RESEARCH ARTICLE lessen the complexityof prenataldiagno-

sis for sickle cell anemia;they may also

be generally applicableto the diagnosis
of other genetic diseases and in the use
of DNA probes for infectious disease
Enzymatic Amplification of P-Globin diagnosis.
Sequenceamplificationby polymerase
Genomic Sequences and Restriction Site chainreaction.We use a two-step proce-
dure for determiningthe p-globingeno-
Analysisfor Diagnosisof Sickle Cell Anemia type of human genomic DNA samples.
First, a small portion of the p-globin
Randall K. Saiki, Stephen Scharf, Fred Faloona, Kary B. Mullis gene sequence spanningthe polymorphic
Dde I restrictionsite diagnosticof the [A
Glenn T. Horn, Henry A. Erlich, Norman Arnheim allele is amplified.Next, the presence or
absence of the Dde I restrictionsite in
the amplifiedDNA sampleis determined
by solution hybridizationwith an end-
Recent advancesin recombinantDNA This disease results from homozygosity labeled complementary oligomer fol-
technology have made possible the mo- of the sickle-cell allele (rS) at the 3- lowed by restrictionendonucleasediges-
lecularanalysisandprenataldiagnosisof globin gene locus. The S allele differs tion, electrophoresis,and autoradiogra-
several human genetic diseases. Fetal from the wild-type allele (3A) by substi- phy.
DNA obtainedby aminocentesisor cho- tution of an A in the wild-typeto a T at The p-globingene segmentwas ampli-
rionicvillus samplingcan be analyzedby the second position of the sixth codon of fied by the polymerase chain reaction
restrictionenzyme digestion,with subse- the p chaingene, resultingin the replace- (PCR)procedureof Mullis and Faloona
quent electrophoresis, Southern trans- ment of a glutamicacid by a valine in the (12) in which we used two 20-baseoligo-
fer, and specific hybridizationto cloned expressedprotein.For the prenataldiag- nucleotide primersthat flank the region
gene or oligonucleotide probes. With nosis of sickle cell anemia, DNA ob- to be amplified. One primer, PC04, is
complementaryto the (+)-strandandthe
other, PC03, is complementaryto the
Abstract. Two new methods were used to establish a rapid and highly sensitive (-)-strand (Fig. 1). The annealing of
prenatal diagnostic test for sickle cell anemia. The first involves the primer-mediated PC04 to the (+)-strand of denaturedge-
enzymatic amplification of specific 3-globin target sequences in genomic DNA, nomic DNA followed by extension with
resulting in the exponential increase (220,000 times) of target DNA copies. In the the Klenow fragmentof Escherichiacoli
second technique, the presence of the 3A and Ps alleles is determined by restriction DNA polymeraseI and deoxynucleotide
endonuclease digestion of an end-labeled oligonucleotide probe hybridized in triphosphatesresultsin the synthesisof a
solution to the amplified 3-globin sequences. The 3-globin genotype can be (-)-strand fragmentcontainingthe target
determined in less than 1 day on samples containing significantly less than I sequence. At the same time, a similar
microgram of genomic DNA. reaction occurs with PC03, creating a
new (+)-strand. Since these newly syn-
thesized DNA strands are themselves
polymorphic DNA markers linked ge- tainedby amniocentesisor chorionicvil- templatefor the PCR primers,repeated
netically to a specific disease locus, seg- lus sampling can be treated with a re- cycles of denaturation,primerannealing,
regation analysis must be carried out striction endonuclease (for example, and extension result in the exponential
with restriction fragment length poly- Dde I and Mst II) that recognizes a accumulationof the 110-basepairregion
morphisms(RFLP's) found to be infor- sequence alteredby the S mutation(8- definedby the primers.
mative by examiningDNA from family 11). This generates pA- and Ps-specific An example of the degree of specific
members(1, 2). restriction fragments that can be re- gene amplificationachieved by the PCR
Many of the hemoglobinopathies, solved by Southerntransferand hybrid- methodis shown in Fig. 2A. Samples of
however, can be detected by more direct ization with a 3-globinprobe. DNA (1 ag)were amplifiedfor 20 cycles
methods in which analysis of the fetus We have developed a procedure for anda fractionof each sample,equivalent
alone is sufficientfor diagnosis. For ex- the detection of the sickle cell mutation to 36 ng of the originalDNA, was sub-
ample, the diagnosis of hydrops fetalis that is very rapid and is at least two jected to alkalinegel electrophoresisand
(homozygous cx-thalassemia) can be ordersof magnitudemore sensitive than transferredto a nylon filter. The filter
madeby documentingthe absence of any standard Southern blotting. There are was then hybridizedwith a 32P-labeled
(x-globingenes by hybridizationwith an two specialfeaturesto this protocol. The 40-base oligonucleotide probe, RS06,
ao-globinprobe (3-5). Homozygosity for first is a method for amplifyingspecific which is complementaryto the target
certain P-thalassemiaalleles can be de- p-globinDNA sequences with the use of sequence (Fig. 1A) but not to the PCR
termined in Southern transfer experi- oligonucleotideprimers and DNA poly- primers.The results, aftera 2-hourauto-
ments by using oligonucleotide probes merase (12). The second is the analysis radiographicexposure, show that a frag-
that form stable duplexes with the nor- of the p-globingenotype by solution hy- ment hybridizingwith the RS06 probe
mal p-globin gene sequence but form bridizationof the amplifiedDNA with a
The authors are in the Departmentof Human
unstable hybrids with specific mutants specific oligonucleotideprobe and sub- Genetics, Cetus Corporation, 1400 Fifty-Third
(6, 7). sequent digestion with a restriction en- Street, Emeryville, California94608. The present
donuclease (13). These two techniques addressfor N.A. is Departmentof Biological Sci-
Sickle cell anemia can also be diag- ences, Universityof SouthernCalifornia,Los Ange-
nosed by direct analysis of fetal DNA. increase the speed and sensitivity, and les 90089-0371.

1350 SCIENCE,VOL. 230

migratesat the position expected of the
amplified target DNA segment (110 RS06:
bases) (lanes 1 and 2). No hybridization
with the RS06probe could be detected in
unamplifiedDNA (lane 4). When PCR PC04:
amplification was performed on a DNA
sample derived from an individualwith B
hereditarypersistence of fetal hemoglo-
bin in which both the 8- and 3-globin
genes are deleted (14), againno 110-base
fragmentwas detected (lane 3). To esti-
matethe yield and efficiencyof 20 cycles
of PCR amplification, we prepared a
Southernblot that contained36 ng of an
amplifiedgenomic DNA sample and a
dilution series consisting of various
amounts of cloned P-globin sequence. Fig. 1. Sequenceof
The efficiency was calculated according P-globinregion. (A) The primerPC03is complementaryto the (-)-strand and the primertarget
to the formula: (1 + X)" = Y, where X is
the mean efficiency per cycle, n is the
number of PCR cycles, and Y is the
extent of amplification (yield) after n
cycles (for example, a 200,000-fold in-
crease after 20 cycles). The amounts of
cloned ,-globin sequences used in this
experimentwere calculatedto represent
efficienciesof 70 to 100 percent.
The reconstructions were prepared
by digesting the P-globin plasmid,
pBR328:: pA, with the restriction en-
zymes Hae III and Mae I. Both of these
cleave the
enzymes ,-globin gene within
or very near to the 20 base regions that
hybridizeto the PCRprimers,generating
a 103-base pair (bp) fragment that is
almost identicalin size and composition
to the 110-bp segment created by PCR
amplification.After hybridizationwith
the RS06probeandautoradiography,the
amplifiedgenomic samplewas compared
with the known standards,and the result
indicatedan overallefficiencyof approx-
imately 85 percent (Fig. 2B), represent-
ing an amplificationof about 220,000 added, and the incubationwas allowed to proceed for an additional2 minutes at 30?C.This
times (1.8520).
Distinguishing the 1A and Ps alleles by
the oligomer restriction method. We have
previouslydescribeda rapidsolutionhy-
bridization method that can indicate
whether a genomic DNA sample con-
tains a specific restriction enzyme site
at, in principle, any chromosomalloca-
tion (13). This method, called oligomer
restriction (OR), involves the stringent
hybridizationof a 32Pend-labeledoligo-
nucleotideprobe to the specific segment
of the denaturedgenomic DNA which
spansthe targetrestrictionsite. The abil-
ity of a mismatchwithin the restriction
site to prevent cleavage of the duplex
formedbetween the probe and the target
genomicsequence is the basis for detect-
ing allelic variants. The presence of the
restrictionsite in the target DNA is re-
vealed by the appearanceof a specific
20 DECEMBER 1985
A
RS1O:
5' CTGACTCCTGAGGAGAAGTCTGCCGTTACTGCCCTGTGGG
3' GACAGAGGTCACCTCTTCAGACGGCAATGACGGGACACCC
3'
5
PC03: 5' ACACAACTGTGTTCACTAGC 3'
3' CCACTTGCACCTACTTCAAC 5'
5' TCTGACACAACTGTGTTCACTAGCAACCTCAAACAGACACCATGGT
.....
5' PC03 3' .....*.
....??** H D
...?' ) ?__---------- + t t
?.***.** GCACCTGACTCCTGAGGAGAAGTCTGCCGTTACTGCCCTGTGGGGCAAGGTGAACGTGGATGAAGTTGGTGG 3'
5' RS06 3' 3' PC04 5'
syntheticoligonucleotideprimersand probeand theirrelationto the
PC04
is complementaryto the (+)-strandof the P-globingene. The probeRS06is complementaryto
the (-)-strand of the wild-type(pA) sequence of P-globin.RS10is the "blockingoligomer",an
oligomer complementaryto the RS06 probe except for three nucleotides, indicated by the
downwardarrows. It is added before enzyme digestion to the OR reaction to anneal to the
excess RS06 oligomerand preventnonspecificcleavageproductsdue to hybridizationof RS06
to nontargetDNA (13). Because of the mismatcheswithinthe Dde I andHinfI restrictionsites,
the RS06/RS1O duplexis not cleaved by Dde I andHinfI digestion.(B) The relationbetweenthe
primers,the probe, and the target 3-globinsequence. The upwardarrowindicatesthe p-globin
initiationcodon. The downwardarrowsindicatenucleotidedifferencesbetweenP- and8-globin.
The polymorphicDde I site (CTCAG)is representedby a singlehorizontaldashedline (D), and
the invariantHinf I (GACTC)site is representedby double horizontaldashed lines (H).
1 2 34 1 3 4 5 6 7 8 pFig.2. Southernanalysis of PCR amplifiedgenomic
DNA with the RS06 probe. (A) Samples (1 Ig) of
genomic DNA were dispensed in microcentrifuge
tubes and adjustedto 100Rl in a buffercontaining10
mM tris, pH 7.5, 50 mM NaCl, 10 mM MgC12,1.5
mM deoxynucleotidetrisphosphate[(dNTP)each of
all four was used], 1 FLMPC03, and 1 pM PC04.
After heatingfor 5 minutesat 950C(to denaturethe
g---- ;
-----i-genomic DNA), the tubes were centrifugedfor 10
--f;- -; - - - - seconds in a microcentrifugeto remove the conden-
i--:X : :-XX:
;-X- " ; sation. The sampleswere immediatelytransferredto
- E -: 0i- a 30?Cheat blockfor 2 minutesto allowthe PC03and
PC04primersto annealto theirtargetsequences. At
the end of this period,2 pl of the Klenowfragmentof
E. coli DNA polymeraseI (Biolabs,0.5 unit/plin 10
mM tris, pH 7.5, 50 mM NaCl, 10 mM MgC12)was
cycle-denaturation, centrifugation,hybridization, and extension-was repeated 19 more
times, except that subsequentdenaturationswere done for 2 instead of 5 minutes. (The final
volumeafter20 cycles was 140 p1.)Thirty-sixnanogramsof the amplifiedgenomicDNA (5 p1)
were applied to a 4 percent Nusieve (FMC) alkaline agarose minigel and subjected to
electrophoresis(50 V), for 2 hours until the bromcresolgreen dye front reached4 cm. After
neutralizationand transferto Genetransnylon membrane(Plasco), the filter was "prehybri-
dized" in 10 ml 3x SSPE ( x SSPE is 0.18M NaCl, 10 mM NaH2PO4,1 mM EDTA,pH 7.4),
5x DET (lx DET is 0.02 percent each polyvinylpyrrolidone,Ficoll, and bovine serum
albumin;0.2 mMtris, 0.2 mMEDTA,pH 8.0), 0.5 percentSDS, and30 percentformamidefor 4
hoursat 42?C.Hybridizationwith 1.0 pmol of phosphorylated(with [y-32P]ATP) RS06(-5 ,iCi/
pmol)in 10 ml of the same bufferwas carriedout for 18 hours at 42?C.The filterwas washed
twice in 2x SSPE, 0.1 percent sodium dodecyl sulfate (SDS) at room temperaturefor 30
minutes,and autoradiographed at -70?C for 2 hourswith a singleintensificationscreen. (Lanes
1 to 3) DNA's isolated from the cell lines Molt4, SCO1,and GM2064,respectively. Molt4 is
homozygousfor the normal, wild-typeallele of 3-globin(pApA), SC-1 is homozygousfor the
sickle cell allele (Pps), and GM2064is a cell line in which the f3-and 8-globingenes have been
deleted (AA)(13). (Lane 4) Contains 36 ng of Molt4 DNA that was not PCR amplified.The
horizontalarrowindicatesthe position of a 114-basemarkerfragmentobtainedby digestionof
pBR328with Nar I. (B) Thirty-sixnanogramsof 20-cycleamplifiedMolt4DNA (see above)was
loadedonto a Nusieve gel alongwith measuredamountsof Hae III-MaeI digestedpBR328:: A
(13) calculatedto representthe molarincreasein P-globintargetsequencesat PCRefficiencies
of 70, 75, 80, 85, 90, 95, and 100 percent (lanes 2 to 8, respectively).DNA was transferredto
Genetransand hybridizedwith the labeled RS06 probe as describedabove. (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::pA,respectively
(20).
1351
labeled fragmentgeneratedby cleavage
of the probe.
For the diagnosisof sickle cell anemia,
the probe was designed to be comple-
mentaryto a region of the p-globingene
locus surroundingthe sixth codon. In the
fA allele, the nucleotide(nt) sequence at
this position contains a Dde I restriction
site, but due to the single base mutation,
this site is absent in the PS allele. Our
strategy for generating specific probe
cleavage products for each allele is
shown in Fig. 3. It is based on the
presence of an invariantHinf I restric-
tion enzyme site immediatelyadjacentto
the polymorphicDde I restriction site.
Resolutionof the labeledoligomercleav-
age productsproducedby sequentialdi-
gestion with Dde I and Hinf I allows us
to distinguishbetween the two alleles. In
an individualhomozygous for the wild-
type P-globinallele AA, Dde I digestion
will produce a labeled octamer (8 nt)
from the probe. Because of its short
length, the 8-nt cleavage product will
dissociatefrom the genomic targetDNA
and the subsequentdigestionwith Hinf I
has no effect. In the case of SS homozy-
gotes, however, Dde I digestiondoes not
cleave the probe since a base pair mis-
matchexists in the recognitionsequence
formed between the probe and target
DNA. The invariantHinf I site will then
be cleaved during Hinf I digestion, re-
leasing a labeled trimer(3 nt). In an AS
heterozygote, both a trimer and an oc-
tamerwould be detected. The resolution
of the intact 40-base probe, the 8-nt and
the 3-ntcleavageproductsis achievedby
polyacrylamidegel electrophoresis. Ex-
periments testing the sequential diges-
tion strategy with plasmids carryingthe
3A and Ps alleles show that, in each case,
the expected probe cleavage products
were produced(Fig. 4).
Analysis of genomic DNA samples by
PCR and OR. Eleven DNA samples de-
rived from lymphoblastoidcell lines or
white blood cells were analyzedfor their
P-globingenotype by standardSouthern
blotting and hybridizationof the Mst II
RFLP (10), identifyingthe genotypes of
the samplesas eitherAA, AS, or SS. Six
of these samples(andone additionalone)
were then amplifiedby PCRfor 20 cycles
starting with 1 ,Ig of DNA each. An
aliquot of the amplified DNA sample
(one-fourteenthof the original1-Kgsam-
ple) was hybridizedto the RS06 probe
and digestedwith Dde I and then Hinf I.
A portion (one-tenth) of this oligomer
restriction reaction was analyzed on a
polyacrylamidegel to resolve the cleav-
age products, and the results obtained
after 6 hours of autoradiographyare
shown Fig. 5. The high sensitivity
achieved with the PCR and OR method
is demonstratedby the strength of the
autoradiographic signal derived from
only 1/140of the original1-jIgsample (7
ng). Each sample determinedto be AA
by RFLP analysis showed a strong 8-nt
fragment while those typed as SS
showed a strong3-nt fragment.Analysis
of the known AS samplesrevealed both
cleavage products.
In the analysis of the AA samples, a
faint3-ntcould be detectedin additionto
the primary8-nt signal. The reasons for
this band remain unclear, although in-
completeDde I cleavageor the occasion-
al failureof the 8-nt fragmentto disasso-
ciate from the targetDNA may contrib-
ute to the nonspecific3-ntproductgener-
ated by Hinf I digestion. In the analysis
of the SS samples, a very faint 8-ntband
was also observed in addition to the
expected 3-nt signal. We have deter-
mined that the background8-nt product
detected in SS samplescan be attributed
to the 8-globin gene, which is highly
homologousto p-globin.The nucleotide
sequence of the two P-globin primers
used for amplificationis shown in Fig. 1.
The downwardpointingarrows indicate
the differences between the 3- and 8-
globin genes. We hypothesized that the
faint 8-nt signalobserved in the SS sam-
ples was due to some amplificationof the
8-globingene by these primers and the
subsequent cross-hybridizationof the
amplified 8 sequences with the RS06
probe used in the OR procedure.8-Glo-
bin has the same Dde I site as normal 3-
globin, and the duplex formed between
an amplified 8 gene segment and the
RS06 probe would be expected to yield
an 8-ntfragmenton Dde I digestioneven

A (pA) B (p S)
GACTCCTGAG CACTCCTGTG
CTGAGGACTC CTGAGGACAC

Denature and anneal Denature and anneal

with probe with probe

*-GACTCCTGAG (RS06) *-GACTCCTGAG (RS06)

CTGAGGACTC CTGAGGACAC

1
(8 nt) *-GACTCC
4 TGAG
Digest with Dde I

* GACTCCTGAG
Digest with Dde I

CTGAGGACT CTGAGGACAC

Digest with Hinfl Digest with Hinfl

*-GACTCC
1 TGAG (3 nt) *-G ACTCCTGAG
r CTGA GGACAC

Fig. 3. Schematicdiagramof oligomerrestrictionby sequentialdigestionto identifyA^- and Ps-specificcleavageproducts.The DNA sequences

shownaretheregionsof the P-globingenomicDNAandtheRS06hybridization
probecontaining HinfI site (GANTC,whereN
theinvariant
andthepolymorphic
anynucleotide)
represents DNAsequencesarerepresented
DdeI site(CTNAG).Theremaining lines.
as solidhorizontal
The asteriskindicatesthe positionof the radioactive32Plabel attachedto the 5'-endof the RS06probewith polynucleotidekinase. (A) Outlineof
the procedure and expected results when RS06 anneals to the normal P-globingene (rA). After denaturationof the genomic DNA and
hybridizationof the labeled RS06 probe to the complementarytargetsequence in the pA gene, digestionof the probe-targethybridwith Dde I
causes the release of a labeled (8-nt)cleavage product.Because of the relativelystringentconditionsduringDde I digestion, the 8-nt cleavage
productdissociatesfromthe genomicDNA andthe subsequentdigestionwith HinfI has no effect. (B) Outlineof Dde I and Hinf I digestionafter
hybridizationof the RS06 probe to the sickle cell allele (pS). As a consequenceof the Ps mutation,the probe-targethybridcontains an A-A
mismatchwithinthe Dde I site and is not cleaved by the Dde I endonuclease.The HinfI site, however,remainsintactand digestionwith thaten-
zyme generatesa labeled3-ntproduct.Thus, the presenceof the pA alleleis revealedby the releaseof a labeled8-ntfragment,whilethe presence
of ps is indicatedby a labeled 3-nt fragment.
1352 SCIENCE, VOL. 230
 though there are sequence differences
(four mismatch out of 40 bases) between
RS06 and 8-globin. It is likely that 8-
globin sequences may be amplified to
some extent and detected weakly with
the RS06 probe in all DNA samples, but
that its contribution to the total signal is
very small and detectable only when the
sample is SS and no 8-nt fragment from
the P-globin gene is expected. We tested
this hypothesis by treating an SS DNA
sample before amplification with the en-
zyme Mbo I. Since there is a recognition
site for this enzyme in the target DNA of
the 8- but not the P-globin gene, cleavage
of the 8 gene between the regions that
hybridize to the PCR primers should
prevent its subsequent amplification (but
not of P-globin). Our results showed that
an SS DNA sample, first digested with
Mbo I, gave only the 3-nt product but not
the 8-nt product, this is consistent with
the hypothesis of 8-globin amplification.
Effect of PCR cycle number on detec-
tion threshold. The strength of the auto-
radiograph signal detected by OR as a
function of PCR cycle number and auto-
radiographic exposure was examined.
The signal intensity after 20 cycles is at
least 20 times as strong as that for 15
cycles and the determination of the p-
globin genotype can be made with an
autoradiographic exposure for only 2
hours (Fig. 6). The observed increase of
>20-fold is consistent with our estimates
of 85 percent efficiency per cycle, calcu-
lated from the data in Fig. 2B
(1.855 = 21.7). Coupled with the time
that it takes to actually carry out the
PCR and OR procedures, a 20-cycle PCR
allows a diagnosis to be made in less than
10 hours with a DNA sample of 1 plg.
Since all of the previous PCR experi-
ments were done with 1 lg of genomic
DNA, we explored the effect of using
significantly smaller amounts of DNA as
template for PCR amplification. The re-
sults obtained with 20 cycles of PCR
amplification on 500, 100, 20, and 4 ng of
DNA from an AS individual are shown in
Fig. 7. After analysis of 1/40 of each
sample by the OR procedure and a 20-
hour autoradiographic exposure, the 3-
globin genotype could be easily deter-
mined on DNA samples of 20 ng or about
100 times less than is needed for a typical
Southern transfer and hybridization ex-
periment. In this experiment, only a
small fraction (1/40) of the starting mate-
rial was placed on the gel; therefore it
should be possible to analyze samples of
less than 20 ng of genomic DNA (20 ng is
equivalent to approximately 6000 hap-
loid genomes) if a larger proportion of
the material was utilized in the OR and
gel electrophoresis steps.
20 DECEMBER 1985
Diagnostic applications of the PCR-OR Because this method includes a liquid
system. When currently available meth- hybridization protocol and involves the
ods are used, the completion of a prena- serial addition of reagents to a single
tal diagnosis for sickle cell anemia takes tube, it is simpler to perform than the
a period of several days after the DNA is standard Southern transfer and hybrid-
isolated. With 1 lg of genomic DNA, the ization procedure. Prior to electrophore-
3-globin genotype can be determined by sis, all of the reactions can be done in
the PCR-OR method in less than 10 two small microcentrifuge tubes and
hours; 20 cycles of amplification requires could readily be automated.
about 2 hours (each full cycle takes 6 to 7 The sensitivity, as well as the speed
minutes in our protocol), the oligomer and simplicity, of this procedure is also
restriction procedure involving liquid hy- important for clinical applications.
bridization and enzyme digestions re- Twenty nanograms of starting material
quire an additional 2 hours, and the can provide an easily detectable result in
electrophoresis takes about an hour. an overnight autoradiographic exposure.
Autoradiographic exposure for 4 hours is This sensitivity makes the PCR-OR
sufficient to generate a strong signal. method particularly valuable in cases
2 3
1 Fig. 4. Demonstrationof OR sequential digestion with
cloned P-globin genes. The sequential digestion strategy
was demonstratedby annealingthe RS06 probe to the P-
globin plasmids pBR328::A and pBR328::p (13). The
methods were similar to those described (13). Cloned P-
globin DNA (45 ng; 0.01 pmol)was placedin a microcentri-
fuge tube, adjustedto 30 cl1with TE buffer(10 mM tris, 0.1
mM EDTA, pH 8.0), overlaid with 0.1 ml of mineraloil.
The DNA was denaturedby heatingfor 5 to 10 minutesat
-8 nt 950C.Ten microlitersof 0.6M NaCIcontaining0.02 pmol of
phosphorylated(with[y-32P]ATP) RS06probeoligomer(-5
tCi/pmol) was addedand annealedfor 60 minutesat 56?C.
UnlabeledRS10O blockingoligomer(4 R1;200 pmol/ml)(Fig.
n3 a"t 1) (13)was then added,andthe hybridizationwas continued
for 5 to 10 minutes.Next, 5 R.1of 100mM MgCI2and 1 p.1of
Dde I (Biolabs, 10 units) was added and incubatedfor 20
minutesat 56?C;1 pL1 of Hinf I (Biolabs, 10units)was added
and digestion was continued for 20 minutes at the same
temperature.The reaction was terminatedby the addition
of 4 pIlof 100mM EDTA and 6 p.1of trackingdye to a final
volume of 61 p.1;a portion (8 p1) (6 ng, 0.0013 pmol) was
applied to a 0.75-mm thick, 30 percent polyacrylamide
minigel(19 acrylamide:1 bis) in a Hoefer SE200apparatusand subjectedto electrophoresis(300
V) for 1 hour until the bromphenolblue dye front reached 3 cm. The top 1.5 cm of the gel,
containingintactRS06, was cut off anddiscarded.The remaininggel was autoradiographed with
a single intensificationscreen for 18 hours at -700C. (Lane 1) six nanogramsof pBR328::PA^;
(lane 2) 3 ng of pBR28::pA plus 3 ng of pBR328:: s; and (lane 3) 6 ng of pBR328:: s.
1 2 3 4
5 6 7 8 9 10 11 15 cycles 20 cycles
1 2 3 4 5 6
-8 nt
-
-3 nt
-8 nt
: ti00 2f :
~::,:., iSfti:
~vS -3 nt
Fig. 5 (left). Determinationof the p-globingenotype in humangenomic DNA with PCR-OR.
Samples(1 p.g)of humangenomic DNA were amplifiedfor 20 cycles (as describedin Fig. 2A).
The amplifiedDNA's (71 ng) were hybridizedto the RS06probeand seriallydigestedwith Dde I
and Hinf I (as described in Fig. 4). Each sample (6 p.l) was analyzed by 30 percent
polyacrylamidegel electrophoresis and autoradiographedfor 6 hours at -70?C with one
intensificationscreen. Each lane contains 7 ng of genomic DNA. (Lane 1) UnamplifiedMolt4
DNA (negativecontrol);(lane 2) amplifiedMolt4(A1pA); (lane 3) SC-1(P3ps);(lane4) GM2064
(AA); (lanes 5 to 11) clinical samples CHI (AA), CH2 (pp), CH3 (1S3 ), CH4 (PSPs), CH7
(pApS),CH8 (ps3s), and CH12 (^pApS),respectively. Fig. 6 (right).Effect of cycle number
on signal strength.GenomicDNA (1 p.g)from the clinical samplesCH2 (pApA), CH12 (pApS),
and CH5 (pSps) were amplifedfor 15 and 20 cycles and equivalentamountsof genomic DNA
(80 ng)were analyzedby oligomerrestriction.(Lanes 1 to 3) DNA (20 ng)fromCH2, CH12,and
CH5, respectively, amplifiedfor 15 cycles; (lanes 4 to 6) DNA (20 ng) from CH2, CH12, and
CH5, respectively, amplifiedfor 20 cycles. Autoradiographicexposure was for 2.5 hours at
-70?C with one intensificationscreen.
1353
where poor DNA yields are obtained
fromprenatalsamples. In addition,DNA
samples of poor quality (very low aver-
age molecularweight) can give excellent
results in the PCR-ORprotocol.
The PCR methodis likely to be gener-
ally applicablefor specific gene amplifi-
cation since a fragmentencoding a por-
tion of the HLA-DQoalocus has recently
been amplifiedwith this procedure(15).
We have carried out PCR amplification
on a 110-bpP-globin sequence with an
overall efficiency per cycle of about 85
percent. We have also amplifiedlonger
p-globin fragments (up to 267 bp), but
the yield was lower under our standard
conditions. Efficient amplificationof a
267-bpfragmentrequiredsome variation
in the PCR procedure. In principle, in-
creasing the number of PCR cycles
should yield even greater amplification
than that reported here (-220,000-fold
after 20 cycles).
Ourmethodfor the diagnosisof sickle
cell anemiainvolves the coupling of the
PCR procedure with that of oligomer
restriction.It was designedto distinguish
between two alleles that differby a poly-
morphic restriction site. The PCR-OR
method is applicableas well to the diag-
nosis of other diseases where the lesion
directly affects a restrictionenzyme site
or where the polymorphic site is in
strong linkage disequilibriumwith the
disease causing locus. If the polymor-
phism is in linkage equilibriumwith the
disease, PCR-ORrequiresfamily studies
to follow the inheritanceof the disease
locus.
In the case of the 3-globinlocus, the
presence of the invariantHinf I restric-
tion site adjacent to the polymorphic
Dde I site allows a sequential digestion
procedureto identifyboth the 1A and S
1354
1 2 3 4 5 number of other procedures including
those involving the hybridization of
small labeled oligomerswhich will form
stableduplexesonly if perfectlymatched
-8 nt (6, 7, 17, 18) and the recently reported
method based on the electrophoretic
shifts of duplexes with base pair mis-
-3 nt matches (19). The ability of the PCR
procedureto amplifya target DNA seg-
ment in genomic DNA raises the possi-
bilitythat its use may extendbeyondthat
Fig. 7. Detection threshold for IPCR-OR. of prenatal diagnosis to other areas of
Fivefold serial dilutions of genom
(500, 100, 20, and 4 ng) from the clinical molecularbiology.
sample CH12 (pApS) were amplifeid by 20
cycles of PCR and one-tenth each reaction Referencesand Notes
(50, 10, 2, and 0.4 ng) was analyzecI by OR. 1. Y. W. Kan and A. M. Dozy, Proc. Natl. Acad.
The gel continued(lane 1)genomicD NA (12.5 U.S.A. 75, 5631 (1978).
ng);(lane2) 2.5 ng; (lane 3) 0.5 ng; (laane 4) 0.1 2. K. E. Davies andD. P. Ellis, Biochem.J. 226, 1
ng; (lane 5) 12.5 ng genomic DNA from the (1985).
globin deletion cell line GM2064.Aiutoradio- 3. Y. W. Kan, M. S. Golbus,A. M. Dozy, N. Eng.
J. Med. 295, 1165(1976).
graphicexposure was for 20 hours at -700C 4. A. M. Dozy et al., J. Am. Med.Assoc. 241, 1610
with an intensificationscreen. (1979).
5. E. M. Rubinand Y. W. Kan, Lancet 1985-I,75
(1985).
6. M. Pirastu et al., N. Engl. J. Med. 309, 284
(1983).
alleles. In principle, this approaich does 7. S. H. Orkin,A. F. Markham,H. H. Kazazian,
Jr., J. Clin. Invest. 71, 775 (1983).
not requirethat the two sites be immedi- 8. R. F. Geever et al., Proc. Natl. Acad. Sci.
ately adjacentbut only that the (cleavage U.S.A. 78, 5081 (1981).
9. J. C. Changand Y. W. Kan, N. Eng. J. Med.
product generated by digestiorn at the 307, 30 (1982).
polymorphicsite dissociate from the tar- 10. S. H. Orkinet al., ibid., p. 32.
11. J. T. Wilson et al., Proc. Natl. Acad. Sci.
get to prevent cutting at the invariant U.S.A. 79, 3628 (1982).
F. Faloonaand K. Mullis,in preparation.
site. Since the restrictionenzyn ie diges-- 12.13. R. Saiki,N. Arnheim,H. Erlich,Biotechnology
tion conditionsused here are fai]rly strin- 3, 1008(1985).
tt 14. D. Tuan et al., Proc. Natl. Acad. Sci. U.S.A.
gent for hybridization, we estirlate that 80, 6937 (1983).
the polymorphic and invariant sites 16. 15. s. Scharf, unpublished data.
R. Saiki et al., unpublisheddata.
could perhapsbe separatedby as much 17. B. J. Conner et al., Proc. Natl. Acad. Sci.
as 20 bp. U.S.A. 80, 278 (1983).
18. V. J. Kidd, R. B. Wallace,K. Itakura,S. L. C.
The applicationof the PCR method to Woo, Nature (London)304, 230 (1983).
ily 19. R. M. Myers, N. Lumelsky,L. S. Lerman,T.
prenataldiagnosis does not nec Maniatis,ibid. 313, 495 (1985).
dependon a polymorphicrestric:tion site 20. Calculation of the amountsof pBR328: A need-
ed as standardsto estimatePCR efficiencywas
or on the use of radioactive priobes.-be Iny
In done in the followingway. If we assume that a
fact, the significantamplificatioin of tar- humanhaploidgenomesize is 3 x 109 bp, 36 ng
of DNA is equivalentto 1.8 x 10-8 pmol. The
get sequences achieved by tlhe PCR extent of amplificationafter 20 cycles at, for
methodallows the use of nonisol toMicall example, 85 percentefficiencyis obtainedwith
(1.8 x 10-8)(1.852?)or 4.0 x 0-3 pmolof plas-
labeled probes (16). Amplifiedt;argetse- mid DNA (Fig. 2B, lane 5).
quences could also be analyze ed by a 20 September1985;accepted 15 November 1985
SCIENCE,VOL. 230