPARATION OF DOUBLE-STRANDED vs.

SINGLE-STRANDED OLIGONUCLEOTIDES ON 35%T/12%C MICROGELS

This example showε that a 35%T/12%C gel concentration haε a pore size such that it will allow migration of single-εtranded (εε) oligonucleotideε, but will effectively stop the migration of the double- stranded (ds) counterparts. In these experiments both a fluorescent sε 34-mer oligonucleotide, and a fluorescent ds 34-mer oligonucleotide (the same 34-mer sequence hybridized with its complement oligonucleotide) were run under exactly the same conditions on a 35%T/12%C microgel. The microgels were 8 mm long, in 15 mm X 1.04 mm (I.D.) glass capillary tubes. Both sampleε were run at 25 voltε/mm for 8 minuteε. Figure 9 εhowε the results (4OX images) . The sε 34-mer haε migrated approximately 500 um into the gel (Figure 9A) , while the dε 34-mer iε effectively excluded from the gel and remainε on the εurface (Figure 9B) . Theεe results indicate that a 35%T/12%C gel has an effective pore diameter of approximately 2 to 3 nm, very close to the diameter of double-stranded DNA (2 n ) . This is an important observation which is consiεtent with the gel pore εize that would be expected for thiε gel concentration (2.8

nm) . This result also corroborates the resultε of pore size determinations made with protein standardε. For important practical purposes, gels at approximately the 35%T/12%C level have pore sizes which excluded ds DNA, but still allow sε DNA moleculeε to be reεolved.

The above obεervation provideε an additional embodiment for the preεent invention, namely a method for εeparating ss nucleic acidε from dε nucleic acidε. In highly restrictive pore diameter-containing gel supports, such as about 1 to 2 nm, ds DNA will not appreciably enter the gel support. Electrophoresiε in these gels will provide a fast system for bulk separation of ss from ds nucleic acid. In less restrictive pores, εuch as 3 to 8 nm, and preferably 3 to 4 nm, ss nucleic acids will more rapidly enter the gel and will fractionate based on εize, and similar ds nucleic acids will enter the gel more slowly, separating from the ss nucleic acids, and will alεo independently fractionate according to εize.

The present invention also contemplates a method for separating εingle-stranded (ss) nucleic acid molecules from double-stranded (ds) nucleic acid molecules present in a heterogeneous composition containing εs and dε nucleic acid moleculeε. The method comprises applying the composition to a gel support and subjecting the composition to an applied electric field oriented along a single axis within the gel for a time period εufficient to effect migration of the εε nucleic acid molecules into the gel support in the direction of the oriented field. The gel support typically has an average pore diameter of 1 to 25 nm, preferably 1 to 8 nm, and more preferably 1 to 4 nm. Where the purpose is to prevent any ds nucleic acid from entering the gel support, a pore diameter of

1 to 2 nm iε preferred, aε demonstrated above. An exemplary gel is the 35 percent acrylamide gel described above. Thus a εs nucleic acid εeparating gel εupport typically haε 30 to 40 percent acrylamide, preferably at leaεt 35 percent acrylamide.

The foregoing description and the examples are intended as illustrative and are not to be taken as limiting. Still other variations within the spirit and εcope of thiε invention are poεεible and will readily preεent themεelves to those skilled in the art. Other embodiments are within the following claims.