A Novel PCR Technique UsingAlu-Specific Primers to Identify Unknown Flanking Sequences from the Human Genome

A Novel PCR Technique UsingAlu-Specific Primers to Identify Unknown Flanking Sequences from the Human Genome
of 6
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  GENOMICS  29,  403–408 (1995) A Novel PCR Technique Using  Alu  -Specific Primers to IdentifyUnknown Flanking Sequences from the Human Genome M ASAHITO  M INAMI , K ARINE  P OUSSIN , C HRISTIAN  B RE´CHOT ,  AND  P ATRIZIA  P ATERLINI 1 INSERM U-370, CHU Necker, Paris, France  Received March 29, 1995; accepted July 14, 1995  et al.,  1991; Williams  et al.,  1992; Screaton  et al.,  1993; Therapidandreproducibleidentificationofnewcel-  Sørensen  et al.,  1993; Malo  et al.,  1994). However, the lular DNA sequences adjacent to known sequences is  number of these methods and modifications indicates difficult to achieve with the currently available proce-  that there is no widely used protocol. Ligation steps in dures. Here we describe a novel approach based on the former and difficulties in optimizing the annealing  the polymerase chain reaction (PCR) using a primer conditions in the latter limit amplification efficiency. specific to the known sequence and another directed  After identification of HBV DNA integrations into to a human  Alu  repeat. To avoid undesirable amplifi- the host genome in hepatoma cell lines and tumor tis- cations between  Alu  sequences, primers are con- sues of HCC (Bre´chot  et al.,  1980; Edman  et al.,  1980), structed withdUTPs anddestroyed byuracil DNAgly- manystudieswereundertakentoexploretheoncogenic cosylase treatment after 10 initial cycles of amplifica- potential of integrated HBV. Among these, some inves- tion. Only desirable fragments are then further tigators have identified HBV integrants that interrupt amplified with specific primers to the known region human genes involved in cellular proliferation, causing  and to a tag sequence introduced in the  Alu -specific modified expression of these genes (Dejean  et al.,  1986; primer.Usingthisprotocol,wehavesuccessfullyiden- de The´  et al.,  1987; Wang   et al.,  1990, 1992; Zhang   et tified cellular sequences flanking integrated hepatitis al.,  1992; Graef   et al.,  1994). Thus, the study of HBV  B virus DNA from the human genome of three hepa- integration sites might be viewed as a tool to identify toma tissues. The method enables a direct specific am- new cellular genes potentially involved in cell differen- plification without any ligation or nonspecific anneal-ingstepsasrequiredbypreviousPCR-basedprotocols.  tiation and proliferation. Actually, only a few tumors This rapid and straightforward approach will be a  have been analyzed in detail for the detection of cellu- powerful tool for the study of viral integration sites,  lar genes adjacent to HBV integrants (Schlu¨ter  et al., but is also widely applicable to other studies of the  1994). Therefore, there is a need for a rapid and effi- human genome.    1995 Academic Press, Inc. cient screening procedure that is applicable to a largenumberofhumanHCCs.Infact,thepreviousapproachof screening genomic libraries from each patient for INTRODUCTION viral–host junctions has several limitations. It is labo-rious and time-consuming, particularly in HBV-in-Polymerase chain reaction (PCR) seems to be suit-fected individuals because coexisting nonintegratedable for analyzing the human genome and for overcom-free HBV sequences can disturb identification of trueing the problem of limited sample size. Although con-integrants. In addition, the requirement of a large tu-ventional PCR requires some knowledge of the targetmor volume makes it difficult to study small cancers,sequence, a number of methods of amplifying unknownprecancerous lesions, and specimens obtained by nee-flanking sequences are currently available. These candle biopsy. Amplification from coexisting free viral se-be broadly classified into two categories: one based onquences is also inevitable with available methods toligation of known sequences to the unknown regionanalyze unknown flanking sequences and competes(CollinsandWeissman,1984;Triglia  etal., 1988;Silverwith desirable products, impairing the amplificationandKeerikatte,1989;Jones andWinistorfer,1992;Du-efficiency and complicating the screening.mas  et al.,  1991; Bloomquist  et al.,  1992; Troutt  et al., To overcome these problems we have developed a1992; Arnold and Hodgson, 1991; Iwahana  et al.,  1994)novel approach to identify unknown flanking se-and the other based on the use of nonspecific annealing quences. This is an application of   Alu -PCR (Nelson  et of primers(Parks  etal., 1991; Wang   etal., 1991; Parker al.,  1989, 1991) to the whole human genome, a tech-nique first developed to amplify human genomes in the 1 To whom correspondence should be addressed at INSERM U- background of nonhuman genomes such as a lambda 370, CHU Necker, 156, rue de Vaugirard, 75730, Paris–Cedex 15,France. Telephone: 1 40615641. Fax: 1 40615581.  phage, yeast cells, and human–rodent hybrid cells. 403 0888-7543/95 $12.00Copyright    1995 by Academic Press, Inc. All rights of reproduction in any form reserved.  MINAMI ET AL. 404 Sequence of the 3   Alu  primer (A3) is the same as that used by Since the method utilizes a primer in an  Alu -repeated Chumakov  et al.  (1992). sequence, which is interspersed in the human genome  PCR using Alu primers.  The first 10 cycles of amplification were at the mean interval of about 4 kb (9  1  10 5 copies carried out in a final volume of 50  m l containing 25 m  M   Tris (pH per the haploid genome), the procedure is focused on 8.9), 40 m  M   potassium acetate, 2.5 m  M   magnesium chloride, 4% avoiding amplifications between  Alu  sequences. In the  glycerol, 200  m  M   each dNTP, 10 pmol of   Alu  primer, 100 pmol of HBVprimer,1unitof  Taq polymerase(LifeTechnologies,Inc.,Gaith- present report we show the validity of our method both ersburg, MD), and 0.04 unit of Vent exo / polymerase (New England fordetectionofHBV–cellularDNAjunctionsand,more Biolabs, Inc., Beverly, MA) in a Perkin–Elmer thermal cycler 480 generally, for studies on the human genome. (Emeryville, CA). A hot start technique was used by adding magne-sium chloride after the temperature reached at 80  C. Cycling condi- MATERIALS AND METHODS  tions consisted of denaturation for 30 s at 94  C, annealing for 30 sat 59  C, and extension for 3 min at 70  C with the initial denaturation  DNApreparation.  Normal liver tissues were obtained at autopsy. for 1 min at 94  C. One unit of uracil DNA glycosylase (UDG) (LifeLiver tumor tissues were obtained surgically from three patients Technologies, Inc.) was then added to each tube, followed by incuba-positive for serum hepatitis B surface antigen (HBsAg). DNA was tion for 30 min at 37  C. After heating for 10 min at 94  C to breakextracted by proteinase K digestion and phenol extraction as pre- DNA strands at apurinic dUTP sites, 10 pmol of each primer for theviouslydescribed(Paterlini  etal., 1990).Weused,asanHBVintegra- next amplification was added. The touchdown PCR technique (Dontion model, clone 20BB (Wang   et al.,  1992) of an 18-kb human geno-  et al.,  1991) was employed for this amplification. Denaturation stepsmic DNA fragment that contains 3 kb of HBV sequence integrated were at 94  C for 30 s, and extension steps were at 70  C for 3 min.inhumancyclinA geneandan  Alu repeatabout 1600bpdownstream Annealing steps started at 65  C for 30 s, and the temperature wasof the viral–host junction. decreased by 1  C every second cycle to 55  C, at which temperature20 cycles were performed with the final extension 8 min at 72  C.  Primers.  For primers used in the first 10 cycles of amplification,Thus, in total 40 cycles were carried out. Products were separateddTTPs were replaced for dUTPs. Primers for  Alu  repeats were de-on 1% agarose gels and were capillary transferred to Gene Screensigned to hybridize to either the 5   (A5, 5   CAGUGCCAAGUGUUU-Plus nylon membranes (DuPont–NEN, Washington, DC) followedGCUGACGCCAAAGUGCUGGGAUUACAG 3  ) or 3   (A3, 5   AGU-by hybridization with the appropriate probes. If necessary, 1  m l of GCCAAGUGUUUGCUGACGACUGCACUCCAGCCUGGGCGACtheproduct wassubjected tonestedor heminestedPCR withinternal3  ) end of the  Alu  sequence and to have a Tag sequence at the 5  primers to obtain discrete bands. These conditions were optimizedend (underlined nucleotides). Highly  Alu -conserved residues werechosen, particularly in the 3   sequences, according to Britten (1994). by using clone 20BB mixed with human genomic DNA to obtain FIG.1.  Amplification strategy.(  A  ) Tagintroducing amplification.Annealing sitesof primersHB1(5   ACAUGAACCUUUACCCCGUUGC3  ), HB2 (5   GCGCTGCAGTGCCAAGTGTTTGCTGACGC 3  ), and A5 are shown. Using a HBV-specific primer (HB1) and an  Alu -specificprimer with a tag sequence (A5), 10 cycles of amplification were performed. These primers are constructed with dUTPs. ( B ) By adding uracil DNA glycosylase, the primers (shown as filled boxes) are destroyed at dUTP sites. cTag represents sequence complementary to thetag sequence. ( C ) Further amplification is carried out by using internal primers HB2 and Tag5 (5   CAAGTGTTTGCTGACGCCAAAG 3  ).DNA strands derived from genomic sequences between two  Alu  repeats will have a tag sequence at only one end (right side) and will bepoorly amplified. It is, however, possible that the Tag primer extension products will anneal to each other: according to our results withand without UDG treatment (see Results), this seems to be a very rare event. In contrast, the DNA fragment that contains the HBV sequence and the tag sequence (left side) is logarithmically amplified.  RAPID AMPLIFICATION OF UNKNOWN FLANKING DNA   405 abouta2-kbbandusingamodificationoflongPCRprotocols(Barnes,1994; Cheng   et al.,  1994). Sequencinganalyses.  TosequencethePCRproducts,aDNAbandof interest was eluted from a 1% agarose gel into 15% polyethyleneglycol(ZhenandSwank,1993),followedbyphenol/chloroformextrac-tion and ethanol precipitation. Purified DNA was subjected to thedideoxy chain termination sequencing using either a CircumVentcycle DNA sequencing kit (New England Biolabs) or an Applied Bio-system 373A sequencer (Foster City, CA) according to the manufac-turer’s instructions. Genomic Southern blot analysis.  Five micrograms of liver DNA from a normal individual and from tumor tissue of a patient withHCC was digested with 25 units of   Hin dIII (Life Technologies, Inc.)at 37  C for 16 h. Digested DNA as well as nondigested DNA wasseparated on an 0.8% agarose gel and capillary transferred to Hy-bond-N / nylon membrane (Amersham, Les Ulis, France). The mem-brane was hybridized to a  32 P-labeled genomic probe isolated by PCR(z5/6 in Fig. 3) and washed twice in 0.1 1  SSC–0.1% SDS at 65  Cfor 15 min. After exposure, the probe was removed by washing inboiling 0.5% SDS, followed by hybridization with a  32 P-labeled HBV probe (nt 1174–1628, numbering from subtype adw by Ono  et al., 1983) and washing under the same conditions. RESULTS  In Vitro Amplification of the Viral–Host Junctions FIG. 2.  (  A  ) Amplification of the viral–host junction. Templates The overall strategy is shown in Fig. 1. To avoid (100 ng human DNA  / 0.6 pg clone 20BB) were first amplified with amplificationbetweentwo  Alu sequences(  Alu–Alu am- various concentrations of primers HB1 and A5 (lane  1,  100 pmol of  plification) in the later step, we introduced a tag se- HB1 and 10 pmol of A5; lane  2,  10 pmol of HB1 and 10 pmol of A5; quence into the 5   end of the  Alu  primer. After the  lane  3,  10 pmol of HB1 and 100 pmol of A5), followed by UDG treat- first10cyclesoftagintroducingamplification(Fig.1A),  ment and PCR with primers HB2 and Tag5. One microliter of thisproduct was subjected to heminested PCR with primers MD60 (5  these primers were destroyed by UDG treatment (Fig. CTGCCGATCCATACTGCGGAAC 3  ) and Tag5. Fifteen microliters 1B), taking advantage of incorporated dUTPs in the of the products was separated on a 1% agarose gel. By staining with primers. Thus, in the following amplification with the ethidiumbromide,theexpected2-kbbandsweredetected.Hybridiza- tag and the internal HBV primers (Fig. 1C),  Alu  se-  tion to an internal oligoprobe in the cyclin A intron confirmed thespecificity of these products (exposure for 30 min). Note that the quencesinthegenomictemplatewerenolongerprimed smaller nonspecific bands do not hybridize to this probe. ( B ) Amplifi- and  Alu–Alu  products were only poorly amplified, cation ofthe viral–host junctionfrom 100 ngof the DNAof hepatoma while the expected product with an HBV sequence at tissue. Primer pairs used were as follows: lane  1,  primers HB1 and oneendandatagsequenceattheotherwaslogarithmi-  A5 and primers HB2 and Tag5; lane  2,  primers HB1 and A3 and cally amplified.  primers HB2 and Tag3 (5   CAAGTGTTTGCTGACGACTGCA 3  );lane  3,  the same primer pairs as lane  1  without UDG treatment; Totestourmethodwetriedtoamplifytheviral–host lane  4,  the same primer pairs as lane  2  without UDG treatment. A   junction of our integration model. We mixed plasmid band of about 1500 bp that hybridizes to a total HBV probe was 20BB, which contains HBV DNA and  Alu  sequence in observed after electrophoresis on a 1% agarose gel (15  m l applied). the cyclin A gene, with normal human liver DNA in ( C ) PCR of tumor DNA was carried out by using primers HB2 and the ratio equivalent to a single copy per cell (0.6 pg of   Zt2 (Fig. 3). Liver DNA (100 ng) from the patient’s tumor tissue(lane  1 ) and from the normal individual (lane  2 ) was amplified by plasmid per 100 ng liver DNA). We obtained an ampli- touchdown PCR protocol. The expected band of 410 bp is seen only fication product of the expected size as shown in Fig. in lane  1.  This band hybridized to a probe z1/2 inside the genomic 2A. Among several combinations tested, 10-fold excess sequence (Fig. 3). of HBV to  Alu  primers greatly enhanced the amplifica-tionefficiency.WealsofoundthatsynthesisoftheHBV weperformedfourindependentreactionswithdifferentprimer with dUTP was essential for successful ampli-sets of primers for each sample; we used each of thefication (data not shown), probably because it avoided  Alu  primers for two directions (A5 and A3) in combina-amplification between  Alu  and nonspecifically an-tion with each of two HBV primers (HBV1, 5   ACA-nealed HBV primers. It was also found that employing UGAACCUUUACCCCGUUGC 3  , and HB3, 5   GAG-touchdown PCR protocol greatly enhanced sensitivityUUCUUCUUCUAGGGGACCU 3  ) for both ends, be-and specificity, which was not achieved by raising thecause the orientation of the nearest  Alu  sequence isannealing temperature or adding dimethylsulfoxamideunknown. Using the above methodology, we obtained(data not shown).part of the viral–host junction sequence in all threeWe then applied our method to liver DNA from threeliver samples tested.patients with HCC to identify unknown viral–host junction  in vivo.  To detect both ends of integrated HBV Theresultforoneendoftheviral–hostjunctionfrom  MINAMI ET AL. 406 FIG. 3.  Structure of the genomic HBV integration fragment. Positions and directions of primers are indicated by arrowheads. Sequencesof the extremities of the fragment are shown. HBV sequences extend up to nt 1371 (numbering from subtype adw by Ono  et al.,  1983) andthen are replaced by the cellular sequence. The presence of the 5   end of the  Alu  sequence confirms the correct priming of the  Alu  primer. patient 1 is shown in Fig. 2B. We observed a discrete toanHBVprobe.Partialdirectsequencingofeachfrag-ment (data not shown) revealed a HBV sequence andband of 1500 bp using the 5  -  Alu  primer (A5) and theprimer HB1 on an ethidium bromide-stained agarose an unknown adjacent sequence that had no homologyto HBV sequences in GenBank.gel (Fig. 2B, lane 1). Omitting the UDG treatment (Fig.2B, lane 3) or using the primer A3 with the HB1 (Fig.2B, lane 2) resulted in smearing of DNA. This band  DISCUSSION hybridized to a  32 P-labeled total HBV probe (Fig. 2B,right side). Complete direct sequencing of the fragment Ourmethodhasseveraladvantagesoverotherproto-revealed 200 bp of HBV sequence replaced by unknown cols used to amplify unknown flanking sequence byadjacent sequence leading to a 5   end of   Alu  sequence PCR. First, the reaction can be performed in a singleflanking the  Alu  primer (Fig. 3). To verify that this tube if a second heminested amplification is not re-flanking sequence is really of host origin, we con- quired. This simplifies the manipulations, reducing thestructed primers in this putative cellular sequence chanceofcontaminationandsavingtime.Second,there(Fig. 3). Amplification with a HBV primer (HB2) and are no ligation and/or dilution steps as used in manya cellular primer (Zt2 in Fig. 3) was successful only other protocols. These procedures are time-consuming from the patient’s DNA but not from the normal liver and limit the efficiency of amplification. Indeed, weDNA (Fig. 2C). In addition, using two primers both in could not obtain the junctional sequence when we ap-the cellular sequence (Fig. 3), we could amplify a band plied the inverse PCR technique to the same patientsoftheexpectedsizefromnormallivergenomicDNAofa used in this study. Finally, the other methods that uti-differentsubject(datanotshown).Theseresultsclearly lize nonspecific annealing of primers or degenerateddemonstrated that the flanking sequence existed adja- primers sometimes amplify fragments from inside thecent to the HBV sequence in the patient’s tumorous known sequence. Trying a modified single primer PCRtissuesandwasnotanartificialproductgenerateddur- (unpublished data), we found that this phenomenon ising PCR. To clarify further the situation  in vivo,  we prominent in searching HBV integration because therestudied DNA from the patient’s tumor and from a nor- are often free viruses coexisting with integrated viralmal liver sample by Southern blot using the flanking  sequences. In addition, a great disadvantage is thatcellular sequence (z5/6 in Fig. 3) as a probe. After  Hin-  this intraviral amplification is often recognized onlydIII digestion, two bands were identified in the tumor aftersequencingtheseproducts.Incontrast,fragmentstissue, while a single band was found in the normal obtained by our protocol should contain hybrid viralliver (Fig. 4A). This lower band in the tumor was due and cellular sequences as long as the annealing of theto HBV insertion as proved by rehybridization of the primers (HBV and  Alu ) is specific.same membrane (Fig. 4B) to an HBV probe (nt 1174– Application of the inverse PCR technique to the de-1628, subtype adw). This result indicated the presence tection of HBV integration sites has been recently re-of an HBV DNA integration in one allele of the clonally ported (Tsuei  et al.,  1994). The efficiency of amplifica-expanded tumor cells. tion was low (two viral–host junctions of 11 HCCsWe also obtained three genomic fragments for pa- tested), probably due to the ligation and dilution steps,tient 2 (250, 700, and 450 bp) and one 1000-bp DNA  competition by intraviral amplification from free vi-fragment for patient 3 that could be seen at the ethid- ruses, and polymorphic false-positive fragments thatwerenoticedonlyaftersequencing.Incontrastwehaveium bromide staining of the agarose gel and hybridized  RAPID AMPLIFICATION OF UNKNOWN FLANKING DNA   407 FIG. 4.  Southern blot of normal and HBV-integrated liver genomic DNA. We migrated on an 0.8% agarose gel 5  m g of undigested tumorDNA of the patient (lane  1 ), 5  m g of   Hin dIII-digested tumor DNA (lane  2 ), 5  m g of undigested liver DNA of a normal individual (lane  3 ),and 5  m g of   Hin dIII-digested liver DNA of a normal individual (lane  4 ), respectively. Hybridization to a genomic probe z5/6 (Fig. 3) and toa HBV probe (nt 1174–1628) are shown in  A   and  B , respectively. applied our method to three patients and successfully  ACKNOWLEDGMENTS obtained at least one viral–host junction from each The authors are grateful to Jean-Baptiste Dumas for helpful dis- sample. cussion. We also thank Chris Ormandy for critical reading of the Themaindrawbackofourprotocolisthatintegration manuscript. M.M. is a recipient of fellowships from the French Gov- sites located far from  Alu sequences might be undetect- ernment and the Fondation pour la Recherche Me´dicale. This work able. It might explain why we could not get amplifica- was supported by grants from INSERM, ARC, LNC, EC, Fegeflux, tion products in two reactions, although it is also possi-  and MRES. ble that primer sequences were deleted in the inte-grated HBV DNAs of these patients, as is often the REFERENCES case with HBV integration. Efficiency of the methodshould be enhanced by fine adjustment and application  Arnold, C., and Hodgson, I. J. (1991). Vectorette PCR: A novel ap- of long PCR protocols; however, long PCR, at present, proach to genome walking.  PCR Methods Appl.  1:  39–42. seems more sensitive to any change in amplification Barnes, W. M. (1994). PCR amplification of up to 35-kb DNA with conditions, including primer design and the sequence  high fidelity and high yield from lambda bacteriophage templates. tobeamplified(communicationin TrendsBiochem.Sci.  Proc. Natl. Acad. Sci. USA  91:  2216–2220. 1994,  19:  341–342 and personal observation). Alterna-  Bloomquist, B. T., Johnson, R. C., and Mains, R. E. (1992). Rapidisolation of flanking genomic DNA using biotin-RAGE, a variation tively, it might also be possible to use repetitive DNA  of single sided polymerase chain reaction.  DNA Cell Biol.  11:  791– sequences other than  Alu  repeats (Jurka, 1990; Jurka 797.  et al.,  1995). Bre´chot, C., Pourcel, C., Louise, A., Rain, B., and Tiollais, P. (1980). Recently, Puska´s  et al.  (1994) have reported the ap- Presence of integrated hepatitis B virus DNA sequences in cellular plication of   Alu  PCR to the whole genome. Their proto- DNA of human hepatocellular carcinoma.  Nature  286:  533–535. col depends on partial inhibition of the elongation from Britten, R. J. (1994). Evolutionary selection against change in many  Alu  primer by another chemically modified oligo-  Alurepeatsequencesinterspersedthroughprimategenomes.  Proc. Natl. Acad. Sci. USA  91:  5992–5996. primer, which appears to need careful adjustment of degree of inhibition and is thus likely to reduce ampli-  Cheng, S., Fockler, C., Barnes, W. M., and Higuchi, R. (1994). Effec-tive amplification of long targets from cloned inserts and human fication efficiency. genomic DNA.  Proc. Natl. Acad. Sci. USA  91:  5695–5699.  Along with this approach, we believe that our Chumakov, I. M., Le Gall, I., Billault, A., Ougen, P., Soularue, P., straightforward and rapid protocol will be a powerful Guillou, S., Rigault, P., Bui, H., De Tand, M. F., Barillot, E., Abd- tool, not only for the study of viral integrations, but errahim, H., Cherif, D., Berger, R., Le Paslier, D., and Cohen, D. also for the study of the human genome, as it enables  (1992). Isolation of chromosome 21-specific yeast artificial chromo-somes from a total human genome library.  Nature Genet.  1:  222– the application of PCR with  Alu -specific primers to the 225. human genomic DNA, otherwise limited to use in Collins, F. S., and Weissman, S. M. (1984). Directional cloning of  lambda phage, hybrid cells, and YAC libraries. For ex- DNA fragments at a large distance from an initial probe: A circu- ample, the application of this procedure, using   Alu  and larization method.  Proc. Natl. Acad. Sci. USA  81:  6812–6816. other repetitive sequences, should allow the rapid se- Dejean, A., Bougueleret, L., Grzeschik, K. M., and Tiollais, P. (1986). quencing of 5   and 3   noncoding regions or exon–intron Hepatitis B virus DNA integration in a sequence homologous to  junctions based on a known cDNA sequence, without  v-erbA and steroid receptor genes in a hepatocellular carcinoma.  Nature  322:  70–72. a laborious cloning step, and to perform chromosomewalking using human genomic DNA.  de The´, H., Marchio, A., Tiollais, P., and Dejean, A. (1987). A novel
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks