Sponsored by

This booklet is brought to you by the AAAS/Science Business Office

Not using ExoSAP-IT?Your loss.

Untreated (–) and ExoSAP-IT treated (+) PCR products were analyzed by gel

electrophoresis. A variety of PCR products of different lengths may be treated with

ExoSAP-IT, with no sample loss.

125 bp 455 bp 1.55 kb 4.6 kb HES-1 numb NRAGE numb M – + – + – + – +

No Sample Loss with ExoSAP-IT®

For more information on ExoSAP-IT®, PN 78200 call 800.321.9322 or visit www.usbweb.com/exosapitIn Europe: +49(0)76 33-933 40 0 or visit www.usbweb.de/exosapit

Stop losing time and samples. ExoSAP-IT® offers quick, one step PCR clean-up with 100% recovery of both short and long PCR products.

ExoSAP-IT• Utilizes Exonuclease I to degrade leftover primers & Shrimp Alkaline

Phosphatase to degrade unused dNTPs• One easy step– just add ExoSAP-IT, incubate, & heat inactivate• Directly use treated sample for sequencing, SNP analysis, etc.

Benefits• Rapid PCR clean-up –15 min to treat, 15 min to inactivate• No sample loss – PCR products are treated & then used directly in the

next application• No tedious columns involved – more free time• Solution-based clean-up – completely scaleable for manual or

automated use

ScienceBook_Ads.indd 2 10/21/07 8:57:57 AM

Contents

© 2007 by The American Association for the Advancement of Science. All rights reserved.30 NOVEMBER 2007

Introductions 2 Setting Off a Chain Reaction Sean Sanders

3 From Hot Springs to Hot Start Michele Paris

Articles 4 Enzymatic Amplification of ß-Globin Genomic Sequences and

Restriction Site Analysis for Diagnosis of Sickle Cell Anemia Randall K. Saiki, Stephen Scharf, Fred Faloona, Kary B. Mullis, Glenn T. Horn, Henry A. Erlich, Norman Arnheim Science20 December 1985 230: 1350–1354

12 The Molecule of the Year Ruth Levy Guyer and Daniel E. Koshland, Jr.

Science22 December 1989 246: 1543–1546

14 A Technique Whose Time Has Come Nigel J.Walker

Science19 April 2002 296: 557–559

18 Accurate Multiplex Polony Sequencing of an Evolved Bacterial Genome Jay Shendure, Gregory J. Porreca, Nikos B. Reppas, Xiaoxia Lin, John P. McCutcheon, Abraham M. Rosenbaum, Michael D. Wang, Kun Zhang, Robi D. Mitra, George M. Church

Science 9 September 2005 309: 1728–1732

25 Microfluidic Digital PCR Enables Multigene Analysis of Individual Environmental Bacteria Elizabeth A. Ottesen, Jong Wook Hong, Stephen R. Quake, Jared R. Leadbetter

Science 1 December 2006 314: 1464–1467

Technical Notes 31 HotStart-ITTM: A Novel Hot Start PCR Method Based on Primer Sequestration USB Corporation

About the Cover:A brief history of PCR, taking us from the first black and white radiographs of amplified DNA to Neanderthal remains from which DNA has been amplified in preparation for sequencing and comparison with humans, all overlaid on a pseudocolored PCR gel.

Copy Editor: Robert BuckDesign and Layout: Amy Hardcastle

Not using ExoSAP-IT?Your loss.

Untreated (–) and ExoSAP-IT treated (+) PCR products were analyzed by gel

electrophoresis. A variety of PCR products of different lengths may be treated with

ExoSAP-IT, with no sample loss.

125 bp 455 bp 1.55 kb 4.6 kb HES-1 numb NRAGE numb M – + – + – + – +

No Sample Loss with ExoSAP-IT®

For more information on ExoSAP-IT®, PN 78200 call 800.321.9322 or visit www.usbweb.com/exosapitIn Europe: +49(0)76 33-933 40 0 or visit www.usbweb.de/exosapit

Stop losing time and samples. ExoSAP-IT® offers quick, one step PCR clean-up with 100% recovery of both short and long PCR products.

ExoSAP-IT• Utilizes Exonuclease I to degrade leftover primers & Shrimp Alkaline

Phosphatase to degrade unused dNTPs• One easy step– just add ExoSAP-IT, incubate, & heat inactivate• Directly use treated sample for sequencing, SNP analysis, etc.

Benefits• Rapid PCR clean-up –15 min to treat, 15 min to inactivate• No sample loss – PCR products are treated & then used directly in the

next application• No tedious columns involved – more free time• Solution-based clean-up – completely scaleable for manual or

automated use

ScienceBook_Ads.indd 2 10/21/07 8:57:57 AM

Setting Off a Chain ReactionIt can be an interesting (and sometimes instructive) thought experiment to speculate where we would be without certain technologies: the airplane, the telephone, the internet. Clearly a concrete conclusion is difficult to obtain, and this is particularly true for the question: Where would scientific research be without the polymerase chain reaction?

Few technologies in the life sciences can claim to have been as pivotal as the polymerase chain reaction (PCR). Some might point to DNA cycle sequencing or the cloning of genetic material as groundbreaking techniques. And they are. However, neither would be possible—or at least practical—without an ini-tial amplification step provided by PCR.

PCR was developed in 1983 by Kary Mullis (for which he won the 1993 Nobel Prize in Chemistry) while working at Cetus Corporation. The original reactions were performed in what would now be regarded as a rather primitive fash-ion, as described in the first paper on the application of PCR in the opening article of this booklet. Reaction tubes containing the necessary components were manually cycled between water baths or heating blocks to achieve high temperature denaturation of DNA, followed by cooler primer binding and po-lymerization conditions. The enzyme that performed the amplification work, E. coli DNA polymerase, was heat sensitive, necessitating the addition of fresh enzyme after each denaturation step.

The advent of the PCR thermal cycler and its release in 1987 automated the tedious task of manual temperature cycling, but it was the novel application of the heat stable Taq polymerase by scientists at Cetus that really allowed PCR to flourish, resulting in Science naming PCR as the “major scientific develop-ment” of 1989. Subsequent discoveries of additional thermostable polymer-ases and the mutation of other enzymes in the polymerase family has enabled the development of high fidelity, long read polymerases, all of which have added to the robustness of the PCR technique.

PCR technology itself has been mutated many times, resulting in the cre-ation of an array of specialized applications. But probably the most widely used and biggest innovation has been quantitative PCR (Q-PCR, also called real-time PCR). The development of this methodology was both an intellectual and technological leap which freed researchers from having to wait until reac-tions were complete before they could perform gel separation and analysis of their samples. It instead effectively gave them a window into the reaction tube, allowing real-time tracking and quantitation of the PCR and inspiring a whole new generation of enabling techniques, technologies, and products. The re-view from Nigel Walker in this booklet nicely summarizes the powerful impact of real-time PCR on molecular biology research.

PCR is currently still used for a vast array of applications, from basic am-plification of DNA fragments for cloning, to amplifying ancient DNA prior to sequencing. It is constantly being modified, improved, and applied in new and interesting ways, as the closing papers of this collection demonstrate. So, for the moment, scientific research is clearly still very much dependent on this technology, a situation that will sustain it as a cornerstone of biomedical re-search well into the 21st century.

Sean Sanders, Ph.D.Commercial Editor, Science

�

3

From Hot Springs to Hot StartReflecting on the magnitude of the PCR application gives rise to a unique perspective here at USB Corporation. USB, formerly known as United States Biochemical, Inc., is probably most well-known in the earlier days of molecu-lar biology for its Sequenase DNA Polymerase. The company may not be as widely known for its role in the early days of DNA amplification. Working with Cetus Corporation in the late 1980s, USB purified the first commercial prepara-tions of Thermus aquaticus (Taq) DNA polymerase, the first enzyme identified to withstand the high temperatures required for PCR.

Thermus aquaticus, from the hot springs of Yellowstone National Park, was first isolated and characterized by Thomas Brock and Hudson Freeze in 1969 at Indiana University. The DNA polymerase of the organism was then character-ized in 1976 at the University of Cincinnati by Alice Chien, David Edgar, and John Trela. Ten years later, scientists from Cetus Corporation made creative use of the enzyme’s thermostability in Kary Mullis’s Nobel Prize winning inven-tion for DNA amplification—the polymerase chain reaction (PCR). USB assisted Cetus with scale-up of the enzyme so it could be produced in quantities large enough to sustain the market. The first preps were native, direct from the deep orange cell pastes provided to us by Cetus. The enzyme was soon cloned and recombinant Taq was more easily purified. At the time, USB had a coexclusive agreement with Perkin-Elmer Cetus Instruments to market the very first PCR enzyme, AmpliTaq®, and its associated kit.

Today USB offers tried and true Taq, FideliTaq™ for long and accurate PCR, RubyTaq™ for direct-load, and HotStart-IT™ for end-point PCR, qPCR, and RT-PCR. The HotStart-IT method is unique because it is based on primer se-questration, not a modification to the Taq enzyme. HotStart-IT uses a DNA binding protein that binds the primers prior to thermocycling and releases them after the first cycling step. The result is higher specificity and sensitivity providing another advancement in our history with PCR.

Do most scientists care about this history? We think so. As scientists inter-ested in biology, we have a need to understand history and apply it to the discoveries of today. After all, it is interesting to know where we have been so we can see where we are going.

Years ago, it may have been difficult to predict the future of PCR. It has become an extremely powerful tool that is deeply entrenched. The technique continues to be key in core applications such as sequencing, cloning, molecu-lar diagnostics, and protein expression. With the progress made in real time PCR, we can now better assess gene expression in a very accurate way. The applications will continue to develop with next generation technologies as in-dicated by the recent articles included in this booklet.

Michele ParisDirector of Marketing, USB Corporation

�

Recentadvances in recombinantDNAtechnology have made possible themolecular analysis and prenatal di-agnosis of several human genetic diseases.Fetal DNA obtained by aminocentesis orchorionic villus sampling can be analyzedby restriction enzyme digestion, with sub-sequent electrophoresis, Southern transfer,andspecifichybridizationtoclonedgeneoroligonucleotide probes. With polymorphicDNAmarkerslinkedgeneticallytoaspecificdisease locus, segregation analysis must becarriedoutwith restriction fragment lengthpolymorphisms (RFLP’s) found to be in-formative by examining DNA from familymembers(1, 2).

Many of the hemoglobinopathies, how-ever, canbedetectedby moredirectmeth-ods in which analysis of the fetus alone issufficient for diagnosis. For example, thediagnosis of hydrops fetalis (homozygousa-thalassemia) can be made by document-ing the absence of any a-globin genes byhybridizationwithana-globinprobe(3-5).Homozygosity for certainß-thalassemiaal-

lelescanbedeterminedinSoutherntransferexperimentsbyusingoligonucleotideprobesthatformstableduplexeswiththenormalß-globingenesequencebutformunstablehy-bridswithspecificmutants(6, 7).

SicklecellanemiacanalsobediagnosedbydirectanalysisoffetalDNA.Thisdiseaseresultsfromhomozygosityofthesickle-cellallele(ßS)attheß-globingenelocus.TheSallelediffersfromthewild-typeallele(ßA)bysubstitutionofanAinthewild-typetoaTatthesecondpositionofthesixthcodonoftheßchaingene,resultinginthereplacementofaglutamicacidbyavalineintheexpressedprotein.Fortheprenataldiagnosisofsicklecellanemia,DNAobtainedbyamniocentesisor chorionic villus sampling can be treatedwitharestrictionendonuclease(forexample,DdeIandMstII)thatrecognizesasequencealteredbytheßSmutation(8-11).Thisgener-atesßA-andßS-specificrestrictionfragmentsthat can be re-solved by Southern transferandhybridizationwithaß-globinprobe.

Wehavedevelopedaprocedureforthede-tectionofthesicklecellmutationthatisveryrapidandisatleasttwoordersofmagnitudemoresensitivethanstandardSouthernblot-ting.There are two special features to thisprotocol.Thefirst isamethodforamplify-ing specific ß-globin DNA sequences withtheuseofoligonucleotideprimersandDNA

Enzymatic Amplification of ß-Globin Genomic Sequences and Restriction Site Analysis for Diagnosis of Sickle Cell AnemiaRandall K. Saiki, Stephen Scharf, Fred Faloona, Kary B. Mullis, Glenn T. Horn, Henry A. Erlich, Norman Arnheim

Two new methods were used to establish a rapid and highly sensitive prenatal diagnostic test for sickle cell anemia. The first involves the primer-mediated enzymatic amplification of specific ß-globin target sequences in genomic DNA, resulting in the exponential increase (220,000 times) of target DNA copies. In the second technique, the presence of the ßA and ßS alleles is determined by restriction endonuclease digestion of an end-labeled oligonucleotide probe hybridized in solution to the amplified ß-globin sequences. The ß-globin genotype can be determined in less than 1 day on samples containing significantly less than 1 microgram of genomic DNA.

The authors are in the Department of Human Genetics, Cetus Corporation, 1400 Fifty-Third Street, Emeryville, California 94608. The pres-ent address for N.A. is Department of Biological Sciences, University of Southern California, Los Angeles 90089-0371.

�

polymerase(12).Thesecondistheanalysisoftheß-globingenotypebysolutionhybrid-izationoftheamplifiedDNAwithaspecificoligonucleotideprobeandsubsequentdiges-tion with a restriction endonuclease (13).Thesetwotechniquesincreasethespeedandsensitivity,andlessenthecomplexityofpre-natal diagnosis for sickle cell anemia; theymay also be generally applicable to the di-agnosisofothergeneticdiseasesandintheuse of DNA probes for infectious diseasediagnosis.

Sequence amplification by polymerase chain reaction. We use a two-step proce-durefordetermining theß-globingenotypeof human genomic DNA samples. First, asmallportionoftheß-globingenesequencespanning thepolymorphicDde I restrictionsitediagnosticof theßAallele isamplified.Next, thepresenceorabsenceof theDde IrestrictionsiteintheamplifiedDNAsampleisdeterminedbysolutionhybridizationwithanend-labeledcomplementaryoligomerfol-

lowedbyrestrictionendonucleasedigestion,electrophoresis,andautoradiography.

Theß-globingenesegmentwasamplifiedbythepolymerasechainreaction(PCR)pro-cedureofMullisandFaloona(12)inwhichweused two20-baseoligonucleotideprim-ersthatflanktheregiontobeamplified.Oneprimer,PC04,iscomplementarytothe(+)-strand and the other, PC03, is complemen-tarytothe(−)-strand(Fig.1).Theannealingof PC04 to the (+)-strand of denatured ge-nomicDNAfollowedbyextensionwiththeKlenow fragment of Escherichia coliDNApolymerase I and deoxynucleotide triphos-phatesresultsinthesynthesisofa(−)-strandfragmentcontainingthetargetsequence.Atthesametime,asimilarreactionoccurswithPC03,creatinganew(+)-strand.Sincethesenewly synthesized DNA strands are them-selvestemplateforthePCRprimers,repeat-edcyclesofdenaturation,primerannealing,andextensionresultintheexponentialaccu-mulationofthe110–basepairregiondefined

Fig. 1. Sequence of synthetic oligonucleotide primers and probe and their relation to the target ß-globin region. (A) The primer PC03 is complementary to the (–)-strand and the primer PC04 is complementary to the (+)-strand of the ß-globin gene. The probe RS06 is complementary to the (-)-strand of the wild-type (ßA) sequence of ß-globin. RS10 is the “blocking oligomer”, an oligomer complementary to the RS06 probe except for three nucleotides, indicated by the downward arrows. It is added before enzyme digestion to the OR reaction to anneal to the excess RS06 oligomer and prevent nonspecific cleavage products due to hybridization of RS06 to nontarget DNA (13). Because of the mismatches within the Dde I and Hinf I restriction sites, the RS06/RS10 duplex is not cleaved by Dde I and Hinf I digestion. (B) The relation between the primers, the probe, and the target ß-globin sequence. The upward arrow indicates the ß-globin initiation codon. The downward arrows indicate nucleotide differences between ß- and δ-globin. The polymorphic Dde I site (CTCAG) is represented by a single horizontal dashed line (D), and the invariant Hinf I (GACTC) site is represented by double horizontal dashed lines (H).

�

bytheprimers.Anexampleofthedegreeofspecificgene

amplification achieved by the PCR methodisshowninFig.2A.SamplesofDNA(1µg)wereamplified for20cyclesanda fractionof each sample, equivalent to 36 ng of theoriginalDNA,wassubjectedtoalkalinegelelectrophoresis and transferred to a nylonfilter.Thefilter was thenhybridizedwith a32P-labeled 40-base oligonucleotide probe,RS06, which is complementary to the tar-get sequence (Fig. IA) but not to the PCRprimers.The results, after a 2-hour autora-diographic exposure, show that a fragmenthybridizingwiththeRS06probemigratesatthepositionexpectedoftheamplifiedtargetDNAsegment (110bases) (lanes 1 and2).NohybridizationwiththeRS06probecould

be detected in unamplified DNA (lane 4).WhenPCRamplificationwasperformedonaDNAsamplederivedfromanindividualwithhereditarypersistenceoffetalhemoglobininwhichboththe8-andß-globingenesarede-leted(14),againno110-basefragmentwasdetected(lane3).Toestimate theyieldandefficiencyof20cyclesofPCRamplification,wepreparedaSouthernblot that contained36ngofanamplifiedgenomicDNAsampleand a dilution series consisting of variousamounts of cloned ß-globin sequence. Theefficiency was calculated according to theformula:(1+X)n=Y,whereX is themeanefficiency per cycle, n is the number ofPCRcycles, andY is theextentofamplifi-cation(yield)afterncycles(forexample,a200,000-fold increase after 20 cycles).The

Fig. 2. Southern analysis of PCR amplified genomic DNA with the RS06 probe. (A) Samples (1 µg) of genomic DNA were dispensed in microcentrifuge tubes and adjusted to 100 µl in a buffer containing 10 mM tris, pH 7.5, 50 mM NaCl, 10 mM MgCl

2, 1.5 mM deoxynucleotide trisphosphate [(dNTP) each of

all four was used], 1 µM PC03, and 1 µM PCO4. After heating for 5 minutes at 95°C (to denature the genomic DNA), the tubes were centrifuged for 10 seconds in a microcentrifuge to remove the condensation. The samples were immediately transferred to a 30°C heat block for 2 minutes to allow the PC03 and PC04 primers to anneal to their target sequences. At the end of this period, 2 µl of the Klenow fragment of E. coli DNA polymerase I (Biolabs, 0.5 unit/µl in 10 mM tris, pH 7.5, 50 mM NaCI, 10 mM

MgCl2) was added, and the incubation was allowed to proceed for an additional 2 minutes at 30°C. This

cycle—denaturation, centrifugation, hybridization, and extension—was repeated 19 more times, except that subsequent denaturations were done for 2 instead of 5 minutes. (The final volume after 20 cycles was 140 µl.) Thirty-six nanograms of the amplified genomic DNA (5 µl) were applied to a 4 percent Nusieve (FMC) alkaline agarose minigel and subjected to electrophoresis (50 V), for 2 hours until the bromcresol green dye front reached 4 cm. After neutralization and transfer to Genetrans nylon mem-brane (Plasco), the filter was “prehybridized” in 10 ml 3x SSPE (I x SSPE is 0.18M NaCl, 10 mM NaH

2PO

4,

1 mM EDTA, pH 7.4), 5x DET (Ix DET is 0.02 percent each polyvinylpyrrolidone, Ficoll, and bovine serum albumin; 0.2 mM tris, 0.2 mM EDTA, pH 8.0), 0.5 percent SDS, and 30 percent formamide for 4 hours at 42°C. Hybridization with 1.0 pmol of phosphorylated (with [γ-32P]ATP) RS06 (~5 µCi/pmol) in 10 ml of the same buffer was carried out for 18 hours at 42°C. The filter was washed twice in 2x SSPE, 0.1 percent sodium dodecyl sulfate (SDS) at room temperature for 30 minutes, and autoradiographed at -70°C for 2 hours with a single intensification screen. (Lanes 1 to 3) DNA’s isolated from the cell lines Molt4, SC01, and GM2064, respectively. Molt4 is homozygous for the normal, wild-type allele of ß-globin (ßAßA), SC-1 is homozygous for the sickle cell allele (ßSßS), and GM2064 is a cell line in which the ß- and δ-globin genes have been deleted (ΔΔ) (13). (Lane 4) Contains 36 ng of Molt4 DNA that was not PCR amplified. The horizontal arrow indicates the position of a 114-base marker fragment obtained by digestion of pBR328 with Nar I. (B) Thirty-six nanograms of 20-cycle amplified Molt4 DNA (see above) was loaded onto a Nusieve gel along with measured amounts of Hae III-Mae I digested pBR328: : ßA (13) calculated to represent the molar increase in ß-globin target sequences at PCR efficiencies of 70, 75, 80, 85, 90, 95, and 100 percent (lanes 2 to 8, respectively). DNA was transferred to Genetrans and hybridized with the labeled RS06 probe as described above. (Lane 1) Molt4 DNA (36 ng); (lanes 2 to 8) 7.3 x 10-4 pmol, 1.3 x 10-3 pmol, 2.3 x 10-3 pmol, 4.0 x 10-3 pmol, 6.8 x 10-3 pmol, 1.1 x 10-2 pmol, and 1.9 x 10-2 pmol of pBR328:: ßA, respectively (20).

�

amountsofclonedß-globinsequencesusedinthisexperimentwerecalculatedtorepre-sentefficienciesof70to100percent.

Thereconstructionswerepreparedbydi-gestingtheß-globinplasmid,pBR328::ßA,withtherestrictionenzymesHaeIIIandMaeI.Bothoftheseenzymescleavetheß-globingenewithinorverynear to the20base re-gionsthathybridizetothePCRprimers,gen-eratinga103–basepair(bp)fragmentthatisalmost identical in size andcomposition tothe110-bpsegmentcreatedbyPCRampli-fication.After hybridization with the RS06probeandautoradiography,theamplifiedge-nomicsamplewascomparedwiththeknownstandards,andtheresultindicatedanoverall

efficiencyofapproximately85percent(Fig.2B), representing an amplificationof about220,000times(1.8520).

Distinguishing the ßA and ßS alleles by the oligomer restriction method.Wehavepreviously described a rapid solution hy-bridizationmethodthatcanindicatewhetheragenomicDNAsamplecontainsa specificrestriction enzyme site at, in principle, anychromosomal location (13). This method,called oligomer restriction (OR), involvesthe stringent hybridization of a 32P end-la-beled oligonucleotide probe to the specificsegment of the denatured genomic DNAwhich spans the target restriction site.Theability of a mismatch within the restriction

Fig. 3. Schematic diagram of oligo-mer restriction by sequential di-gestion to identify ßA_ and ßS-spe-cific cleavage products. The DNA sequences shown are the regions of the ß-globin genomic DNA and the RS06 hybridization probe con-taining the invariant Hinf I site (GANTC, where N represents any nucleotide) and the polymorphic Dde I site (CTNAG). The remaining DNA sequences are represented as solid horizontal lines. The as-terisk indicates the position of the radioactive 32P label attached to the 5’-end of the RS06 probe with polynucleotide kinase. (A) Outline of the procedure and expected results when RS06 anneals to the normal ß-globin gene (ßA). After denaturation of the genomic DNA and hybridization of the labeled RS06 probe to the complementary target sequence in the ßA gene, di-gestion of the probe-target hybrid with Dde I causes the release of a labeled (8-nt) cleavage product. Because of the relatively stringent conditions during Dde I digestion, the 8-nt cleavage product dissoci-ates from the genomic DNA and the subsequent digestion with Hinf I has no effect. (B) Outline of Dde I and Hinf I digestion after hybridiza-tion of the RS06 probe to the sickle cell allele (ßS). As a consequence of the ßS mutation, the probe-target hybrid contains an A-A mismatch within the Dde I site and is not cleaved by the Dde I endonuclease. The Hinf I site, however, remains intact and digestion with that enzyme generates a labeled 3-nt product. Thus, the presence of the ßA allele is revealed by the release of a labeled 8-nt fragment, while the pres-ence of ßS is indicated by a labeled 3-nt fragment.

�

sitetopreventcleavageoftheduplexformedbetweentheprobeandthetargetgenomicse-quenceisthebasisfordetectingallelicvari-ants.The presence of the restriction site inthetargetDNAisrevealedbytheappearanceofaspecific labeled fragmentgeneratedbycleavageoftheprobe.

For the diagnosis of sickle cell anemia,theprobewasdesignedtobecomplementaryto a region of the ß-globin gene locus sur-rounding the sixth codon. In the ßA allele,thenucleotide(nt)sequenceatthispositioncontainsaDde I restrictionsite,butdue tothesinglebasemutation,thissiteisabsentintheßSallele.Ourstrategyforgeneratingspe-cificprobecleavageproductsforeachalleleisshowninFig.3.ItisbasedonthepresenceofaninvariantHinfIrestrictionenzymesiteimmediately adjacent to the polymorphicDde I restriction site. Resolution of the la-beledoligomercleavageproductsproducedbysequentialdigestionwithDdeIandHinfI allows us to distinguish between the twoalleles.Inanindividualhomozygousforthewild-type ß-globin alleleAA, Dde I diges-tion will produce a labeled octamer (8 nt)fromtheprobe.Becauseofitsshortlength,the 8-nt cleavage product will dissociatefromthegenomictargetDNAandthesubse-quentdigestionwithHinfIhasnoeffect.In

thecaseofSShomozygotes,however,DdeIdigestiondoesnot cleave theprobe sinceabasepairmismatchexistsintherecognitionsequenceformedbetweentheprobeandtar-getDNA.TheinvariantHinfIsitewillthenbecleavedduringHinfIdigestion,releasinga labeled trimer (3 nt). In anAS heterozy-gote,bothatrimerandanoctamerwouldbedetected.Theresolutionoftheintact40-baseprobe, the8-nt and the3-nt cleavageprod-uctsisachievedbypolyacrylamidegelelec-trophoresis.Experimentstestingthesequen-tialdigestionstrategywithplasmidscarryingtheßAandßSallelesshowthat,ineachcase,the expected probe cleavage products wereproduced(Fig.4).

Analysis of genomic DNA samples by PCR and OR. Eleven DNA samples de-rivedfromlymphoblastoidcelllinesorwhiteblood cells were analyzed for their ß-glo-bingenotypebystandardSouthernblottingandhybridizationof theMst IIRFLP(10),identifyingthegenotypesof thesamplesaseitherAA,AS,orSS.Sixof thesesamples(andoneadditionalone)werethenamplifiedbyPCRfor20cyclesstartingwith1µgofDNAeach.AnaliquotoftheamplifiedDNAsample (one-fourteenth of the original l-µgsample) was hybridized to the RS06 probeanddigestedwithDdeIandthenHinfI.A

Fig. 4. Demonstration of OR sequential digestion with cloned ß-globin genes. The sequential digestion strategy was demonstrated by anneal-ing the RS06 probe to the ß-globin plasmids pBR328::ßA and pBR328::ßS (13). The methods were similar to those described (13). Cloned ß-globin DNA (45 ng; 0.01 pmol) was placed in a microcentrifuge tube, adjusted to 30 µl with TE buffer (10 mM tris, 0.1 mM EDTA, pH 8.0), overlaid with 0.1 ml of mineral oil. The DNA was denatured by heating for 5 to 10 minutes at 95°C. Ten microliters of 0.6M NaCl containing 0.02 pmol of phosphorylated (with [-y-32P]ATP) RS06 probe oligomer (~5 µCi/pmol) was added and annealed for 60 minutes at 56°C. Unla-beled RS10 blocking oligomer (4 µI; 200 pmol/ml) (Fig. 1) (13) was then added, and the hybridization was continued for 5 to 10 minutes. Next, 5 µ of 100 mM MgCl

2 and 1 µI of Dde I (Biolabs, 10 units) was added and

incubated for 20 minutes at 56°C; 1 µl of Hinf I (Biolabs, 10 units) was added and digestion was continued for 20 minutes at the same tem-perature. The reaction was terminated by the addition of 4 µl of 100 mM EDTA and 6 µl of tracking dye to a final volume of 61 µl; a portion (8 µI) (6 ng, 0.0013 pmol) was applied to a 0.75-mm thick, 30 percent polyacrylamide minigel (19 acrylamide: 1 bis) in a Hoefer SE200 appa-

ratus and subjected to electrophoresis (300 V) for 1 hour until the bromphenol blue dye front reached 3 cm. The top 1.5 cm of the gel, containing intact RS06, was cut off and discarded. The remaining gel was autoradiographed with a single intensification screen for 18 hours at -70°C. (Lane 1) six nanograms of pBR328:: ßA; (lane 2) 3 ng of pBR328::ßA plus 3 ng of pBR328:: ßS; and (lane 3) 6 ng of pBR328:: ßS.

�

portion(one-tenth)of thisoligomerrestric-tion reaction was analyzed on a polyacryl-amidegel to resolve thecleavageproducts,and the results obtained after 6 hours ofautoradiographyareshownFig.5.Thehighsensitivity achieved with the PCR and ORmethod is demonstrated by the strength ofthe autoradiographic signal derived fromonly1/140oftheoriginall-µgsample(7ng).EachsampledeterminedtobeAAbyRFLPanalysisshowedastrong8-ntfragmentwhilethosetypedasSSshowedastrong3-ntfrag-ment.AnalysisoftheknownASsamplesre-vealedbothcleavageproducts.

IntheanalysisoftheAAsamples,afaint3-ntcouldbedetectedinadditiontothepri-mary8-ntsignal.Thereasonsfor thisbandremain unclear, although incomplete Dde Icleavageor theoccasional failureof the8-nt fragment to disassociate from the targetDNAmaycontributetothenonspecific3-ntproductgeneratedbyHinfIdigestion.Intheanalysis of the SS samples, a very faint 8-ntbandwasalsoobservedinadditiontotheexpected 3-nt signal. We have determinedthat the background 8-nt product detectedin SS samples can be attributed to the δ-globin gene, which is highly homologous

toß-globin.Thenucleotidesequenceofthetwoß-globinprimersusedforamplificationis shown inFig.1.Thedownwardpointingarrows indicate thedifferencesbetween theß- and δ-globin genes. We hypothesizedthatthefaint8-ntsignalobservedintheSSsamples was due to some amplification oftheδ-globingeneby theseprimers and thesubsequentcross-hybridizationoftheampli-fiedδ sequenceswith theRS06probeusedintheORprocedure.δ-GlobinhasthesameDdeIsiteasnormalß-globin,andtheduplexformed between an amplified δ gene seg-mentandtheRS06probewouldbeexpectedtoyieldan8-ntfragmentonDdeIdigestioneven though there are sequence differences(four mismatch out of 40 bases) betweenRS06andδ-globin.Itislikelythatδ-globinsequencesmaybeamplifiedtosomeextentanddetectedweaklywiththeRS06probeinall DNA samples, but that its contributionto the total signal isverysmallanddetect-ableonlywhenthesampleisSSandno8-ntfragmentfromtheß-globingeneisexpected.Wetested thishypothesisby treatinganSSDNA sample before amplification with theenzymeMboI.SincethereisarecognitionsiteforthisenzymeinthetargetDNAofthe

Fig. 5. Determination of the ß-globin genotype in human genomic DNA with PCR-OR. Samples (1 µg) of human genomic DNA were amplified for 20 cycles (as described in Fig. 2A). The amplified DNA’s (71 ng) were hybridized to the RS06 probe and serially digested with Dde I and Hinf I (as de-scribed in Fig. 4). Each sample (6 µl) was analyzed by 30 percent polyacrylamide gel electrophoresis and autoradiographed for 6 hours at -70°C with one intensification screen. Each lane contains 7 ng of genomic DNA. (Lane 1) Unamplified Molt4 DNA (negative control); (lane 2) amplified Molt4 (ßAßA); (lane 3) SC-1 (ßSßS); (lane 4) GM2064 (ΔΔ); (lanes 5 to 11) clinical samples CH1 (ßAßA), CH2 (ßAßA), CH3 (ßSßS), CH4 (ßSßS), CH7 (ßAßS), CH8 (ßSßS), and CH12 (ßAßS), respectively.

Fig. 6. Effect of cycle number on signal strength. Genomic DNA (1 µg) from the clinical samples CH2 (ßAßA), CH12 (ßAßS), and CH5 (ßSßS) were amplifed for 15 and 20 cycles and equivalent amounts of genomic DNA (80 ng) were analyzed by oligomer restriction. (Lanes 1 to 3) DNA (20 ng) from CH2, CH12, and CH5, respectively, am-plified for 15 cycles; (lanes 4 to 6) DNA (20 ng) from CH2, CH12, and CH5, respectively, amplified for 15 cycles; (lanes 4 to 6) DNA (20ng) from CH2, CH12, and CH5, respectively, amplified for 15 cy-cles; (lanes 4 to 6) DNA (20 ng) from CH2, CH12, and CH5, respectively, amplified for 20 cycles. Autoradiographic exposure was for 2.5 hours at -70°C with one intensification screen.

10

δ-butnottheß-globingene,cleavageoftheδ gene between the regions that hybridizeto thePCRprimers shouldprevent its sub-sequentamplification (butnotofß-globin).OurresultsshowedthatanSSDNAsample,first digested with Mbo I, gave only the 3-nt product but not the 8-nt product, this isconsistent with the hypothesis of δ-globinamplification.

Effect of PCR cycle number on detec-tion threshold.The strength of the autora-diographsignaldetectedbyORasafunctionofPCRcyclenumberandautoradiographicexposurewasexamined.Thesignalintensityafter20cyclesisatleast20timesasstrongasthatfor15cyclesandthedeterminationoftheß-globingenotypecanbemadewithanautoradiographicexposure foronly2hours(Fig.6).Theobserved increaseof>20-foldisconsistentwithourestimatesof85percentefficiencypercycle,calculatedfromthedatainFig.2B(1.855=21.7).Coupledwiththetime that it takes to actually carry out thePCRandORprocedures,a20-cyclePCRal-lowsadiagnosistobemadeinlessthan10hourswithaDNAsampleof1µg.

SinceallofthepreviousPCRexperimentsweredonewith1µgofgenomicDNA,weexplored the effect of using significantlysmaller amounts of DNA as template forPCRamplification.Theresultsobtainedwith

20cyclesofPCRamplificationon500,100,20,and4ngofDNAfromanASindividualare shown in Fig. 7.After analysis of 1/40ofeachsampleby theORprocedureanda20-hour autoradiographic exposure, the ß-globingenotypecouldbeeasilydeterminedonDNAsamplesof20ngorabout100timesless than is needed for a typical Southerntransferandhybridizationexperiment.Inthisexperiment, only a small fraction (1/40) ofthestartingmaterialwasplacedon thegel;therefore it should be possible to analyzesamplesoflessthan20ngofgenomicDNA(20ng is equivalent to approximately6000haploid genomes) if a larger proportion ofthematerialwasutilizedin theORandgelelectrophoresissteps.

Diagnostic applications of the PCR-OR system. When currently available methodsareused,thecompletionofaprenataldiag-nosis for sickle cell anemia takes a periodof several days after the DNA is isolated.With 1 µg of genomic DNA, the ß-globingenotypecanbedeterminedbythePCR-ORmethod in less than 10 hours; 20 cycles ofamplification requires about 2 hours (eachfull cycle takes 6 to 7 minutes in our pro-tocol), the oligomer restriction procedureinvolving liquid hybridization and enzymedigestionsrequireanadditional2hours,andtheelectrophoresistakesaboutanhour.Au-toradiographic exposure for 4 hours is suf-ficient to generate a strong signal.Becausethis method includes a liquid hybridizationprotocol and involves the serial additionofreagentstoasingletube,itissimplertoper-formthanthestandardSoutherntransferandhybridizationprocedure.Priortoelectropho-resis,allofthereactionscanbedoneintwosmallmicrocentrifugetubesandcouldread-ilybeautomated.

The sensitivity, as well as the speed andsimplicity,ofthisprocedureisalsoimportantforclinicalapplications.Twentynanogramsof starting material can provide an easilydetectable result in an overnight autoradio-graphicexposure.ThissensitivitymakesthePCR-OR method particularly valuable incases where poor DNA yields are obtainedfrom prenatal samples. In addition, DNAsamples of poor quality (very low average

Fig. 7. Detection threshold for PCR-OR. Fivefold serial dilutions of genomic DNA (500, 100, 20, and 4 ng) from the sample CH12 (ßAßS) were amplified by 20 cycles of PCR and one-tenth each reaction (50, 10, 2, and 0.4 ng) was analyzed by OR. The gel continued (lane 1) genomic DNA; (12.5 ng); (lane 2) 2.5 ng; (lane 3) 0.5 ng; (lane 4) 0.1 ng; (lane 5) 12.5 ng genomic DNA from the globin dele-tion cell line GM2064. Autoradiographic exposure was for 20 hours at –70°C with an intensification screen.

11

molecularweight)cangiveexcellentresultsinthePCR-ORprotocol.

The PCR method is likely to be gener-ally applicable for specific gene amplifica-tion since a fragment encoding a portionof the HLA-DQa locus has recently beenamplifiedwiththisprocedure(15).WehavecarriedoutPCRamplificationona110-bpß-globinsequencewithanoverallefficiencypercycleofabout85percent.Wehavealsoamplified longerß-globin fragments (up to267bp),buttheyieldwaslowerunderourstandardconditions.Efficientamplificationofa267-bp fragment requiredsomevaria-tioninthePCRprocedure.Inprinciple,in-creasing thenumberofPCRcyclesshouldyieldevengreateramplificationthanthatre-portedhere(~220,000-foldafter20cycles).

Our method for the diagnosis of sicklecellanemiainvolvesthecouplingofthePCRprocedurewiththatofoligomerrestriction.Itwasdesignedtodistinguishbetweentwoallelesthatdifferbyapolymorphicrestric-tionsite.ThePCR-ORmethodisapplicableas well to the diagnosis of other diseaseswherethelesiondirectlyaffectsarestrictionenzymesiteorwhere thepolymorphicsiteisinstronglinkagedisequilibriumwiththediseasecausinglocus.Ifthepolymorphismis in linkage equilibrium with the disease,PCR-OR requires family studies to followtheinheritanceofthediseaselocus.

Inthecaseoftheß-globinlocus,thepres-enceof the invariantHinf I restriction siteadjacent to the polymorphic Dde I site al-lows a sequential digestion procedure toidentifyboththeßAandßSalleles.Inprin-ciple,thisapproachdoesnotrequirethatthetwositesbe immediatelyadjacentbutonlythatthecleavageproductgeneratedbydiges-tionatthepolymorphicsitedissociatefromthetargettopreventcuttingattheinvariantsite.Sincetherestrictionenzymedigestionconditionsusedherearefairlystringentforhybridization,weestimatethatthepolymor-phic and invariant sites could perhaps beseparatedbyasmuchas20bp.

The application of the PCR method toprenataldiagnosisdoesnotnecessarilyde-pend on a polymorphic restriction site orontheuseofradioactiveprobes.Infact,the

significantamplificationoftargetsequencesachieved by the PCR method allows theuse of nonisotopically labeled probes (16).Amplified target sequences could also beanalyzed by a number of other proceduresincluding those involving the hybridizationofsmalllabeledoligomerswhichwillformstableduplexesonlyifperfectlymatched(6, 7, 17, 18)andtherecentlyreportedmethodbased on the electrophoretic shifts of du-plexeswithbasepairmismatches(19).TheabilityofPCRprocedureotamplifyatargetDNA segment in genomic DNA raises thepossibility that its use may extend beyondthat of prenatal diagnosis to other areas ofmolecularbiology.

References and Notes 1. Y. W. Kan and A. M. Dozy, Proc. Natl. Acad. U.S.A. 75, 5631 (1978). 2. K. E. Davies and D. P. Ellis, Biochem. J. 226, 1 (1985). 3. Y. W. Kan, M.S. Golbus, A. M. Dozy, N. Eng. J. Med,

295, 1165 (1976). 4. A. M. Dozy et al., J. Am. Med. Assoc. 241, 1610 (1979). 5. E. M. Rubin and Y. W. Kan, Lancet 1985-I, 75 (1985). 6. M. Pirastu et al., N. Eng. J. Med. 309, 284 (1983). 7. S. H. Orkin, A. F. Markham, H. H. Kazazian, Jr., J. Clin. Invest. 71, 775 (1983). 8. R. F. Geever et al., Proc. Natl. Acad. Sci. U.S.A. 78, 5081 (1981). 9. J. C. Chang and Y. W. Kan, N. Eng. J. Med. 307, 30 (1982).10. S. H. Orkin et al., ibid., p. 32. 11. J. T. Wilson et al., Proc. Natl. Acad. Sci. U.S.A. 79,

3628 (1982).12. F. Faloona and K. Mullis, in preparation.13. R. Saiki, N. Arnheim, H. Erlich, Biotechnology 3,

1008 (1985).14. D. Tuan et al., Proc. Natl. Acad. Sci. U.S.A. 80,

6937 (1983).15. S. Scharf, unpublished data.16. R. Saiki et al., unpublished data.17. B. J. Conner et al., Proc. Natl. Acad. Sci. U.S.A. 80, 278 (1983).18. V. J. Kidd, R. B. Wallace, K. Itakura, S. L. C. Woo, Nature (London) 304, 230 (1983).19. R. M. Myers, N. Lumelsky, L. S. Lerman, T.

Maniatis, Nature (London) 313, 495 (1985).20. Calculation of the amounts of pßR328::ßA need- ed as standards to estimate PCR efficiency was

done in the following way. If we assume that a human haploid genome size is 3 x 109 bp, 36 ng of DNA is equivalent to 1.8 x 10-8 pmol. The

extent of amplification after 20 cycles at, for example, 85 percent efficiency is obtained with (1.8 x 10-8) (1.8520) or 4.0 x 10-3 pmol of plasmid DNA (Fig. 2B, lane 5).

20 September 1985; accepted 15 November 1985

1�

Science HAS SELECTED THE POLY-MERASE CHAIN REACTION AS themajor scientific development of 1989and has chosen for its first “Molecule ofthe Year” the DNA polymerase moleculethatdrivesthereaction.Thelistfromwhichthe polymerase chain reaction (PCR) waschosen included an impressive array of ac-complishmentsinmanyareasofscienceandtechnology; additional kudos are thereforeconferredto17oftheotherbig“stories”thatmade1989anexcitingyearforscientistsandfor followers and beneficiaries of science.AlthoughthePCRprocedurewasintroducedseveralyearsago,useofthetechniquetrulyburgeoned in1989; inmuch thesameway,thefullpotentialsofmanyoftheinteresting“runner-up” scientific achievements of thisyeararelikelytoberealizedsometimeintheyearstocome.

The first PCR papers were published in1985. Since that time PCR has grown intoanincreasinglypowerful,versatile,anduse-fultechnique.ThePCR“explosion”of1989can be seen as the result of a combinationofimprovementsinandoptimizationofthemethodology,introductionofnewvariationsonthebasicPCRtheme,andgrowingaware-nessbyscientistsofwhatPCRhastooffer.WithPCR,tinybitsofembedded,oftenhid-den, genetic information can be amplifiedinto large quantities of accessible, identifi-able, and analyzablematerial.A single cellprovides enough material for analysis; asinglehair canbeused to identify an indi-vidual.

The basic PCR reaction. The startingmaterial for PCR, the “target sequence,” isa gene or segment of DNA. In a matter ofhours, this target sequence can be ampli-fiedamillionfold.Howthisisaccomplishedis shown in the accompanying figure. Thecomplementarystrandsofadouble-strandedmoleculeofDNAareseparatedbyheating.Two small pieces of synthetic DNA, eachcomplementing a specific sequence at one

end of the target sequence, serve as prim-ers.Eachprimerbindstoitscomplementarysequence. Polymerases start at each primerandcopythesequenceofthatstrand.Withinashorttime,exactreplicasofthetargetse-quence have been produced. In subsequentcycles, double-stranded molecules of boththe original DNA and the copies are sepa-rated;primersbindagaintocomplementarysequences and the polymerase replicatesthem.Attheendofmanycycles,thepoolisgreatlyenrichedinthesmallpiecesofDNAthathavethetargetsequences,andthisam-plifiedgenetic informationis thenavailableforfurtheranalysis.

Evolving PCR. Many improvements onthe original PCR method have been made.One of the first was the substitution of aheat-stable enzyme for the original DNApolymerase, which was heat-labile and hadtobereplenishedaftereachcycle.Thestable“Taqpolymerase,”whichcomesfrombacte-riathatliveinhotsprings,continuesworkingalmostindefinitelydespitetheheatingsteps.Taq polymerase improved the yield, gener-atedmorespecificandlongerproducts,andfacilitatedautomation.

Newstrategieshavealsobeendevisedforflanking unknown sequences with definedprimersites.ForstandardPCR,thesequenc-esatbothendsofatargetsequencehavetobeknown.“Inverse”PCRprovidesawayofsequencing DNA outside the primer sitesratherthanbetweentwoprimersites.Primermolecules are synthesized with their se-quencesreversed.ThetargetDNAiscutandcircularized, and, when the polymerase ex-tendstheprimer,itdoessoaroundthecirclein the direction opposite that which wouldhave been taken by standard PCR primers.“Anchored” PCR was developed for study-ing genes that encode proteins for whichpartial sequences are known. For anchoredPCR, only one defined primer sequence isneeded,nottwo.

Theimplicationsofinverseandanchored

The Molecule of the YearRuth Levy Guyer and Daniel E. Koshland, Jr.

13

PCR for DNA sequencing are astound-ing: enormous stretches of DNA can besequenced once a tiny bit of sequence isknown.BothtechniquesmakeitpossibletoproceedalongtheDNA,continuallyredefin-ing“ends”towhichsyntheticprimerscanbeboundandthenextended.

Applications of PCR. The basic PCRprocedurehasbeenvaluableindiseasediag-nosis because specific DNA sequences canbe amplifiedenormously (theneedle in thehaystack). One of the first uses led to im-proveddiagnosisofageneticdisease(sicklecellanemia),becausethePCRtechniquede-pended on much less clinical material thanstandardprocedures.(BecausePCRisexqui-sitelysensitive,unusualcareistakentoavoidtheamplificationofcontaminants.)PCRcanalsobeusedtoamplifytraceamountsofge-neticmaterialof infectiousagents inblood,cells, water, food, and other clinical andenvironmental samples. PCR-based testsare especially valuable for detecting patho-gens that are difficult or impossible to cul-ture,suchastheagentsofLymediseaseandAIDS. For cancer diagnosis and cancer re-search,PCRcanindicatewhatgenesareex-pressedorturnedoff,becausethemessengerRNAmoleculesassociatedwithsuchgenescanbeconvertedintocomplementaryDNAsequencesthatthencanbeamplified.

DNA samples in trace materials (semen,

blood, hairs) found at the scene of a crimehavebeencomparedwithDNAsamplesfromcrime suspects; both acquittals and convic-tionshave resulted fromsuchcomparisons.Missing persons have also been positivelyidentified through PCR-based comparisons.The resolution of paternity cases has beenaidedbycomparingDNAfromachildwiththat of the alleged father. And matches oftransplantdonorsandrecipientsarefacilitat-edwithPCR.“Universal”primersarebeingusedtodeterminetheextentofhomologyinthesequencesofconservedgenesfromdif-ferent samples. Such comparisons, whichhelp to establish evolutionary relationsamongorganisms, can even include extinctorganisms,becauseDNAsamplesextractedfrom mummies, bones, and other archivalmaterialscanbeused.

PCRmaysoonreplacegenecloningastheamplificationmethodofchoiceforgenese-quencing,forwhichlargeamountsofDNAareneeded.PCR isalsoprovidingnewop-tions inmoleculargeneticsstudies foradd-inggeneticinformationtotargetmaterialsorforalteringwhatisalreadythere.

The rate at which new PCR-based tech-niques have been developed suggests thatthistechnologyisproliferatingasrapidlyasitsTaqpolymerasemoleculesreplicatetargetsequences.

1�

Developed in the mid 1990s for theanalysis and quantification of nu-cleicacids, real-timePCR isamo-lecularbiological techniquegaining rapidlyinpopularity.Itisbasedonthetechniqueofthe polymerase chain reaction (PCR) thatwasfirst envisionedbyKaryMullis almost20yearsago,duringamoonlitdrivethroughtheredwoodhillsofCalifornia(1).Thetech-nology of PCR (2) has become one of themostinfluentialdiscoveriesofthemolecularbiologyrevolutionandoneforwhichMullisreceived the Nobel Prize in 1993. Becauseof the impactofPCRand the thermostableTaq DNA polymerase (the enzyme respon-sible for the PCR revolution), the pair wasnamedasthefirst“MoleculeoftheYear”byScience in1989(3).Inmanyways,therecentdevelopmentofreal-timePCRseemsset tochangethegeneraluseofPCR.

The advancement provided by the real-timeversionofPCRisduetoitsuniqueabil-itytomonitorthecompleteDNAamplifica-tionprocess.DuringconventionalPCR,thetwostrandsofaDNAmoleculearesubjectedtoaseriesofheatingandcoolingcyclesthatresultinDNAstrandseparation,oligonucle-otide primer annealing, and thermostableTaq DNA polymerase–directed primer ex-tension, ultimately generating two identicaldaughter strands. Iterative cycling of theprocess exponentially amplifies thenumberoforiginalDNAmolecules,hencethetermPCR(4).AftercompletionofthePCRreac-tion,amplificationproductsareanalyzedbysize-fractionation of the amplified samplewiththeuseofgelelectrophoresis.

In the mid 1990s, researchers showedthatthe5′ nucleaseactivityoftheTaq DNA

A Technique WhoseTime Has ComeNigel J.Walker

polymerasecouldbeexploitedasamethodtoindirectlyassessthelevelofDNAampli-ficationwith theuseof specificfluorescentprobes(5),eliminatingtheneedforelectro-phoresis. Around the same time, research-ersshowedthatreal-timemonitoringof theDNAamplificationwithinthePCRreactiontube during the PCR could be achieved byusingfluorescentDNAbindingdyes,whichis knownaskineticPCR (6).The couplingof these twoprocesses (7,8) led to today’stechnology of fluorescence detection real-timePCR.

Ingeneral,analysisofamplificationdur-ingreal-timePCRhasbeenachievedbyde-tectingthefluorescencethatiseitherdirectlyor indirectly associated with the accumula-tionofthenewlyamplifiedDNA(seefigure,page15).Thedetectionsystemthatisalmostsynonymouswithreal-timePCRisthe“Taq-man”system(8),whichusesafluorescenceresonanceenergytransfer(FRET)probeasareportersystem.AFRETprobeisashortoli-gonucleotidethatiscomplementarytooneofthestrands.Theprobecontainsa“reporter”and a “quencher” fluorescent molecule atthe5′ and3′ endoftheprobe,respectively.Thisprobeisincludedinthereal-timePCRreactionalongwiththerequiredforwardandreverse PCR primers.The quencher fluoro-chrome on the probe, because it is in suchclose proximity to the reporter, is able toquench thefluorescenceof the reporter.AstheTaq DNApolymeraseenzymereplicatesthenewstrandofDNA,thenucleaseactiv-itydegrades theFRETprobeat the5′ end,which isbound to templateDNAstrand, inamannermuchlikethePacManvideogamecharacter.This degradation releases the re-porterfluorochromefromitsproximitytothequencher,resultinginfluorescenceofthere-porter.Accumulationoffluorescentreportermolecules,asaresultofamplificationofthe

The author is in the Environmental Toxicology Program, National Institute of Environmental Health Sciences, Research Triangle Park, NC 27709, USA. E-mail: walker3@niehs.nih.gov

1�

target,canthenbedetectedbyanappropriateopticalsensingsystemsuchastheTaqman,an “indirect” system that detects the accu-mulationoffluorescenceratherthantheam-plifiedDNAitself.Incontrast,acommonlyused “direct” method uses a fluorescentDNA (SYBR green) that binds nonspecifi-cally to double-stranded DNA, and the ac-cumulationofthefluorescenceboundtotheamplifiedDNAtargetismeasured.The“Mo-lecularBeacon”technologyisanotherdirectapproach that uses FRET-based fluorescentprobestobindtheamplifiedDNA.Intheun-boundstate,thequencherandreporterfluo-rochromesaremaintainedincloseproximityviaahairpinloopdesignedintothesequenceof theprobe.Bindingof theprobeata tar-get sequence–specific region to its comple-mentarystrandontheamplifiedtargetDNAseparatesthetwofluorochromes,therebyal-leviating the FRET interference and allow-ingthereportertofluoresce.RelatedsystemsthatuseFRET-basedPCRprimers incorpo-ratedintotheamplifiedDNAhavealsobeendeveloped(9).Inthesesystems,thereporterandquencherfluorochromesaremaintainedin a hairpin loop structure via a sequencethatisaddedtothe5′ endofoneofthePCRprimers.Disruptionofthehairpinloopstruc-tureduringincorporationoftheprimerintotheamplifiedDNAproductresultsinlossofthe FRET interference, leading to fluores-

cence of the reporter molecule. Choosingadetectionsystemisamajorconsiderationin developing a real-time PCR assay, andeach type of system has its pros and cons.Thedecisionisoftenacompromisebetweendesiredspecificity,assaydevelopment time,andcostperassay.

An inherent property of PCR that is ex-ploited in real-time PCR is that the morecopies of nucleic acid one starts with, thefewer cycles of template amplification ittakestomakeaspecificnumberofproducts.Therefore, thenumberofcyclesneededforthe amplification-associated fluorescenceto reachaspecific threshold levelofdetec-tion(theC

Tvalue)isinverselycorrelatedto

the amount of nucleic acid that was in theoriginal sample (see figure, top page 16).Because the progress of amplification ismonitored throughout the PCR process inreal-timePCR,aC

Tvaluecanbedetermined

during the exponential phase of a PCR re-action,whenamplification ismost efficientand least affected by reaction-limiting con-ditions.ThequantityofDNAinthesamplecan thenbeobtainedby interpolationof itsC

TvalueversusalinearstandardcurveofC

T

valuesobtainedfromaseriallydilutedstan-dard solution (see figure, middle page 16).Inpractice,suchcurvesarelinearovermorethan five orders of magnitude, a dynamicrange that isunsurpassedbyothermethods

Detection systems. Fluorescent detection of amplification can be achieved using double-stranded DNA binding dyes like SYBR green (green starbursts) or with FRET-based probes such as Taqman 5′nuclease-sensitive probes or DNA binding probes (red starbursts).

1�

forquantitativeDNAanalysis.It isalsopos-sible to quantify differences in nucleic acidlevelswithoutsuchcurves.Forexample,dur-ingtheexponentialphaseofPCR,thenumberof DNA strands theoretically doubles duringeach cycle (assuming amplification is 100%efficient). Consequently, a sample that hastwicethenumberofstartingcopiescomparedwith another sample would require one lesscycleofamplification togenerateanequiva-lentnumberofproductstrands.Byusingthedifference in the C

T values for two samples,

therefore, one can mathematically determinetherelativedifferenceinthelevelofthenucle-

ic acid of interest of differentsamples(10).

Instrumentationsystemsforreal-timePCRhaveundergoneextensivechangesalready.Ini-tiallytheywerelargemachinesthat took up almost an entirelab bench and were incapableofa trueanalysis inreal time.As the technology improved,systems became available thatallow the PCR reaction to bemonitoredasitoccurs.Thesizeand cost of these systems hasbeen reduced so dramaticallythat they are now reasonablypriced units with very smallfootprints.Giventheaddedca-pabilitiesofthereal-timetech-nology, it is likely that in thenear future such systems willbecomethestandardPCRplat-forminthegenerallaboratory,inmuchthesamewaythatthe96- and 384-microwell PCRmachineshavesupersededini-tiallarge-tubeformatunits.

In the laboratory, real-timereverse transcriptase–PCR(RT-PCR) has become themethodofchoicefortherapidand quantitative examinationof the expression of specificgenes. Users of the methodcan rapidly, reproducibly, andstatistically determine evensmall(twofold)changesinthe

expressionofhundredsof samplesperday.Such analyses would be very difficult andlaboriouswiththeuseoftraditionalhybrid-ization techniques, such as Northern blot-ting or RNase protection assays. The lowRNAquantitiesrequired(nanograms)makethis assay more suitable for the analysis ofsamplesobtainedfromlasermicrodissectedtissue. In addition, realtime approaches areperfectlyadaptableforhigh-throughputandquantitative gene expression studies, espe-ciallywith the recentdevelopmentofhigh-throughput 384-well real-time PCR instru-ments.Theseadvancesaretimely,giventhe

Amplification time for fluorescence detection. Curves representing the cycle-dependent fluorescence associated with amplification of a specific gene product from 0.1 to 100 ng total RNA are shown. The C

T value for each sample is determined from

each curve as the cycle at which the fluorescence achieves a specific threshold value. ΔRn, normalized fluorescence.

Dynamic range of real-time PCR. Fluorescent detection of the amplification of the gene product is linear over five orders of magnitude. Values were calculated on the basis of C

T values from

amplification figure. Slope, –3.538; y intercept, 31.806; correlation coefficient, 0.991

1�

completionof theHumanGenomeProject.Recentadvancesinmicrotechnologies,suchasDNAchiptechnology,coupledwithpro-teomics and metabonomics (11) are estab-lishing the new science of systems biology(12).This division of biology seeks to useinformation from the integrated analysis ofgenesorproteinsandmetabolismtodevelopcomputational models of cellular functionand physiology. Such models are enhancedthroughtheavailabilityofquantitativeinfor-mationontheexpressionofgenes,proteins,or metabolites (13). Quantitative real-timePCRanalysisofgeneexpressionislikelytoplayakeyroleinthisburgeoningfield.

Though PCR has influenced drasticallythe way in which molecular biological re-searchisconducted,itsimpactonourevery-dayliveshasyettobefullyrealized.Becauseofitsreduceddetectiontimesandsimplifica-tionofquantitation, real-timePCRsystemsarelikelytohavethegreatestimpactonthegeneralpublic inenvironmentalmonitoringand nucleic acid diagnostics. Most uses ofPCR-baseddiagnosticsstill requirespecial-izedlaboratoryservices,butseveralcompa-nies are developing devices that will allowsuch rapid PCR analysis to be performedinthenearfutureatthepointthesampleiscollected instead of in a far-removed labo-ratory.Devicescouplingmicrofluidtechnol-ogy with realtime PCR analysis of nucleicaciddetectionaresettomakeahugeimpactinmany facetsofour lives asweenter thegenomicinformationera(14).Suchdeviceswould incorporate nucleic acid extractionandPCRdetectioninsmall,disposablesys-

tems that are readable through attachmentsto a personal computer. This technologycould potentially enable inhome testing fornucleicacidsfrombacterialorviralpatho-gens.Asmethodsforthequantitativeanaly-sisofgeneexpressionandDNAlevelscon-tinuetoevolveinsophistication,devicesthatincorporate the concepts of real-time PCRwill likely herald the era of individualizedgenomicsandgenetictesting.

References 1. K. B. Mullis, Sci.Am. 262, 56 (1990). 2. R. K. Saiki et al., Science 230, 1350 (1985); R. K.

Saiki et al., Nature 324, 163 (1986). 3. R. L. Guyer, D. E. Koshland Jr., Science 246, 1543

(1989). 4. B. A.White, Ed., PCR Protocols: Current Methods

and Applications (Humana Press, Totowa, NJ, 1993), vol. 15.

5. P. M. Holland, R. D. Abramson, R.Watson, D. H. Gelfand, Proc. Natl. Acad. Sci. U.S.A. 88, 7276 (1991).

6. R. Higuchi, C. Fockler, G. Dollinger, R. Watson, Biotechnology 11, 1026 (1993); R. Higuchi, G. Dollinger, P. S.Walsh, R. Griffith, Biotechnology 10, 413 (1992).

7. U. E. Gibson, C. A. Heid, P. M.Williams, Genome Res. 6, 995 (1996).

8. C. A. Heid, J. Stevens, K. J. Livak, P. M.Williams, Genome Res. 6, 986 (1996).

9. N. J.Walker, J. Biochem. Mol. Toxicol. 15, 121 (2001).

10. M.W. Pfaffl, Nucleic Acids Res. 29, E45 (2001).11. E. K. Lobenhofer, P. R. Bushel, C. A. Afshari, H. K.

Hamadeh, Environ. Health Perspect. 109, 881 (2001); J. K. Nicholson, J. C. Lindon, E. Holmes, Xenobiotica 29, 1181 (1999).

12. W. W. Gibbs, Sci. Am. 285, 52 (2001); T. Ideker, T. Galitski, L. Hood, Annu. rev. Genomics Hum. Genet. 2, 343 (2001).

13. T. Ideker et al., Science 292, 929 (2001).14. P. Mitchell, Nature Biotechnol. 19, 717 (2001).

1�

The ubiquity and longevity of Sangersequencing(1)areremarkable.Analo-gous to semiconductors,measuresofcost and production have followed expo-nential trends (2). High-throughput centersgeneratedataataspeedof20rawbasesperinstrument-second and a cost of $1.00 perrawkilobase.Nonetheless,optimizationsofelectrophoretic methods may be reachingtheir limits. Meeting the challenge of the$1000 human genome requires a paradigmshiftinourunderlyingapproachtotheDNApolymer(3).

Cyclic array methods, an attractive classofalternativetechnologies,are“multiplex”inthattheyleverageasinglereagentvolumeto enzymatically manipulate thousands tomillions of immobilized DNA features inparallel.Readsarebuiltupover successivecycles of imaging-based data acquisition.Beyondthiscommonthread,thesetechnolo-giesdiversify inapanoplyofways:single-moleculeversusmultimoleculefeatures,or-dered versus disordered arrays, sequencingbiochemistry, scale of miniaturization, etc.

(3).Innovativeproof-of-conceptexperimentshavebeenreported,butaregenerallylimitedin termsof throughput, featuredensity,andlibrarycomplexity(4–9).Arangeofpracti-calandtechnicalhurdlesseparatethesetestsystems from competing with conventionalsequencingongenomic-scaleapplications.

Our approach to developing a more ma-ture alternative was guided by several con-siderations. (i) An integrated sequencingpipeline includes library construction, tem-plate amplification, and DNA sequencing.We therefore sought compatible protocolsthat multiplexed each step to an equivalentorder of magnitude. (ii)As more genomesare sequenced de novo, demand will likelyshift towardgenomic resequencing; e.g., tolookatvariationbetweenindividuals.Forre-sequencing,consensusaccuracyincreasesinimportancerelativetoreadlengthbecauseareadneedonlybe longenough tocorrectlyposition it on a reference genome. How-ever, a consensus accuracy of 99.99%, i.e.,the Bermuda standard, would still result inhundreds of errors in a microbial genomeand hundreds of thousands of errors in amammaliangenome.Toavoidunacceptablenumbersoffalse-positives,aconsensuserrorrateof1×10-6isamorereasonablestandardforwhichtoaim.(iii)Wesoughttodevelopsequencingchemistriescompatiblewithcon-ventional epifluorescence imaging. Diffrac-

3. S. Georgi, www.batteriesdigest.com/id380.htm (accessedJune 2005).

4. R. M. Alexander, J. Exp. Biol. 160, 55 (1991).5. G. A. Cavagna, N. C. Heglund, C. R. Taylor, Am. J.

Physiol. 233, R243 (1977).6. J. Drake, Wired 9, 90 (2001).7. S. Stanford, R. Pelrine, R. Kornbluh, Q. Pei, in Proceedings

of the 13th International Symposium on UnmannedUntethered Submersible Technology (AutonomousUndersea Systems Institute, Lee, NH, 2003).

8. T. Starner, J. Paradiso, in Low Power Electronics Design(CRC Press, Boca Raton, FL, 2004), p. 45–1.

9. G. A. Cavagna, M. Kaneko, J. Physiol. 268, 647 (1977).10. S. A. Gard, S. C. Miff, A. D. Kuo, Hum. Mov. Sci. 22,

597 (2004).11. Supporting material is available on Science Online.12. Because it is a prototype, there has been no attempt

to reduce the weight of the backpack—indeed, it issubstantially ‘‘overdesigned.’’ Further, the 5.6 kgincludes the weight of six load cells and one 25-cm-long transducer, each with accompanying bracketsand cables, as well as other components that will notbe present on a typical pack. In future prototypes, weestimate that the weight will exceed that of a normalbackpack by no more than 1 to 1.5 kg.

13. Under high-power conditions (5.6 km hourj1 with20- and 29-kg loads and 4.8 km hourj1 with a 38-kgload), power generation on the incline was the sameas on the flat. Under low-power conditions (4.8 kmhourj1 with 20- and 28-kg loads), electricity gener-ation on the incline was actually substantially greaterthan that on the flat (table S1).

14. R. Margaria, Biomechanics and Energetics of MuscularExercise (Clarendon, Oxford, 1976).

15. R. A. Ferguson et al., J. Physiol. 536, 261 (2001).16. G. A. Cavagna, P. A. Willems, M. A. Legramandi, N. C.

Heglund, J. Exp. Biol. 205, 3413 (2002).17. A. Grabowski, C. T. Farley, R. Kram, J. Appl. Physiol.

98, 579 (2005).18. J. M. Donelan, R. Kram, A. D. Kuo, J. Exp. Biol. 205,

3717 (2002).19. J. M. Donelan, R. Kram, A. D. Kuo, J. Biomech. 35, 117

(2002).20. J. S. Gottschall, R. Kram, J. Appl. Physiol. 94, 1766 (2003).21. Because this savings in metabolic energy represents

only 6% of the net energetic cost of walking with thebackpack (492 W) (table S3) (17, 18), accurate de-terminations of the position and movements of thecenter of mass, as well as the direction and magnitudeof the ground reaction forces, are essential to discernthe mechanism. This will require twin–force-platformsingle-leg measurements, as well as a complete kine-matics and mechanical energy analysis (19, 20). Theenergy analysis is made more complex because theposition of the load with respect to the backpackframe and the amount of energy stored in the back-pack springs vary during the gait cycle. Finally, elec-tromyogram measurements are also important totest whether a change in effective muscle momentarms may have caused a change in the volume ofactivated muscle and hence a change in metaboliccost (20, 27, 28).

22. K. Schmidt-Nielsen, Animal Physiology: Adaptationand Environment (Cambridge Univ. Press, Cambridge,ed. 3, 1988).

23. This assumes that electronic devices are being pow-ered in real time. If there were a power loss of 50%associated with storage (such as in batteries) and re-

covery of electrical energy, then these factors wouldbe halved.

24. When not walking, the rack can be disengaged andthe generator cranked by hand or by foot. Electricalpowers of È3 W are achievable by hand, and higherwattage can be achieved by using the leg to power it.

25. R. Kram, J. Appl. Physiol. 71, 1119 (1991).26. A. E. Minetti, J. Exp. Biol. 207, 1265 (2004).27. A. A. Biewener, C. T. Farley, T. J. Roberts, M. Temaner,

J. Appl. Physiol. 97, 2266 (2004).28. T. M. Griffin, T. J. Roberts, R. Kram, J. Appl. Physiol.

95, 172 (2003).29. This work was supported by NIH grants AR46125 and

AR38404. Some aspects of the project were supportedby Office of Naval Research grant N000140310568and a grant from the University of Pennsylvania Re-search Foundation. The authors thank Q. Zhang, H.Hofmann, W. Megill, and A. Dunham for helpful dis-cussions; R. Sprague, E. Maxwell, R. Essner, L. Gazit, M.Yuhas, and J. Milligan for helping with the experimen-tation; and F. Letterio for machining the backpacks.

Supporting Online Materialwww.sciencemag.org/cgi/content/full/309/5741/1725/DC1Materials and MethodsSOM TextFigs. S1 and S2Tables S1 to S4References

14 February 2005; accepted 25 July 200510.1126/science.1111063

Accurate Multiplex Polony

Sequencing of an Evolved

Bacterial GenomeJay Shendure,1*. Gregory J. Porreca,1*. Nikos B. Reppas,1

Xiaoxia Lin,1 John P. McCutcheon,2,3 Abraham M. Rosenbaum,1

Michael D. Wang,1 Kun Zhang,1 Robi D. Mitra,2 George M. Church1

We describe a DNA sequencing technology in which a commonly available,inexpensive epifluorescence microscope is converted to rapid nonelectrophoreticDNA sequencing automation. We apply this technology to resequence an evolvedstrain of Escherichia coli at less than one error per million consensus bases. Acell-free, mate-paired library provided single DNA molecules that were amplifiedin parallel to 1-micrometer beads by emulsion polymerase chain reaction.Millions of beads were immobilized in a polyacrylamide gel and subjected toautomated cycles of sequencing by ligation and four-color imaging. Cost perbase was roughly one-ninth as much as that of conventional sequencing. Ourprotocols were implemented with off-the-shelf instrumentation and reagents.

The ubiquity and longevity of Sanger sequenc-

ing (1) are remarkable. Analogous to semicon-ductors, measures of cost and production have

followed exponential trends (2). High-throughputcenters generate data at a speed of 20 raw bases

per instrument-second and a cost of $1.00 perraw kilobase. Nonetheless, optimizations of elec-

trophoretic methods may be reaching their lim-

its. Meeting the challenge of the $1000 humangenome requires a paradigm shift in our under-

lying approach to the DNA polymer (3).Cyclic array methods, an attractive class

of alternative technologies, are Bmultiplex[ inthat they leverage a single reagent volume to

enzymatically manipulate thousands to mil-lions of immobilized DNA features in paral-

lel. Reads are built up over successive cyclesof imaging-based data acquisition. Beyond

this common thread, these technologies di-versify in a panoply of ways: single-molecule

versus multimolecule features, ordered versusdisordered arrays, sequencing biochemistry,

scale of miniaturization, etc. (3). Innovative

proof-of-concept experiments have been re-ported, but are generally limited in terms of

throughput, feature density, and library com-plexity (4–9). A range of practical and tech-

nical hurdles separate these test systems fromcompeting with conventional sequencing on

genomic-scale applications.Our approach to developing a more mature

alternative was guided by several consider-ations. (i) An integrated sequencing pipeline

includes library construction, template ampli-fication, and DNA sequencing. We therefore

sought compatible protocols that multiplexedeach step to an equivalent order of magnitude.

(ii) As more genomes are sequenced de novo,demand will likely shift toward genomic rese-

quencing; e.g., to look at variation between in-dividuals. For resequencing, consensus accuracy

increases in importance relative to read lengthbecause a read need only be long enough to

correctly position it on a reference genome.

However, a consensus accuracy of 99.99%, i.e.,the Bermuda standard, would still result in hun-

dreds of errors in a microbial genome and hun-dreds of thousands of errors in a mammalian

genome. To avoid unacceptable numbers offalse-positives, a consensus error rate of 1

10j6 is a more reasonable standard for whichto aim. (iii) We sought to develop sequencing

chemistries compatible with conventional epi-fluorescence imaging. Diffraction-limited optics

with charge-coupled device detection achievesan excellent balance because it not only pro-

vides submicrometer resolution and high sen-sitivity for rapid data acquisition, but is also

inexpensive and easily implemented.

1Department of Genetics, Harvard Medical School,Boston, MA 02115, USA. 2Department of Genetics,3Howard Hughes Medical Institute, Washington Uni-versity, St. Louis, MO 63110, USA.

*These authors contributed equally to this work..To whom correspondence should be addressed.E-mail: shendure@alumni.princeton.edu (J.S.),gregory_porreca@student.hms.harvard.edu (G.J.P.)

R E P O R T S

9 SEPTEMBER 2005 VOL 309 SCIENCE www.sciencemag.org1728

Accurate Multiplex Polony Sequencing of an EvolvedBacterial GenomeJay Shendure,1*† Gregory J. Porreca,1*† Nikos B. Reppas,1 Xiaoxia Lin,1 John P. McCutcheon,2,3 Abraham M. Rosenbaum,1 Michael D. Wang,1 Kun Zhang,1 Robi D. Mitra,2 George M. Church1

We describe a DNA sequencing technology in which a commonly available, inexpensive epifluorescence microscope is converted to rapid nonelectrophoretic DNA sequencing automation. We apply this technology to resequence an evolved strain of Escherichia coli at less than one error per million consensus bases. A cell-free, mate-paired library provided single DNA molecules that were amplified in parallel to 1-micrometer beads by emulsion polymerase chain reaction. Millions of beads were immobilized in a polyacrylamide gel and subjected to automated cycles of sequencing by ligation and four-color imaging. Cost per base was roughly one-ninth as much as that of conventional sequencing. Our protocols were implemented with off-the-shelf instrumentation and reagents.

1�

tion-limited optics with charge-coupled de-vicedetectionachievesanexcellentbalancebecauseitnotonlyprovidessubmicrometerresolutionandhighsensitivityforrapiddataacquisition,butisalsoinexpensiveandeas-ilyimplemented.

Conventional shotgun libraries are con-structed by cloning fragmented genomicDNAof adefined size range into anEsch-erichia colivector.Sequencingreadsderivedfrom opposite ends of each fragment aretermed “mate-pairs.” To avoid bottlenecksimposed by E. coli transformation, we de-velopedamultiplexed,cell-freelibrarycon-struction protocol. Our strategy (Fig. 1A)uses a type IIs restriction endonuclease tobringsequencesseparatedonthegenomeby~1 kb into proximity. Each ~135–base pair(bp) library molecule contains two mate-paired17-to18-bptagsofuniquegenomicsequence,flankedandseparatedbyuniversalsequencesthatarecomplementarytoamplifica-tionorsequencingprimersusedinsubsequentsteps.The in vitro protocol (Note S1) resultsin a library with a complexity of ~1 million

unique,mate-pairedspecies.Conventionally, template amplification has

beenperformedbybacterialcoloniesthatmustbe individually picked. Polymerase colony, orpolony,technologiesperformmultiplexampli-ficationwhilemaintainingspatialclusteringofidenticalamplicons(10).Theseincludeinsitupolonies (11), in situ rolling circle amplifica-tion(RCA)(12),bridgepolymerasechainreac-tion(PCR)(13),picotiterPCR(9),andemul-sion PCR (14). In emulsion PCR (ePCR), awater-in-oilemulsionpermitsmillionsofnon-interacting amplifications within a milliliter-scale volume (15–17).Amplification productsof individual compartments are captured viainclusion of 1-µm paramagnetic beads bear-ing one of the PCR primers (14).Any singlebeadbearsthousandsofsingle-strandedcopiesofthesamePCRproduct,whereasdifferentbeads bear the products of different com-partmentalizedPCRreactions(Fig.1B).ThebeadsgeneratedbyePCRhavehighlydesir-ablecharacteristics:highsignaldensity,geo-metricuniformity,strongfeatureseparation,andasizethatissmallbutstillresolvableby

Conventional shotgun libraries are con-

structed by cloning fragmented genomic DNAof a defined size range into an Escherichia coli

vector. Sequencing reads derived from oppositeends of each fragment are termed Bmate-pairs.[

To avoid bottlenecks imposed by E. colitransformation, we developed a multiplexed,

cell-free library construction protocol. Ourstrategy (Fig. 1A) uses a type IIs restriction

endonuclease to bring sequences separated onthe genome by È1 kb into proximity. Each

È135–base pair (bp) library molecule containstwo mate-paired 17- to 18-bp tags of unique ge-

nomic sequence, flanked and separated by uni-

versal sequences that are complementary toamplification or sequencing primers used in

subsequent steps. The in vitro protocol (NoteS1) results in a library with a complexity of È1

million unique, mate-paired species.Conventionally, template amplification has

been performed by bacterial colonies that mustbe individually picked. Polymerase colony, or

polony, technologies perform multiplex ampli-fication while maintaining spatial clustering of

identical amplicons (10). These include in situpolonies (11), in situ rolling circle amplification

(RCA) (12), bridge polymerase chain reaction(PCR) (13), picotiter PCR (9), and emulsion

PCR (14). In emulsion PCR (ePCR), a water-in-oil emulsion permits millions of noninteract-

ing amplifications within a milliliter-scalevolume (15–17). Amplification products of in-

dividual compartments are captured via in-clusion of 1-mm paramagnetic beads bearing

one of the PCR primers (14). Any single beadbears thousands of single-stranded copies of the

same PCR product, whereas different beads bearthe products of different compartmentalized

PCR reactions (Fig. 1B). The beads generatedby ePCR have highly desirable characteristics:

high signal density, geometric uniformity, strongfeature separation, and a size that is small but

still resolvable by inexpensive optics.Provided that the template molecules are

sufficiently short (fig. S1), an optimized versionof the ePCR protocol described by Dressman

et al. (14) robustly and reproducibly amplifiesour complex libraries (Note S2). In practice,

ePCR yields empty, clonal, and nonclonal

beads, which arise from emulsion compartmentsthat initially have zero, one, or multiple template

molecules, respectively. Increasing templateconcentration in an ePCR reaction boosts the

fraction of amplified beads at the cost of greaternonclonality (14). To generate populations in

which a high fraction of beads was both ampli-fied and clonal, we developed a hybridization-

based in vitro enrichment method (Fig. 1C). Theprotocol is capable of a fivefold enrichment of

amplified beads (Note S3).Iterative interrogation of ePCR beads (Fig.

1D) requires immobilization in a format compat-ible with enzymatic manipulation and epifluo-

rescence imaging. We found that a simpleacrylamide-based gel system developed for in

situ polonies (6) was easily applied to ePCR

beads, resulting in a È1.5-cm2 array of dis-ordered, monolayered, immobilized beads (Note

S4, Fig. 2A).With few exceptions (18), sequencing bio-

chemistries rely on the discriminatory capaci-ties of polymerases and ligases (1, 6, 8, 19–22).

We evaluated a variety of sequencing protocolsin our system. A four-color sequencing by

ligation scheme (Bdegenerate ligation[) yieldedthe most promising results (Fig. 2, B and C). A

detailed graphical description of this method isshown in fig. S7. We begin by hybridizing an

Banchor primer[ to one of four positions

(immediately 5¶ or 3¶ to one of the two tags).We then perform an enyzmatic ligation reaction

of the anchor primer to a population of degen-erate nonamers that are labeled with fluorescent

dyes. At any given cycle, the population ofnonamers that is used is structured such that the

identity of one of its positions is correlated withthe identity of the fluorophore attached to that

nonamer. To the extent that the ligase discrim-inates for complementarity at that queried po-

sition, the fluorescent signal allows us to infer

the identity of that base (Fig. 2, B and C). After

performing the ligation and four-color imaging,the anchor primer:nonamer complexes are

stripped and a new cycle is begun. With T4DNA ligase, we can obtain accurate sequence

when the query position is as far as six basesfrom the ligation junction while ligating in

the 5¶Y3¶ direction, and seven bases fromthe ligation junction in the 3¶Y5¶ direction.

This allows us to access 13 bp per tag (ahexamer and heptamer separated by a 4- to

5-bp gap) and 26 bp per amplicon (2 tags 13 bp) (fig. S7).

Although the sequencing method presented

here can be performed manually, we benefitedfrom fully automating the procedure (fig. S3).

Our integrated liquid-handling and microscopysetup can be replicated with off-the-shelf com-

ponents at a cost of about $140,000. A detaileddescription of instrumentation and software is

provided in Notes S5 and S7.As a genomic-scale challenge, we sought a

microbial genome that was expected, relative toa reference sequence, to contain a modest num-

ber of both expected and unexpected differences.

Fig. 1. A multiplex approach to genome sequencing. (A) Sheared, size-selected genomic fragments(yellow) are circularized with a linker (red) bearing Mme I recognition sites (Note S1). Subsequentsteps, which include a rolling circle amplification, yield the 134- to 136-bp mate-paired librarymolecules shown at right. (B) ePCR (14 ) yields clonal template amplification on 1-mm beads (NoteS2). (C) Hybridization to nonmagnetic, low-density ‘‘capture beads’’ (dark blue) permits enrichmentof the amplified fraction (red) of magnetic ePCR beads by centrifugation (Note S3). Beads areimmobilized and mounted in a flowcell for automated sequencing (Note S4). (D) At each sequencingcycle, four-color imaging is performed across several hundred raster positions to determine thesequence of each amplified bead at a specific position in one of the tags. The structure of eachsequencing cycle is discussed in the text, Note S6, and fig. S7.

R E P O R T S

www.sciencemag.org SCIENCE VOL 309 9 SEPTEMBER 2005 1729

Fig. 1. A multiplex ap-proach to genome se-quencing. (A) Sheared, size-selected genomic fragments (yellow) are circularized with a linker (red) bearing Mme I recognition sites (Note S1). Subsequent steps, which include a rolling circle amplifica-tion, yield the 134- to 136-bp mate-paired li-brary molecules shown at right. (B) ePCR (14) yields clonal template amplification on 1-µm beads (Note S2). (C) Hybridization to non-magnetic, low-density ‘‘capture beads’’ (dark blue) permits enrich-ment of the amplified fraction (red) of mag-netic ePCR beads by centrifugation (Note S3). Beads are immobilized and mounted in a flowcell for automated sequencing (Note S4). (D) At each sequencing cycle, four-color imaging is performed across several hundred raster positions to determine the sequence of each amplified bead at a specific position in one of the tags. The structure of each sequencing cycle is discussed in the text, Note S6, and fig. S7.

�0

inexpensiveoptics.Provided that the templatemoleculesare

sufficiently short (fig. S1), an optimizedversion of the ePCR protocol described byDressmanet al.(14)robustlyandreproduc-ibly amplifies our complex libraries (NoteS2).Inpractice,ePCRyieldsempty,clonal,andnonclonalbeads,whicharisefromemul-sion compartments that initially have zero,one,ormultipletemplatemolecules,respec-tively. Increasing template concentration inanePCRreactionbooststhefractionofam-plifiedbeadsatthecostofgreaternonclonal-ity(14).Togeneratepopulationsinwhichahigh fraction of beads was both amplifiedand clonal, we developed a hybridization-basedinvitroenrichmentmethod(Fig.1C).Theprotocoliscapableofafivefoldenrich-mentofamplifiedbeads(NoteS3).

Iterative interrogation of ePCR beads(Fig. 1D) requires immobilization in a for-mat compatible with enzymatic manipula-tionandepifluorescenceimaging.Wefoundthat a simple acrylamide-based gel systemdevelopedforinsitupolonies(6)waseasilyappliedtoePCRbeads,resultingina~1.5-cm2 array of disordered, monolayered, im-mobilizedbeads(NoteS4,Fig.2A).

With few exceptions (18), sequencingbiochemistriesrelyonthediscriminatoryca-pacitiesofpolymerasesandligases(1, 6, 8, 19–22).Weevaluatedavarietyofsequenc-ing protocols in our system. A four-colorsequencingbyligationscheme(“degenerateligation”)yieldedthemostpromisingresults(Fig. 2,BandC).Adetailedgraphicalde-scriptionofthismethodisshowninfig.S7.Webeginbyhybridizingan“anchorprimer”tooneof fourpositions (immediately5´or3´tooneofthetwotags).Wethenperformanenyzmaticligationreactionoftheanchorprimer to a population of degenerate nona-mers thatare labeledwithfluorescentdyes.At any given cycle, the population of no-namers that is used is structured such thatthe identity of one of its positions is cor-related with the identity of the fluorophoreattached to thatnonamer.To theextent thatthe ligase discriminates for complementar-ity at that queried position, the fluorescentsignalallowsus to infer the identityof that

base(Fig.2,BandC).Afterperformingtheligation and four-color imaging, the anchorprimer:nonamercomplexesarestrippedandanewcycleisbegun.WithT4DNAligase,we can obtain accurate sequence when thequerypositionisasfarassixbasesfromtheligationjunctionwhileligatinginthe5´→3´direction,andsevenbasesfromtheligationjunctioninthe3´→5´direction.Thisallowsus to access 13 bp per tag (a hexamer andheptamerseparatedbya4-to5-bpgap)and26bpperamplicon(2tags×13bp)(fig.S7).

Although the sequencing method pre-sentedherecanbeperformedmanually,webenefited from fully automating the proce-dure(fig.S3).Ourintegratedliquid-handlingandmicroscopysetupcanbereplicatedwithoff-the-shelf componentsat a costof about$140,000.A detailed description of instru-mentationandsoftwareisprovidedinNotesS5andS7.

Asagenomic-scalechallenge,wesoughtamicrobialgenomethatwasexpected,rela-tive to a reference sequence, to contain amodestnumberofbothexpectedandunex-pected differences. We selected a derivativeof E. coli MG1655, engineered for deficien-cies in tryptophan biosynthesis and evolvedfor ~200 generations under conditions ofsyntrophic symbiosis via coculture with atyrosine biosynthesis–deficient strain (23).Specificphenotypesemergedduringthelabo-ratoryevolution,leadingtotheexpectationofgenetic changes in addition to intentionallyengineereddifferences.

An in vitro mate-paired library was con-structedfromgenomicDNAderivedfromasinglecloneoftheevolvedTrp

vstrain.Tose-

quencethislibrary,weperformedsuccessiveinstrument runs with progressively higherbead densities. In an experiment ultimatelyyielding 30.1 Mb of sequence, 26 cyclesof sequencing were performed on an arraycontainingamplified,enrichedePCRbeads.At each cycle, data were acquired for fourwavelengthsat20×opticalmagnificationbyrastering across each of 516 fields of viewonthearray(Fig.1D).AdetaileddescriptionofthestructureofeachsequencingcycleisprovidedinNoteS6.Intotal,54,696images(14bit,1000×1000)werecollected.Cycle

�1

timesaveraged135minperbase(~90minforreactionsand~45minforimaging),foratotalof~60hoursperinstrumentrun.

Imageprocessingandbasecallingalgo-rithmsaredetailedinNoteS7.Inbrief,allimagestakenatagivenrasterpositionwerealigned. Two additional image sets wereacquired: brightfield images to robustlyidentifybeadlocations(Fig.2A)andfluo-rescentprimerimagestoidentifyamplifiedbeads.Ouralgorithmsdetected14million

objectswithinthesetofbrightfieldimages.Onthebasisofsize,fluorescence,andover-allsignalcoherenceoverthecourseofthese-quencingrun,wedetermined1.6milliontobewell-amplified,clonalbeads(~11%).Foreach cycle, mean intensities for amplifiedbeadswereextractedandnormalizedtoa4Dunitvector(Fig.2,BandC).TheEuclideandistance of the unit vector for a given rawbasecalltothemediancentroidofthenear-estcluster servesasanaturalmetricof the

We selected a derivative of E. coli MG1655,

engineered for deficiencies in tryptophan bio-

synthesis and evolved for È200 generations

under conditions of syntrophic symbiosis via

coculture with a tyrosine biosynthesis–deficient

strain (23). Specific phenotypes emerged during

the laboratory evolution, leading to the expec-

tation of genetic changes in addition to inten-

tionally engineered differences.

An in vitro mate-paired library was con-

structed from genomic DNA derived from a

single clone of the evolved Trpvj strain. To

sequence this library, we performed successive

instrument runs with progressively higher bead

densities. In an experiment ultimately yielding

30.1 Mb of sequence, 26 cycles of sequencing

were performed on an array containing ampli-

fied, enriched ePCR beads. At each cycle, data

were acquired for four wavelengths at 20

optical magnification by rastering across each

of 516 fields of view on the array (Fig. 1D). A

detailed description of the structure of each

sequencing cycle is provided in Note S6. In

total, 54,696 images (14 bit, 1000 1000)

were collected. Cycle times averaged 135 min

per base (È90 min for reactions and È45 min

for imaging), for a total of È60 hours per

instrument run.

Image processing and base calling algo-

rithms are detailed in Note S7. In brief, all

images taken at a given raster position were

aligned. Two additional image sets were ac-

quired: brightfield images to robustly identify

bead locations (Fig. 2A) and fluorescent primer

images to identify amplified beads. Our algo-

rithms detected 14 million objects within the

set of brightfield images. On the basis of size,

fluorescence, and overall signal coherence over

the course of the sequencing run, we deter-

mined 1.6 million to be well-amplified, clonal

beads (È11%). For each cycle, mean inten-

sities for amplified beads were extracted and

normalized to a 4D unit vector (Fig. 2, B and

C). The Euclidean distance of the unit vector

for a given raw base call to the median cen-

troid of the nearest cluster serves as a natural

metric of the quality of that call.

The reference genome consisted of the E.

coliMG1655 genome (GenBank accession code

U00096.2) appended with sequences corre-

sponding to the cat gene and the lambda Red

prophage, which had been engineered into the

sequenced strain to replace the trp and bio

operons, respectively. To systematically assess

our power to detect single-base substitutions,

we introduced a set of 100 random single-

nucleotide changes into the reference sequence

at randomly selected positions (Bmock SNCs[)

(Table 1).

An algorithm was developed to place the

discontinuous reads onto the reference sequence

(Note S7). The matching criteria required the

paired tags to be appropriately oriented and

located within 700 to 1200 bp of one anoth-

er, allowing for substitutions if exact matches

were not found. Of the 1.6 million reads, we

were able to confidently place È1.16 million

(È72%) to specific locations on the reference

genome, resulting in È30.1 million bases of

resequencing data at a median raw accuracy

of 99.7%. At this stage of the analysis, the

data were combined with reads from a pre-

vious instrument run that contributed an addi-

tionalÈ18.1 million bases of equivalent quality

(Fig. 2D). In this latter experiment, È1.8 mil-

lion reads were generated from È7.6 million

objects (È24%), of which È0.8 million were

confidently placed (È4