Anal. Chem. 1995, 67, 1197-1203
`
`Rapid Sizing of Short Tandem Repeat Alleles Using
`Capillary Array Electrophoresis and
`Energy-Transfer Fluorescent Primers
`Yiwen Wang,™ Jingyue Ju,f Brooke A. Carpenter,* Jeanette M. Atherton,*
`George F. Sensabaugh,* and Richard A. Mathies™
`Department of Chemistry and Forensic Science Group, School of Public Health, University of California,
`Berkeley, California 94720
`
`Genetic typing of the short tandem repeat (STR) polymor-
`phism HUMTHOl has been performed using capillary
`array electrophoresis and energy-transfer fluorescent dye-
`labeled polymerase chain reaction primers. Target alleles
`were amplified by use of primers labeled with one
`fluorescein at the 5' end and another fluorescein at the
`position of the 15th (modified) base to produce fragments
`that fluoresce in the green (/max = 525 nm). Unknown
`alleles were electrophoretically separated together with a
`standard ladder made up of alleles having 6, 7, 8, and 9
`four-base pair repeats, each of which was amplified with
`an energy-transfer primer having a donor fluorescein at
`the 5' end and a rhodamine acceptor at the position of
`the 7th (modified) base to produce standard fragments
`fluorescing in the red (>590 nm). Separations were
`performed on arrays of hollow fused-silica capillaries filled
`with a replaceable sieving matrix consisting of 0.8%
`hydroxyethyl cellulose plus 1 piM 9-aminoacridine to
`enhance the resolution. The labeled DNA fragments were
`excited at 488 nm, and the fluorescence was detected with
`a two-color confocal fluorescence scanner.
`Separations
`are complete in less than 20 min and allow sizing with
`an average absolute error or accuracy of less than 0.4 base
`pair and an average standard deviation of ~0.5 base pair
`with no correction for mobility shift and cross-talk between
`the fluorescence channels. This work establishes the
`feasibility of high-speed, high-throughput STR typing of
`double-stranded DNA fragments using capillary array
`electrophoresis.
`
`DNA sequences containing di-, tri-, tetra-, and pentanucleotide
`repeats are often genetically polymorphic.* 1 2"3 Over 2000 of these
`short tandem repeat (STR) polymorphisms have been mapped
`on the human genome,4 and it is estimated that thousands more
`remain to be discovered.5 Because of the abundance of this type
`of polymorphism and the relative ease of STR detection following
`
`f Department of Chemistry.
`* Performed as a part of the Ph.D. research of Y.W. in the Graduate Group in
`Biophysics.
`§ Forensic Science Group.
`(1) Tautz, D. Nucleic Acids Res. 1989, 17, 6463-6471.
`(2) Edwards, A; Civitello, A; Hammond, H. A; Caskey, C. T. Am. J. Hum. Genet.
`1991, 49, 746-756.
`(3) Charlesworth, B.; Sniegowski, P.; Stephan, W. Nature 1994, 371, 215-
`220.
`(4) Gyapay, G.; Morissette, J.; Vignal, A; Dib, C.; Fizames, C.; Millasseau, P.;
`Marc, S.; Bemardi, G.; Lathrop, M.; Weissenbach, J. Nat. Genet. 1994, 7,
`246-339.
`
`amplification by the polymerase chain reaction (PCR), STRs have
`found widespread use as markers in gene mapping studies4 and
`are emerging as potential markers for use in testing for paternity
`and personal identity.6 *"9
`The analysis of STR markers in gene mapping and in the
`development of population polymorphism data banks is typically
`performed by slab gel electrophoresis. As the number of inter-
`rogated STR markers increases and as one desires to have a more
`complete polymorphism data base, the speed, throughput, and
`sample handling associated with slab gels becomes limiting. One
`approach for increasing the speed and throughput of
`these
`electrophoretic separations is through the use of capillary array
`the speed of
`In this technique,
`electrophoresis (CAE).10"12
`electrophoretic separations is increased ~10-fold through the use
`of high electric fields, and the throughput is increased by using
`large numbers of capillaries in an array. High sensitivity detection
`of fluorescently labeled DNA fragments is achieved through the
`use of a laser-excited, confocal-fluorescence scanner.13 An ad-
`ditional advantage of CAE is that samples are easily loaded in
`parallel through standard electrokinetic injection techniques. Thus
`far, CAE has been used for DNA sequencing1011 and for DNA
`fragment sizing,1415 but it has not been used to analyze STRs.
`In this paper, we explore the development of methods for rapid
`to determine
`STR sizing on capillary arrays. One goal was
`whether accurate sizing can be performed by separating ds-DNA
`fragments on capillaries containing easily replaceable nondena-
`If suc-
`turing hydroxyethyl cellulose (HEC) sieving solutions.
`cessful, this would dramatically increase the speed of separations
`and simplify the separation procedures by eliminating the need
`for denaturing columns. We show here that, by using the
`appropriate separation buffer and labeling methods, sizing of
`
`(5) Beckmann, J. S.; Weber, J. L. Genomics 1992, 12, 627-631.
`(6) Hammond, H. A; Jin, L.; Zhong, Y.; Caskey, C. T.; Chakraborty, R Am. J.
`Hum. Genet. 1994, 55, 175-189.
`(7) Kimpton, C. P.; Gill, P.; Walton, A; Urquhart, A; Millican, E. S.; Adams, M.
`PCR Methods Appl. 1993, 3, 13-22.
`(8) Fregeau, C. J.; Foumey, R M. BioTechniques 1993, 15, 100-119.
`(9) Alford, R L.; Hammond, H. A; Coto, I.; Caskey, C. T. Am. J. Hum. Genet.
`1994, 55, 190-195.
`(10) Huang, X. C.; Quesada, M. A; Mathies, R A Anal. Chem. 1992, 64, 967-
`972.
`(11) Huang, X. C.; Quesada, M. A; Mathies, R A Anal. Chem. 1992, 64,2149-
`2154.
`(12) Mathies, R A; Huang, X. C. Nature (London) 1992, 359, 167-169.
`(13) Quesada, M. A; Rye, H. S.; Gingrich, J. C.; Glazer, A N.; Mathies, R A
`BioTechniques 1991, 10, 616—625.
`(14) Clark, S. M.; Mathies, R A Anal. Biochem. 1993, 215, 163-170.
`(15) Zhu, H.; Clark, S. M.; Benson, S. C.; Rye, H. S.; Glazer, A N.; Mathies, R
`A Anal. Chem. 1994, 66, 1941-1948.
`
`0003-2700/95/0367-1197$9.00/0 © 1995 American Chemical Society
`
`Analytical Chemistry, Vot. 67, No. 7, April 1, 1995 1197
`
`Downloaded via UNIV OF VIRGINIA on November 6, 2018 at 17:48:19 (UTC).
`
`See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
`
`Agilent Exhibit 1272
`Page 1 of 7
`
`

`

`alleles of the human tyrosine hydroxylase locus THOl differing
`by a single base pair can be achieved. A second goal was to
`develop methods for two-color labeling and detection so that a
`standard STR ladder could be electrophoresed along with the
`unknown amplified alleles. We have accomplished this by
`amplifying the standard THOl allelic ladder and the unknown with
`the energy-transfer (ET) dye-labeled primers recently developed
`by Ju and co-workers.16 ET primers are valuable because they
`absorb strongly at a common laser wavelength while also exhibit-
`ing intense and nonoverlapping emissions. The THOl allelic
`sizing experiments presented here validate the use of energy-
`transfer dye-labeled primers together with capillary array elec-
`trophoresis for rapid and high-throughput STR sizing.
`
`EXPERIMENTAL SECTION
`Instrumentation. Capillary array electrophoresis separations
`were detected with the laser-excited, confocal-fluorescence scan-
`ner as previously described by Huang et al.10,11 Briefly, excitation
`light at 488 nm from an argon ion laser is reflected by a long-
`pass dichroic beam splitter, passed through a 32 x, NA 0.4
`microscope objective, and brought to a 10 wm diameter focus
`within the 75 pm i.d. capillaries in the capillary array. The
`fluorescence is collected by the objective, passed back through
`the first beam splitter to a second dichroic beam splitter that
`separates the red (A > 565 nm) and green (A < 565 nm) detection
`channels. The emission is then focused on 400 urn diameter
`confocal pinholes, spectrally filtered by a 590-nm long-pass filter
`(red channel) or a 20-nm band-pass filter centered at 520 nm
`followed by photomultiplier detection. The
`(green channel),
`output is preamplified, filtered, digitized, and then stored in an
`IBM PS/2 computer. A computer-controlled stage is used to
`translate the capillary array past the optical system at 20 mm/s.
`The fluorescence is sampled unidirectionally at 1500 Hz/channel.
`The scanner
`construction and operation have recently been
`described in detail.17 Postacquisition image processing was
`performed with the programs IPLab, KaleidaGraph, and Canvas.
`fused-
`Capillary Electrophoresis. Polyacrylamide-coated,
`silica capillaries were prepared using a modification of
`the
`procedure described by Hjerten.18 A 2-3 mm wide detection
`window was produced by burning off the polyimide coating with
`a hot wire followed by cleaning the external surface with ethanol.
`The detection window was placed 25 cm from the injection ends
`of the 75 /rm i-d., 350 /rm o.d., 50 cm long fused-silica capillaries
`(Polymicro Technologies, Phoenix, AZ). The inner walls of the
`capillaries were incubated with 1 N NaOH for 30 min at room
`followed by rinsing with deionized water. The
`temperature,
`capillaries were then treated overnight at room temperature with
`•/-methacryloxypropyltrimethoxysilane (1:250 dilution with H2O
`adjusted to pH 3.5 with acetic acid) to derivatize the walls for
`polyacrylamide binding. Freshly made 4% T acrylamide solution
`in V2X TBE buffer (45 mM tris, 45 mM boric acid, 1 mM EDTA,
`pH 8.3) was filtered with a 0.2-^m syringe filter and degassed
`under vacuum for 30 min. NJVJf /V’-Tetramethylethylenediamine
`CTEMED; 1 piL) and 10 uL of 10% ammonium persulfate (APS)
`solution were added to 1 mL of gel solution. The solution was
`immediately forced into the capillary with a 100-mL syringe. After
`
`(16) Ju, J.; Ruan, C.; Fuller, C. W.; Glazer, A. N.; Mathies, R. A Proc. Natl. Acad.
`Sci. U.S.A., in press.
`(17) Mathies, R. A; Scherer, J. R.; Quesada, M. A; Rye, H. S.; Glazer, A N. Rev.
`Sci. Instrum. 1994, 65, 807—812.
`(18) Hjerten, S./. Chromatogr. 1985, 347, 191-198.
`
`1198 Analytical Chemistry, Vol. 67, No. 7, April 1, 1995
`
`30 min, the polyacrylamide solution was flushed out with deionized
`water and capillaries were filled with buffer consisting of hydroxy-
`ethyl cellulose (HEC; M?. = 438 000, Aqualon Co. Hopewell, VA)
`dissolved in 'Ax TBE. The separation buffer was prepared by
`adding 0.8 g of HEC to 100 mL of 'Ax TBE and dissolved by
`stirring overnight at room temperature.14 The HEC buffer was
`degassed under vacuum for 30 min, centrifuged for 20 min on a
`tabletop centrifuge, and drawn into a lOO/tL syringe, and 3 uL
`was injected into each capillary. Capillaries were prerun at 80
`V/cm for 5 min before each experiment. Diluted and deionized
`PCR samples were injected by inserting the capillary in a 5-«L
`sample volume held in an Eppendorf tube followed by electroki-
`netic injection (80 V/cm for 3 s). After injection, the sample tubes
`were replaced with tubes containing 0.8% HEC plus V2x TBE
`buffer. Electrophoresis was performed at 80 V/cm using five-
`capillary arrays held at ambient temperature (22 °C). The low
`(80 V/cm) electrophoresis voltage was used to avoid undersam-
`pling of the bands with our current detection system, which is
`limited to 1-Hz scan rates. When the experiments were complete,
`capillaries were flushed with water and then with methanol,
`followed by drying. Our coated capillaries could be refilled 20-
`25 times before the quality of the separations deteriorated.
`Methods for the further extension of the lifetime of capillary
`columns have been described.19
`PCR Amplification of THOl Loci. DNA was isolated from
`blood by using standard methods.20 The human tyrosine hy-
`location llpl5.5,
`droxylase locus HUMTHOl, chromosomal
`contains a polymorphic four-base STR sequence (AATG) in intron
`l.21 PCR amplification of this polymorphic region produces allelic
`fragments designated 5-11, according to the number of AATG
`repeats; an additional allele designated 9.3 differs from allele 10
`by a single base deletion. The primer sequences used for PCR
`are 5'-ATTCAAAGGGTATCTGGGCTCTGG-3' (THOl-A) and 5'-
`GTGGGCTGAAAAGCTCCCGATTAT-3' (THOl-B).2 PCR ampli-
`fications were performed in 50-mL volumes by using 10 ng of
`genomic DNA template, 0.5 pM of each primer, 5 units of Taq
`DNA polymerase, 50 mM KC1,1.5 mM MgCl2,10 mM Tris-HCl
`at pH 8.3, and 200 «M dNTPs (final concentrations indicated).
`The PCR cycle protocol using a Perkin-Elmer Cetus Model 480
`(1) melting at 95 °C for 5 min; (2) 30 cycles of 95
`was as follows:
`°C for 1 min, 58 °C for 1 min, and 72 °C for 1 min; (3) 72 °C for
`7 min to complete extention. The PCR sample was then dialyzed
`for 30 min by pipeting 8-10 uL onto a 0.10-^m VCWP membrane
`filter (Millipore, Bedford, MA), which was floated on deionized
`water in a beaker held at 4 °C. Dialysis was used to remove
`salts
`which can interfere with sample injection. Following dialysis, the
`samples were diluted with deionized water 100-1000 times
`(depending on product concentration) before electrokinetic injec-
`tion. The amplifications with fluorescently labeled primers were
`Initially, four sets of fluores-
`also performed as described above.
`cent primers were used for PCR amplification and the mobility
`shifts of the products were evaluated with capillary electrophore-
`sis. For these mobility shift experiments, 1-2 ng of unlabeled
`reamplified by 20 PCR cycles using the
`PCR product was
`
`(19) Schmalzing, D.; Piggee, C. A; Foret, F.; Carrilho, E.; Karger, B. L. J.
`Chromatogr. 1993, 652, 149-159.
`(20) Sambrook, J.; Fritsch, E. F.; Maniatis, T. Molecular cloning: a laboratory
`manual, 2nd ed.; Cold Spring Harbor Laboratory: Cold Spring Harbor, NY,
`1989.
`(21) Puers, C.; Hammond, H. A; Jin, L.; Caskey, C. T.; Schumm, J. W. Am. J.
`Hum. Genet. 1993, 53, 953-958.
`
`Agilent Exhibit 1272
`Page 2 of 7
`
`

`

`Chart 1. Structures of the Six PCR Primers Used
`for the Amplification of the HUMTH01 Loci"
`F10F
`FAM-5'-ATTCAAAGGGT"ATCTGGGCTCTGG-3'
`
`(CH)2(CO)-NH-(CH2)g-NH-FAM
`
`F14F
`
`FAM-5'-GTGGGCTGAAAAGCTCCCGATTAT-3‘
`
`(CH)2(CO)-NH-(CH2)6-NH-FAM
`
`F2R
`
`F6R
`
`FAM-5-ATj’CAAAGGGTATCTGGGCTCTGG
`(CH)2(CO)-NH-(CH2)6-NH-ROX
`
`-3'
`
`FAM-5-GTGGGCT'GAAAAGCTCCCGATTAT-3'
`<i!;h)2(co)-nh-(ch2)6-nh-rox
`
`TH01 -A
`
`5'-ATTCAAAGGGT ATCTGGGCTCTGG -3'
`
`TH01-B
`5'-GTGGGCTGAAAAGCTCCCGATTAT -3'
`a The fluorescent primers are labeled with a common fluores-
`cein donor (F) at the 5' end and either a second fluorescein or a
`rhodamine (R) acceptor at the indicated locations of a modified
`T in the sequence. The number of nucleotides between the two
`fluorophores is indicated in the primer designation
`
`appropriate fluorescent primers. THOl types for all samples used
`in this study were independently determined by analysis on slab
`gels essentially as described by Puers.21 Standard reference alleles
`were determined by sequence analysis (M. Savill and G. Sensa-
`baugh, unpublished).
`Design and Synthesis of PCR Printers. Chemicals were
`purchased from Applied Biosystems (Foster City, CA). Oligode-
`oxynucleotides were synthesized by the phosphoramidite method
`on an Applied Biosystems 392 DNA synthesizer. The structures
`of two unlabeled (THOl-A and IHOl-B) and four energy-transfer
`dye-labeled PCR primers are presented in Chart 1. The nomen-
`clature and procedures for preparing and purifying energy-transfer
`fluorescent primers have been described by Ju et al.16 The F10F
`and F14F primers required for the studies reported here were
`synthesized with the incorporation of two 5-carboxyfluorescein
`(FAM) fluorophores at the locations indicated in Chart 1. The
`F2R and F6R primers were similarly prepared with the incorpora-
`tion of FAM as a donor and 6-carboxy-X-rhodamine (ROX) as an
`acceptor.
`The energy-transfer dye-labeled primers are advantageous for
`two-color fragment sizing because the 488-nm exciting light is
`optimally absorbed by the FAM chromophore in these primers
`followed by enhanced emission at the FAM wavelength in the
`case of F10F and F14F or very distinctively Stokes-shifted emission
`following energy transfer in the case of F2R and F6R. Absorption
`spectra of the primers were measured on a Perkin-Elmer Lambda
`6 UV-visible spectrophotometer and fluorescence emission
`spectra were taken on a Perkin-Elmer Model MPF 44B spectro-
`fluorimeter. As representative examples, spectra of F14F and the
`energy-transfer dye-labeled primer F6R are presented in Figure
`1. F14F exhibits strong absorption at ~488 nm and intense
`fluorescence emission with a maximum at 525 nm.
`F6R also
`exhibits intense absorption at 488 nm, but because of
`the
`fluorescein-to-rhodamine fluorescence energy transfer, the emis-
`sion maximum is shifted out to ~600 nm. Primers were dissolved
`in 10 mM Tris-HCl, 1 mM EDTA buffer at a final concentration
`of 10 pmol/wL for PCR reactions.
`
`Figure 1. Absorption (—) and fluorescence emission (—) spectra
`of the fluorescently labeled THOl primers F14F and F6R measured
`in 1 x TBE. F14F exhibits strong absorption at ~488 nm and intense
`fluorescence emission with a maximum at 525 nm. F6R also exhibits
`intense absorption at ~488 nm, but the maximum emission is shifted
`out to ~600 nm.
`
`Figure 2. Electropherograms of (A) the TH01 standard ladder
`consisting of the 6, 7, 8, and 9 alleles which have been mixed and
`then coinjected with a OX-174 RF DNA-H/nd I digest. (B) Standard
`allelic ladder spiked with “unknown” allele 7. (C) Standard allelic ladder
`spiked with allele 9. (D) Standard allelic ladder spiked with alleles 7
`and 8. (E) Standard allelic ladder spiked with alleles 9 and 9.3. These
`separations were run with 0.8% HEC, V2X TBE, and 1 fiM thiazole
`orange in the running buffer and detected in the green channel. The
`additional weak peaks appearing at detection times greater than ~17
`min are due to heteroduplex formation.
`
`Analytical Chemistry, Vol. 67, No. 7, April 1, 1995
`
`1199
`
`Agilent Exhibit 1272
`Page 3 of 7
`
`

`

`Time (minutes)
`Figure 3. Comparison of the mobility shift using five different
`methods for attaching energy-transfer-coupled fluorophores to TH01
`target alleles 6 and 9.3. The green line indicates the fluorescence
`intensity in the green channel, and the red line indicates the intensity
`in the red channel. The structures of the primer sets used are
`indicated. Electrophoresis was performed using 0.8% HEC, 'fax TBE,
`and 1
`9-AA in the running buffer.
`
`Time (minutes)
`Figure 4. Electropherograms of standard TH01 allelic ladder
`separations using three different running buffers. (A) 0.8% HEC plus
`'fax TBE. (B) 0.8% HEC plus 'fax TBE and 1 tiM thiazole orange.
`The additional peaks at 15.2 and 17.1 min in this run are due to OX-
`174 RF DNA-H/ndl DNA. (C) 0.8% HEC plus 'fax TBE and 1 pM
`9-AA. The standard ladder was amplified with F6R and detected in
`the red channel.
`
`RESULTS AND DISCUSSION
`One-Color Sizing of THOl Alleles. Figure 2 presents
`capillary array electrophoresis separations of the standard THOl
`ladder that has been labeled on-column with the intercalating
`fluorophore, thiazole orange (TO).22 To form this standard ladder,
`two individual heterozygote samples containing alleles 6 + 9 and
`7 + 8 were amplified and mixed. Trace A presents the electro-
`pherogram of the allelic ladder which was injected along with a
`d>X-174 RF DNA-tfmdl restriction digest (Pharmacia, Pisca-
`taway, NJ) as a control. The four alleles which are 4 bp apart are
`nearly baseline-resolved with a separation time of less than 17
`In trace B, an “unknown” PCR-amplified sample is added to
`min.
`the allelic ladder and coinjected. The increased intensity of allele
`7 identifies the unknown allele. Traces C—E in Figure 2 similarly
`demonstrate the detection of samples containing alleles 9, 7 + 8,
`and 9 + 9.3. This work demonstrates the feasibility of producing
`the THOl ladder for allelic typing with capillary array electro-
`phoresis using nondenaturing HEC separation matrices. The use
`of a standard allelic ladder to perform accurate sizing has
`previously been reported.2'21 High-resolution separation of ds-
`DNA fragments on capillary columns has also been demonstrated
`
`(22) Rye, H. S.; Quesada, M. A; Peck, K.; Mathies, R. A; Glazer, A N. Nucleic
`Acids Res. 1991, 19, 327-333.
`
`1200 Analytical Chemistry, Vol. 67, No. 7, April 1, 1995
`
`«
`
`Figure 5. Resolution of the TH01 ladder as a function of 9-ami-
`noacridine concentration. The resolution of last two adjacent peaks
`(8, 9) in the TH01 allelic ladder was calculated using the equation R
`(T2 - Ti) x 1.18/(fwhhi + fwhh2), where Tand fwhh are the peak
`migration time and full width at half height, respectively. The TH01
`ladder was amplified with F6R as described in the text and separated
`on 0.8% HEC at 80 V/cm.
`
`=
`
`by McCord and co-workers.23'24 However, for one-color labeling
`of the allelic ladder and unknown to work reliably and precisely,
`it is necessary to have good control over
`the concentration of the
`
`Agilent Exhibit 1272
`Page 4 of 7
`
`

`

`67e9
`
`Red Channel
`4
`2
`3
`
`1
`
`5
`
`1
`
`Green Channel
`4
`2
`3
`
`5
`
`15 min
`
`17 min
`
`19 min
`
`Images of the fluorescence from a five-capillary array separation of THOI alleles. The left image presents the fluorescence signal
`Figure 6.
`as a function of time detected in the red (>590 nm) channel, and the right image presents the fluorescence signal from the green (Amax = 525
`nm) channel. The standard THOI allelic ladder (6 + 7 + 8 + 9) was amplified with the red-emitting ET primer F6R and detected in the red
`channel; unknown alleles were amplified with the green-emitting primer F14F and detected in the green channel. These images have been
`adjusted for the 1 -2% capillary-to-capillary variance in mobility by shifting the time axes so that the allelic ladder is detected at the same time
`9-AA in the running buffer at 80 V/cm.
`in all capillaries. These separations were performed with 0.8% HEC and 1
`
`DNA in the unknown samples. Alternatively, one could use a
`nonallelic ladder as the standard but there could be some loss of
`accuracy in the resulting interpolation.
`Evaluation of Energy-Transfer Primer Labeling. For
`routine sizing experiments, it is desirable to perform two-color
`detection where the allelic standards are amplified with one
`fluorescent primer and the unknowns are amplified with a second
`fluorescent primer having a distinctive emission. Ju and co-
`workers16 recently reported the synthesis and use of a new class
`of fluorescent primers labeled with pairs of dye molecules that
`are coupled by fluorescence energy transfer. These ET primers
`have the advantage of providing strong absorption at a common
`laser excitation wavelength (488 nm). Following the fluorescence
`energy transfer, the ET primers emit at a Stokes-shifted wave-
`length determined by the properties of the acceptor. Thus, the
`fluorescence emission of the ET primers is very intense, and the
`emission spectra of the different ET dye-labeled primers are
`distinctively Stokes-shifted. These primers have been shown to
`
`(23) McCord, B. R.: McClure, D. L; Jung, J. M. /. Chmmatogr. A 1993, 652,
`75-82.
`(24) McCord, B. R.; Jung, J. M.; Holleran. E. A ]. I.iq. Chmmatogr. 1993, 16,
`1963-1981.
`
`provide 2-6 times the signal strength compared to conventional
`single dye-labeled fluorescent primers in DNA sequencing ap-
`plications.16 Furthermore, the mobility shift of DNA fragments
`generated with ET primers depends on the spacing between the
`dyes. We therefore performed experiments to evaluate the
`mobility shift of the amplified fragments for all combinations of
`singly and doubly labeled targets.
`In trace A of Figure 3, both strands of the amplified 6 and 9.3
`In one case, the F10F primer is
`targets have been labeled.
`extended to form the (+) strand and the F14F primer is extended
`to form the (-) strand producing the green-emitting fragments.
`In the second case, the ET dye-labeled primer F2R is extended
`to produce the (+) strand and F6R is used to extend the (-)
`strand producing the red-emitting fragments. These fragments
`were mixed and electrophoresed on HEC-filled capillaries in the
`presence of 1 /<M 9-aminoacridine (9AA, see below). The mobility
`shift between the green fragments and the red fragments in trace
`found to be ~2 bp. We therefore decided to evaluate
`A was
`amplifying with just one fluorescent primer per ds-DNA fragment
`to see if a particular combination of labels would reduce the
`mobility shift. Trace B in Figure 3 shows that amplifying the (-)
`
`Analytical Chemistry, Vol. 67, No. 7, April 1, 1995
`
`1201
`
`Agilent Exhibit 1272
`Page 5 of 7
`
`

`

`strand with the F14F primer or with the F6R primer generates
`fragments having almost no mobility shift (<0.3 bp) between the
`green- and the red-labeled fragments. The two other labeling
`methods shown in traces C and D resulted in larger mobility shifts.
`However, amplifying the (-) strand with F6R and the (+) strand
`with F10F also produced fragments having almost no mobility
`shift (trace E). We decided to perform the subsequent experi-
`ments using the labeling method illustrated in trace B because
`these fragments are labeled on the same strand and can thus also
`be sized under denaturing conditions if necessary.
`Resolution Enhancement with 9-Aminoacridine. To achieve
`satisfactory resolution of the THOl allelic ladder, we found it
`necessary to include an intercalating dye in the running buffer.
`Trace A of Figure 4 shows that poor resolution is obtained when
`the allelic ladder (amplified with the F6R primer) is run in 0.8%
`HEC alone. Trace B shows the separation of the same ladder in
`0.8% HEC plus 1 «M of the intercalating dye thiazole orange. The
`resolution of the separation in the presence of TO is dramatically
`enhanced. Electrophoresis in the presence of the intercalator
`ethidium bromide has also been shown to improve the electro-
`phoretic resolution of ds-DNA25-26 Unfortunately, TO contributes
`in the green channel of our
`to the signal
`two-color detection
`system, rendering it unsuitable for use in the desired two-color
`It is thus necessary to use a nonfluorescent
`labeling scheme.
`In electrophoretic separa-
`intercalator to improve the resolution.
`tions of preformed dimeric dye-DNA complexes, Zhu et al.15
`observed that the addition of the nonfluorescent dye 9-aminoacri-
`dine (9-AA) can be used to dramatically improve ds-DNA separa-
`tions much like TO and ethidium. We therefore decided to
`evaluate the effect of 9-AA Trace C of Figure 4 presents a
`separation of the THOl allelic ladder when the column is filled
`with 0.8% HEC and 1 pM 9-AA This separation is as good as
`that obtained in the presence of TO. The dependence of the
`resolution of the THOl ladder separation on 9-AA concentration
`is presented in Figure 5. The resolution improves significantly
`up to 1 pM 9-AA and is only slighdy better at 5 «M. 9-AA
`concentrations above 50 /<M were found to quench the fluores-
`cence.
`Two-Color THOl Sizing with Capillary Array Electro-
`phoresis. Figure 6 presents the results of a typical THOl sizing
`experiment performed using two-color capillary array electro-
`phoresis, The standard allelic ladder was amplified with F6R and
`detected in the red channel, while the unknown alleles were
`amplified with F14F and detected in the green channel. The signal
`detected as a function of time is presented as two images:
`the
`left image is the signal in the red channel and the right image is
`that detected in the green channel. The alleles appear ~17 min
`after injection. The signal in the red channel is predominantly
`from the red-labeled standard ladder, and the expected four band
`patterns are seen in all capillaries. The unknown, amplified with
`the green-emitting primer,
`is detected in the green channel.
`Allelic bands are identified as the intense green bands coinciding
`in mobility with allelic ladder bands in the red channel. To
`illustrate, the right image of Figure 6 reveals intense green bands
`corresponding to allele 7 in lane 1, alleles 9 + 9.3 in lane 2, alleles
`6 + 9.3 in lane 3, alleles 7 -I- 8 in lane 4, and allele 9 in lane 5.
`Figure 7 presents electropherograms derived from the image in
`
`(25) Schwartz. H. E.: Ulfelder, K.; Sunzeri, F. J.; Busch, M. P.; Brownlee, R G.
`/. Chromatogr. 1991, 559, 267-283.
`(26) Guttman, A.: Cooke, N. Anal. Otem. 1991, 63, 2038-2042.
`
`1202 Analytical Chemistry, Vol. 67, No. 7, April 1, 1995
`
`Time (minutes)
`Figure 7. Electropherograms of the THOl fragment sizing separa-
`is from the unknown
`tions presented in Figure 6. The green signal
`alleles, and the red signal is from the standard TH01 ladder. Traces
`A-E correspond to lanes 1 -5 in Figure 6.
`
`Figure 6. Very clear discrimination is observed between the
`green-labeled fragments and the red-labeled fragments with almost
`In some cases,
`no cross-talk between the green and red channels.
`the amplification products of heterozygote samples contain het-
`eroduplex bands that migrate behind the allelic ladder. Additional
`weak “noise” bands may represent nontemplated base addition7
`known to occur with PCR amplification of THOl (M. Savill and
`G. Sensabaugh, unpublished data).
`The accuracy and precision of allelic sizing using CAE was
`tested by performing multiple runs on 11 different samples. These
`results are summarized in Table 1. Since a linear relationship
`exists between molecular weight and migration time in the region
`of interest, the allelic ladder was used with a linear regression
`analysis to size the unknown fragments. The calculated sizes of
`unknown alleles are compared to true sizes based on sequence
`analysis and verified by denaturing polyacrylamide gel electro-
`phoresis. The average absolute difference of the determined allele
`70% of the
`sizes from the true allele sizes was 0.41, and over
`determined values were within 0.5 bp of the true value. The
`reproducibility is excellent (relative standard deviation less than
`no ambiguity in allele
`0.4% for each allele) and there was
`assignments. Two alleles, 9.3 and 10, which differ by a single
`base pair deletion cannot be electrophoretically resolved when
`paired but can be correctly assigned when in combination with
`It should, however, be possible to separate these
`any other allele.
`two fragments on columns containing higher concentrations of
`HEC 0. Bashkin and R. Johnston, personal communication).
`
`Agilent Exhibit 1272
`Page 6 of 7
`
`

`

`Table 1. Statistical Analysis of THOI Fragment Sizing*
`no. of detns
`allele
`mean size*
`length (bp)
`SDC (%)
`183
`187
`191
`195
`198
`199
`
`183.1
`187.0
`191.2
`195.4
`198.3
`199.0
`
`0.61 (0.33)
`0.41 (0.22)
`0.69 (0.36)
`0.60 (0.31)
`0.52 (0.26)
`0.31 (0.15)
`
`6
`7
`8
`9
`9.3
`10
`
`6
`23
`15
`11
`8
`6
`
`“ Eleven different amplified samples (7, 8), (6, 9.3), (9, 9.3), (6, 9)
`(7, 8), (8, 9), (7, 9.3), (7, 10), (6, 9.3), (7). and (9) were run 8, 2, 4, 3,
`5, 2, 1, 6, 1, 3, and 2 times, respectively.b Mean PCR product size as
`determined by linear regression using the allelic ladder as the sizing
`in terms of base pairs for the
`standard.c Standard deviation (SD)
`indicated number of determinations. The percent relative SD is given
`in parentheses.
`
`CONCLUSIONS
`The purpose of this work was to develop methods for high-
`speed, high-throughput sizing of short tandem repeats using
`capillary array electrophoresis. This can be broken down into
`the steps of (1) developing methods for achieving single-base (or
`near single base) resolution of ds-DNA fragments using replace-
`able separation matrices and (2) developing methods for multiplex
`labeling to achieve on-column sizing of unknown STRs in the
`presence of an allelic standard. We demonstrate here that these
`objectives can be achieved using ET primers in combination with
`CAE on HEC capillary supports containing 9-AA The advantages
`
`(27) Bashkin, J. S.; Roach, D.; Rosengaus, E.; Barker, D. L. Human Genome
`Program, Contractor-Grantee Workshop III, U.S. Department of Energy
`1993; p 116.
`
`of the method presented here are that ET primers provide higher
`sensitivity and better color discrimination than that possible with
`conventional single dye-labeled fluorescent PCR primers. Second,
`the HEC solutions used as the sieving media provide high speed
`(<20 min) separations on columns that can be easily refilled and
`rerun. Finally, the CAE format provides the ability to run 50 or
`more separations in parallel.27 The amplification of multiple STR
`targets with different fragment lengths in the 100-500-bp size
`range plus the use of up to four different ET primers16 should
`provide an additional increase in the throughput of STR typing.
`ACKNOWLEDGMENT
`We thank the members of the Berkeley High-Sensitivity DNA
`Analysis Group for their support, Alexander N. Glazer and Steven
`M. Clark for valuable discussions, and Alan Taur for expert
`preparation of the figures. This research was supported by a grant
`from the Director, Office of Energy Research, Office of Health
`and Environmental Research of the U.S. Department of Energy
`under Contract DE-FG-91ER61125 to RAM. and by a grant from
`the National Institute of Justice (93-U-EX-0010) to G.F.S. JJ. was
`supported by a Human Genome Distinguished Postdoctoral
`Fellowship sponsored by the U.S. Department of Energy, Office
`of Health and Environment Research, and administrated by the
`Oak Ridge Institute for Science and Educatio