Available online at www.sciencedirect.com

s 70 (2008) 897–902www.elsevier.com/locate/jbbm

J. Biochem. Biophys. Method

PCR-free mutation detection of BRCA1 on a zip-codemicroarray using ligase chain reaction

Agnishwar Girigoswami, Cheulhee Jung, Hyo Young Mun, Hyun Gyu Park ⁎

Department of Chemical and Biomolecular Engineering, Korea Advanced Institute of Science and Technology (KAIST),373-1 Guseong-dong, Yuseong-gu, Daejeon, 305-701, Republic of Korea

Received 5 September 2007; received in revised form 7 October 2007; accepted 9 January 2008

Abstract

We describe here ligation-based strategy to detect mutations in BRCA1 utilizing zip-code microarray technology. In our first approach, PCRwas performed to amplify the genomic regions containing the mutation sites. The PCR products were then used as templates in a subsequentligation reaction using two ligation primers that flanked the mutation site. The primary allele-specific primer is designed to contain a base ofmutation site at its 3′ end with 5′ complementarity to the respective zip-code sequence while the secondary common primer is modified by biotinat its 3′ end. Depending on the genotype of samples at the mutation site, the nick between the two ligation primers can be sealed in the presence ofDNA ligase. The ligation products were then hybridized on the zip-code microarray followed by staining with streptavidine-cy3 to generate afluorescent signal. Using this strategy we successfully genotyped selected Korean-specific mutation sites in exon 11 of BRCA1 with a wild typeand two heterozygote mutant samples. Furthermore, we also demonstrated that ligase chain reaction using unamplified genomic DNA as directtemplates is enough to generate sufficient signals for correct genotypings in a multiplexed manner, verifying first that PCR is not essential for thismicroarray-based strategy.© 2008 Elsevier B.V. All rights reserved.

Keywords: Mutation detection; Ligase chain reaction; Zip-code microarray; BRCA mutation

1. Introduction

Genetic mutations are responsible for attacks of variousdiseases and even cause cancer. Most typical examples aremutations in BRCA1 and BRCA2, which are associated with agreater risk of developing breast, ovarian, prostate and othercancers [1]. The incidence of breast cancer and its associateddeath rate in Korea are increasing at a more rapid rate than theworld average [2–4]. Therefore, we are in great need to developan efficient tool to detect BRCA mutations for the medicalbenefit of breast cancer patients.

Wide arrays of methods and technologies have beenproposed for screening of BRCA mutations and some of themare currently used in clinical and research laboratories world-wide. They include single-strand conformation polymorphism

⁎ Corresponding author. Tel.: +82 42869 3932; fax: +82 42869 3910.E-mail address: hgpark@kaist.ac.kr (H.G. Park).

0165-022X/$ - see front matter © 2008 Elsevier B.V. All rights reserved.doi:10.1016/j.jprot.2008.01.005

assay, direct DNA sequencing, clamped denaturing gel electro-phoresis, heteroduplex analysis, and protein truncation assay[5–10]. Recently, DNA microarray-based strategy [11–14] hasemerged as one of the most powerful technologies for routineclinical detection of such human genetic mutations due to itshigh power of multiplexity, high processing speed, and lowcost. Of several microarray-based methods, zip-code microarraytechnology offers many promising opportunities due to itsunique advantages over the conventional technology [15]. Thezip-code microarray contains unique zip-code sequences withsimilar Tm values which are used to monitor products generatedfrom allele-specific reaction, such as single base extension andligation reactions. Most attractive feature of this approach is thatonce developed, the zip-code microarray can be used fordifferent sets of human genetic mutations. Therefore, over thepast few years, several methods for mutation genotyping havebeen reported based on the zip-code microarray technology[16,17].

898 A. Girigoswami et al. / J. Biochem. Biophys. Methods 70 (2008) 897–902

Along with the single base extension reaction, ligationreaction by DNA ligase can also be efficiently combined toachieve a diagnostic method for human genetic mutations,which was mainly developed by Barany group [18]. Thisligation-based technology utilizes the ability of DNA ligase toseal adjacent oligonucleotides hybridized to target DNA onlywhen the perfect complementation is present at the nickjunction. Barany group very successfully demonstrated thatseveral genetic variations can be detected by employing thistechnology [19–21]. In their studies, so called PCR/LDR(polymerase chain reaction/ligase detection reaction) strategy isusually employed, in which PCR precedes LDR to generatesufficient templates for the main ligation reaction.

Since the LDR method can provide an elegant tool for thedetection of point mutations, we herein, describe ligation-basedgenotyping strategy on the zip-code microarray with a wild typeand heterozygote mutant samples from breast cancer patients.Two Korean-specific mutation sites were selected in BRCA1exon11 (1942 and 3459) for this study [4,22]. First, wesuccessfully genotyped two mutation sites using PCR productsas templates for the main ligation reaction, which is a similarapproach with other reports [20,21]. However, we found that thePCR amplification is not essential with this strategy andunamplified genomic DNA can be directly used to generatesufficient ligated products by ligase chain reaction. Finally,using this advanced PCR-free strategy, we correctly genotypedthe Korean-specific BRCA mutations in a multiplexed manner.

2. Materials and methods

2.1. Oligonucleotides

Oligonucleotides for PCR primers, zip-code sequences, andligation primers (Tables 1 and 3) were synthesized and purifiedby Bioneer (Seoul, Korea). Concentrations of all nucleotideswere measured by CARY 100 Conc UV–Vis spectrophotometer(Varian, Australia) at λ260 nm based on the calculated molarextinction coefficient and their identities were confirmed bymatrix-assisted laser desorption/ionization — time-of-flight(MALDI-TOF).

2.2. Microarray fabrication

The zip-code sequences were diluted in 200 mM carbonatebuffer (pH 9) with 10% (v/v) dimethyl sulfoxide (Sigma-Aldrich) and spotted on aldehyde-activated glass slides (CEL-Associates, Inc. Pearland, Texas) by using a VersArray ChipWriter Compact Systems (BioRad Laboratories (Canada) Ltd.)according to the manufacturer's instructions. They were then

Table 1DNA zip-code sequences used in this study

Zip-code Sequence (5′→3′)

Z1 NH2 — TGCGGGTAATCGZ2 NH2 — ATCGTGCGACCTZ3 NH2 — ATCGGGTATGCGZ4 NH2 — ACCTGACCATCG

allowed to covalently bind to the glass surface for 12 h in ahumid chamber. The spots were about 200 μm in diameter andthe centre to centre distance between adjacent spots was300 μm. To verify the reproducibility six spots were printed foreach oligonucleotide on the same glass slide.

2.3. DNA isolation, sequencing, and PCR amplification

We prepared genomic DNA (gDNA) from whole blood of anapparently healthy subject and two breast cancer patients byusing genomic DNA extraction kit (Qiagen, Hilden, Germany)and our laboratory standard protocol [14]. After PCRamplification, direct sequencing was performed using an ABIDye Terminator Cycling Sequencing Kit (Applied Biosystems)according to the manufacturer's instructions, followed byanalysis on an ABI3700 DNA sequencer (Applied Biosystems).

Each PCR amplification was carried out in a thermocycler(Applied Biosystems) using a 50 μl total reaction volumecontaining 100 ng gDNA, 0.5 μM each primer [for 1942mutation site: 5′-AGATTTGGCAGTTCAAAAGA-3′ (for-ward), 5′-TGCAATTCAGTACAATTAGG-3′ (reverse), for3459 mutation site: 5′-TGGAAGTAATTGTAAGCATCCT-GAAATAAAAA-3′ (forward), 5′-GGGAAGCTCTTCATCCT-CACTAGATAA-3′ (reverse)], 5 μl 10× Pyrobest (Takara BioInc, Japan) reaction buffer (20 mM tris–HCl, pH 8.3 at 37 °C,10 mM KCl, 2 mM MgCl2), 0.2 mM dNTPs (Pyrobest) and0.5 U Taq DNA polymerase (Pyrobest). The reaction tubeswere heated at 94 °C for 5 min, followed by 35 cycles of 30 s at94 °C, 30 s at 57 °C and 1 min at 72 °C. The reaction ended with5 min extension at 72 °C and then cooling at 4 °C forpreservation. After the amplification, the PCR products werepurified by Qiaquic purification kit (Qiagen) and amplificationwas confirmed by agarose-gel electrophoresis.

2.4. Ligation reaction

Ligation reactions were performed by using PCR products ordirect gDNA as template DNA (tDNA) in a 20 μl volumecontaining tDNA (40 ng), two primary primers correspondingto both of wild and mutant sequences (5 pmol), one commonsecondary primer (5 pmol), reaction buffer (30.5 mM Tris–HCl,27.5 mM KCl, 10 mM MgCl2, pH 8.3), 1 mM DTT, 2.5 mMNAD+ and 5 units Tfi (thermus filiformis)DNA ligase (Bioneer,Korea). The reaction tubes were heated at 95 °C for 5 minfollowed by 25 cycles of 30 s at 95 °C and 2 min at 65 °C andthen a final 5 min extension at 95 °C. Finally the reactionmixture was cooled at 4 °C for preservation.

2.5. Hybridization and staining

10 μl ligated products were mixed with 20–30 μl 1× SSPETbuffer (0.01 M phosphate buffer, 0.15 M NaCl, 0.001 M EDTA,pH 7.4, Triton X-100 0.01%) and applied to the zip-codemicroarray followed by an incubation at 38–45 °C for 6 h.Then, the hybridized slides were washed by 6× SSPET buffer(0.9 M NaCl, pH 7.4, Triton X-100 0.005%) and distilled waterfor 8 min and 5 min respectively, and finally air dried at room

899A. Girigoswami et al. / J. Biochem. Biophys. Methods 70 (2008) 897–902

temperature. Streptavidin-cy3 (Sigma-Aldrich) was added into1× SSPET buffer to give the final concentration of 1 μg/ml,applied to the hybridized glass slide for labeling, and incubatedat 25 °C for 3 h followed by washing with 6× SSPET for 8 minand distilled water for 5 min. After drying, the glass slides werescanned at 532 nm by a special Genepix 4000A MicroarrayScanner (Axon Instruments, USA).

3. Results

3.1. Overall strategy

We examined the ability of the ligation-based approach todetect Korean-specific BRCA1 mutations on a zip-code micro-array. Scheme 1 shows whole strategy of our detection method.First, the template DNA for ligation reaction was prepared fromthe sample under interrogation. As the template DNA of theligation reaction, either PCR product or unamplified genomicDNA can be used. To generate the PCR template, PCRamplification of genomic regions containing mutation siteswas performed using gene-specific primers. Two kinds ofligation primers were then allowed to hybridize to the templatesequence. The primary allele-specific primer was designed tohybridize to the 3′ part of the template sequence and end with abase corresponding to the mutation base of the template at its 3′end. It contains on its 5′ end a zip-code complement that directsthe ligation product to the respective zip-code position. For each

Scheme 1. Schematic representation of genotypin

mutation site, there are two different primary primers being usedwhich correspond to both of wild type and mutant sequences.Only difference between the two primary primers is the last baseat their 3′ end (Table 3). The secondary common primer wasdesigned to hybridize to the 5′ part of the template sequence andlabeled with biotin at the 3′ end.

All zip-code sequences were randomly selected from a zip-code pool generated by shuffling three different tetramersequences according to the method developed by Gerry et al.[16]. For this type of detection method, nonspecific binding is acommon interference to be overcome and length of the ligationprimers is a key factor in determining the specificity of thisstrategy by reducing the nonspecific interaction. Through thepreliminary test, our designed primary primers were 27–28 merfor 1942 site and 30mer for 3459 site. Secondary primers were 21and 19 base long for 1942 and 3459 mutation sites, respectively.

3.2. Detection of mutations using PCR product-based ligationon a zip-code microarray

To demonstrate the diagnostic validity of the strategy,mutation detection analysis was performed for two selectedKorean-specific mutation sites in BRCA1 exon 11 (Table 2).One of the two mutation sites contains single base deletion(1942) i.e. base ‘A' is deleted in mutant sample and the othercontains a base substitution (3459) i.e. base ‘G’ in wild typesample is substituted by base ‘T’ in mutant sample. Following

g by ligase chain reaction on zip-code array.

Table 2Korean-specific BRCA1 mutations detected in this study

Region Position Mutation effect

Exon 11 1942 delA:611XExon 11 3459 GNT:1114X

900 A. Girigoswami et al. / J. Biochem. Biophys. Methods 70 (2008) 897–902

the conventional approach [20,21], we carried out the genotyp-ing analysis with the human samples using PCR product asligation template. For the genotyping of 1942 wild type andmutant samples, ligation reaction was carried out for eachsample in the presence of two primary primers (P1A and P1B)and one common secondary primer (P2AB) (Table 3). After thehybridization of the ligation product on the zip-code microarrayfollowed by the staining, we finally obtained good fluorescencesignals at the desired place as shown in Fig. 1. The wild typesample generated the fluorescence signal only at the Z1 positionwhereas the mutant sample generated fluorescence at both Z1and Z2 positions. In the same way the ligation reaction wascarried out with two primary primers (P1C and P1D) and acommon secondary primer (P2CD) in the presence of 3459 wildtype and mutant sample. The 3459 wild type sample producedfluorescence at Z3 position and mutant sample gave fluores-cence at both Z3 and Z4 positions.

3.3. PCR-free mutation detection

After successful genotyping of the two BRCA1 mutationsites using the PCR-amplified genomic regions, we envisionedthat the strategy would be much more beneficial if the PCRamplification step can be avoided. Therefore, we tried togenotype the same mutation sites using the genomic DNA asdirect ligation template and we obtained the fluorescent signalsat the desired positions on the zip-code microarray and thesignal intensities were more than sufficient for correctgenotyping. For this study, we spotted four zip-code sequences(Z1 and Z2 for 1942 and Z3 and Z4 for 3459) in four differentrows on a glass slide (Fig. 2). Then we performed multiplexligation reactions using the reaction mixture containing all fourprimary primers and two secondary primers corresponding toboth the two mutation sites, but unamplified gDNAwas directlyused as ligation template instead of PCR product. When thewild type sample was tested, we obtained fluorescence spots at

Table 3Ligation primers for the detection of mutations in BRCA1 exon11

Mutation position Name Probe sequence (5′→3′)

1942 P1A CGATTACCCGCA AATTAAATAAAAGCACCTAAAAA

P1B AGGTCGCACGATAATATCCACCCTAAAA

P2AB GAATAGGCTGAGGAGGAAGT3459 P1C CGCATACCCGATAGCATCCTG

GCAAGAATATGP1D CGATGGTCAGGTAGCATCCTG

GCAAGAATATTP2CD AAGAAGTAG TTCAGACTGT-

The sequence complementary to the zip-code sequence is underlined and bold letter

Z1 for 1942 and at Z3 for 3459 (Fig. 2(a)). With the 1942mutant sample, we obtained signals at Z1, Z2 and Z3 positions(Fig. 2(b) while 3459 mutant sample generated fluorescencesignals at Z1, Z3 and Z4 positions (Fig. 2(c).

4. Discussion

In the present study, we have successfully genotyped twoselected Korean-specific BRCA mutation sites by combiningthe specificity of DNA zip-code microarray and the high fidelityof the ligation reaction. We mainly used the activity ofthermostable monomeric protein Tfi DNA ligase to catalyzethe sealing of 5′-phosphate and 3′-hydroxyl termini between theprimary and secondary primers [23]. The ligated samples werethen hybridized with zip-code on microarray followed bystaining. For the signaling of the ligated products, a directlabeling method can also be adopted using a direct fluores-cence-labeled secondary primer (e.g. Cy3-primer). In thismethod, no further staining step is required. Biotin-labeledprimers, however, are much less expensive and are supposed tobe better ligated by the enzyme due to the small size of thebiotin molecule. Thus, this two-step signaling strategy adoptedin this study is considered to be more efficient and economical[13,14].

To confirm the ligation-based strategy, we first usedconventional PCR amplified template for the ligation. Afterhybridization and staining the 1942 wild type sample generatedthe fluorescence signal only at the Z1 position (Fig. 1(a))indicating that only P1A primer ligated with P2AB in thepresence of the wild type template. On the other hand, thesignals were obtained from both wild type Z1 and mutant Z2positions with the 1942 mutant sample (Fig. 1(b)), correctlyindicating that the sample has heterozygote mutant genotype at1942 site. Similarly, proper genotyping determinations weresuccessfully achieved with the 3459 wild type and mutantsamples (Fig. 1(c) and (d)), demonstrating the clinicalusefulness of this strategy.

To demonstrate that the PCR is not essential with this strategy,we performed the ligation reaction in the presence of genomicDNA and we obtained more than enough fluorescence signals forcorrect genotyping. Signals at Z1 and Z3 (Fig. 2(a)) for wild typesample indicate that two pairs of wild type-specific primers (P1Aand P2AB and P1C and P2CD) were ligated. This result indicates

Corresponding sample Zip-code

T CCACAATTCA Wild Z1

AATTCAAAAGCA Mutant Z2

C-biotin Wild and mutantAAATAAAAAA Wild Z3

AAATAAAAAA Mutant Z4

biotin Wild and mutant

indicates a base at mutation sites.

Fig. 2. Fluorescence images of the ligation product from the gDNA template usingall probes — P1A, P1B, P1C, P1D, P2AB and P2CD — a) wild type sample;b) 1942 mutant sample; and c) 3459 mutant sample.

Fig. 1. Fluorescence images of the ligation product from the PCR-amplifiedtemplate — a) 1942 wild type sample using P1A, P1B and P2AB; b) 1942mutant sample using P1A, P1B and P2AB; c) 3459 wild type sample using P1C,P1D and P2CD; and d) 3459 mutant sample P1C, P1D and P2CD.

901A. Girigoswami et al. / J. Biochem. Biophys. Methods 70 (2008) 897–902

that the sample has homozygote wild genotypes both at the 1942and 3459 mutation sites, which is correct genotyping. The 1942mutant sample has heterozygote mutant sequence at 1942, butwild type sequence at 3459, which was confirmed by priorsequencing. With this mutant sample, we obtained signals at bothZ1 and Z2 for 1942 site while fluorescence was generated only atZ3 for 3459 position (Fig. 2(b)) indicating correct determinationof the heterozygote mutant genotype at 1942 and homozygotewild genotype at 3459. Similarly, we correctly got threefluorescence spots for the mutant 3459 sample at Z1, Z3, andZ4 (Fig. 2(c)). Again, the fluorescence signal at Z1 indicateshomozygotewild genotype at 1942, and both signals at Z3 and Z4indicate heterozygote mutant genotype at 3459. These resultshighlighted that this PCR-free strategy could be reliablyemployed for the diagnosis of BRCA mutations in a multiplexedmanner. Particularly, by demonstrating that PCR is not essential

with this strategy and unamplified genomic DNA is enough toserve as direct ligation template, we made our strategy morepowerful in its clinical utility for human mutation screening orSNP genotyping, in many aspects including reduced detectiontime, lower reaction steps, and more reduced cost. The currentwork contained only two limited mutation sites because this is aproof of concept study for the PCR-free diagnosis, but higherdegree of multiplexing can be achieved by increasing the numberof ligation primers and their corresponding zip-code sequences onthe microarray.

5. Simplified description of the method and its(future) application

The present approach reported here is a proof-of-principlefor genotyping assay method that can detect SNPs through theligase chain reaction. The two oligonucleotides are joined

902 A. Girigoswami et al. / J. Biochem. Biophys. Methods 70 (2008) 897–902

covalently by the action of DNA ligase on genomic DNAtemplate, provided that the nucleotides at the junction arecorrectly base paired. It is beyond doubt that PCR/LCR/zip-code array combination is a very powerful tool for the samepurpose. But the current strategy excludes PCR and success-fully generated enough fluorescent signals at the right positionon the DNA-array for the correct genotyping of real samplecollected from the patient. The exclusion of PCR permits therapid and cost-saving identification of SNP directly fromgenomic DNA in a multiplex fashion which could be a powerfulclinical tool in a near future.

Acknowledgements

This work was supported by the Brain Korea 21 (BK21)program, the Korea Research Foundation and the Centre forUltramicrochemical Process Systems.

References

[1] Lux MP, Fasching PA, Beckmann MW. Hereditary breast and ovariancancer: review and future perspectives. J Mol Med 2006;84:16–28.

[2] Seo JH, Cho DY, Ahn SH, Yoon KS, Kang CS, Cho HM, et al. BRCA1 andBRCA2 germline mutations in Korean patients with sporadic breast cancer.Human Mutat 2004;24:350.

[3] Son BH, Kwak BS, Kim JK, Kim HJ, Hong SJ, Lee JS, et al. Changingpatterns in the clinical characteristics of Korean patients with breast cancerduring the last 15 years. Arch Surg 2006;141:155–60.

[4] Ahn SH, Son BH, Yoon KS, Noh DY, Han W, Kim SW, et al. BRCA1 andBRCA2 germline mutations in Korean breast cancer patients at high risk ofcarrying mutations. Cancer Lett 2007;245:90–5.

[5] Castilla LH, Couch FJ, ErdosMR.Mutations in theBRCA1 gene in familieswith early onset breast and ovarian cancer. Nat Genet 1994;8:387–91.

[6] Friedman LS, Ostermeyer EA, Szabo CL, Dowd P, Lynch ED, Rowell SE,et al. Confirmation of BRCA1 by analysis of germline mutations linked tobreast and ovarian cancer inten families. Nat Genet 1994;8:399–404.

[7] Simard J, Tonin P, Durocher F, Morgan K, Rommens J, Gingras S, et al.Common origins of BRCA1 in Canadian breast and ovarian cancerfamilies. Nat Genet 1994;8:392–8.

[8] Hogervorst FBL, Cornelis RS, Bout M, Vanvliet M, Oosterwijk JC, OlmerR, et al. Rapid detection of BRCA1mutations by the protein truncation test.Nat Genet 1995;10:208–12.

[9] Gayther SA,WarrenW, Mazoyer S, Russell PA, Harrington PA, ChianoM,et al. Germline mutations of the BRCA1 gene in breast and ovarian cancer

families provide evidence for genotype–phenotype correlation. Nat Genet1995;11:428–33.

[10] Gayther SA, Harrington PA, Russell PA, Kharkevich G, Garkavtseva RF,Ponder BAJ, et al. Rapid detection of regionally clustered germlineBRCA1 mutations by multiplex heteroduplex analysis. Am J Hum Genet1996;58:451–6.

[11] Tyagi S, Bratu DP, Kramer FR. Multicolor molecular beacons for allelediscrimination. Nat Biotechnol 1998;16:49–53.

[12] WangDG, Fan JB, Siao CJ, BernoA, Young P, Sapolsky R, et al. Large-scaleidentification, mapping, and genotyping of single-nucleotide polymorphismin the human genome. Science 1998;280:1077–82.

[13] Song JY, Park HG, Jung SO, Park J. Diagnosis of HNF-1alpha mutationson a PNA zip-code microarray by single base extension. Nucleic Acids Res2005;33:e19.

[14] Yim SC, Park HG, Chang HN, Cho DY. Array-based mutationdetection of BRCA1 using direct probe/target hybridization. AnalBiochem 2005;337:332–7.

[15] Zabarovsky ER, Petrenko L, Protopopov A, Vorontsova O, Kutsenko AS,Zhao Y, et al. Restriction site tagged (RST) microarrays: a novel techniqueto study the species composition of complex microbial systems. NucleicAcids Res 2003;31:e95.

[16] Gerry NP, Witowski NE, Day J, Hammer RP, Barany G, Barany F. UniversalDNA microarray method for multiplex detection of low abundance pointmutation. J Mol Biol 1999;292:251–62.

[17] Hirschhorn JN, Sklar P, Lindblad-Toh K, Lim YM, Ruiz-Gutierrez M,Bolk S, et al. SBE-TAGS: an array-based method for efficient single-nucleotide polymorphism genotyping. 2000;97:12164–9.

[18] Barany F. The ligase chain reaction in a PCR world. PCR Methods Appl1991;1:5–16.

[19] Favis R, Day JP, Gerry NP, Phelan C, Narod S, Barany F. Universal DNAarray detection of small insertions and deletions in BRCA1 and BRCA2.Nat Biotechnol 2000;18:561–4.

[20] Hashimoto M, Hupert ML, Murphy MC, Soper SA, Cheng YW, Barany F.Ligase detection reaction/hybridization assays using three-dimensionalmicrofluidic networks for the detection of low-abundant DNA pointmutations. Anal Chem 2005;77:3243–55.

[21] Hashimoto M, Barany F, Soper SA. Polymerase chain reaction/ligasedetection reaction/hybridization assays using flow-through microfluidicdevices for the detection of low-abundant DNA point mutations. BiosensBioelectron 2006;21:1915–23.

[22] Kang HC, Kim IJ, Park JH, Kwon HJ, Won YJ, Heo SC, et al. Germlinemutations of BRCA1 and BRCA2 in Korean breast and/or ovarian cancerfamilies. Human Mutat 2002;20:235.

[23] Lehnman IR. DNA ligase: structure, mechanism, and function. Science1974;186:790–7.