Genomic structure of HOXD13 gene: a nine polyalanine duplication causes synpolydactyly in two unrelated families
Genomic structure of HOXD13 gene: a nine polyalanine duplication causes synpolydactyly in two unrelated familiesA. Nurten Akarsu, Ivaylo Stoilov, Engin Yilmaz, B. Sitki Sayli1 and Mansoor Sarfarazi*
Surgical Research Center, Department of Surgery, University of Connecticut Health Center, Farmington, Connecticut, USA and 1Department of Medical Biology and Genetics, University of Ankara, Faculty of Medicine, Sihhiye, Ankara, Turkey
Received April 16, 1996;Revised and Accepted May 3, 1996
Synpolydactyly (SPD) is a limb malformation that shows a characteristic manifestation in both hands and feet. This condition is inherited as an autosomal dominant trait with reduced penetrance. We have recently mapped this locus centromeric to the HOXD8 intragenic marker and suggested the HOXD13 gene as a potential candidate for this condition. The genomic structure of HOXD13 established in this study consists of two exons that encodes a polypeptide of 335 amino acids. The downstream exon at the 3' end of this gene contains the homeodomain sequences that are highly conserved. Sixty-three bp upstream of this exon lies a stretch of intronic CA-repeats that proved to be polymorphic in two different populations. The upstream exon encodes 75% of the entire protein and contains a stretch of 15 normal alanines at its 5' end. Sequence comparison at this position in the homozygous affected individuals identified a total of 24 alanine residues that resulted from a duplication of nine polyalanines. In two unrelated SPD families, this duplication was directly transmitted from the affected parents to their affected, but not unaffected, offspring; in one family its size has remained constant for at least 150 years spanning over seven generations. The presence of this duplication confirmed the status of four normal gene carriers, one incomplete penetrance and two affected individuals who were recombinants for HOXD8 or HOXD13-CA repeat markers. This duplication was not present in 150 chromosomes of unrelated healthy subjects of two different populations.
Synpolydactyly (SPD; MIM # 186000) is an autosomal dominant condition with a clinical presentation that, in general, bilaterally affects the 3rd and 4th fingers of hands and 5th toes. This condition is characterized by an increased number of digits that is usually associated with syndactyly in the mesoaxial line of hands and postaxial part of feet. However, proximal-distal and anterior-posterior patterning is normal in the heterozygous affected individuals (1 ). Recently, we reported a family from the village of Derbent in Turkey that consisted of 425 subjects, 182 of whom are affected with SPD (2 ). This kindred also included a total of seven severely affected homozygous individuals exhibiting a clear bony reduction, proximal-distal duplication and an increased number of digits in the anterior-posterior direction (3 ). We subsequently used this kindred and mapped the SPD locus to the 2q31 region and reported a tight linkage with an intragenic marker at the HOXD8 locus (4 ). Another SPD family from Turkey also showed linkage to the same region but with different haplotype composition, from those observed in the Derbent kindred, suggesting that they may have resulted from different mutations. The overall linkage study of these two kindreds placed the SPD locus 1.2 cM (Z = 41.23) centromeric to the HOXD8 intragenic marker and within an estimated region of 2.4 cM that is flanked by D2S1238. In order to reduce the size of the SPD critical region, four CA-repeats newly identified by the Genethon group (5 ) that map within the candidate region (i.e., D2S2188, D2S2307, D2S2257 and D2S2314) were genotyped in the original recombinant branches of the Derbent kindred. However, none of these provided any new information and we were, therefore, unable to narrow down the SPD candidate region further.
Our previously reported recombination with the HOXD8 intragenic marker in an affected individual (4 ) excluded the most 3' end of the HOXD cluster and implicated other genes at the 5' end (i.e., HOXD9 to HOXD13, EVX2 and DLX1/DLX2) as potential candidates for this condition. We further suggested that the HOXD13 within this cluster is more likely to be involved in this phenotype. This was based on the previously reported pattern of HOXD13 gene expression (6 -7 ) and on the phenotypic presentation of the affected homozygous subjects compared to the heterozygotes (2 -3 ). Furthermore, as the HOXD13 gene structure was unknown, we initiated the genomic identification of its normal coding sequences, intron-exon boundaries and mutation screening of this gene in our two SPD families.
We determined the intron-exon structure and the complete coding sequence of the human HOXD13 gene by using direct sequencing, polymerase chain reaction (PCR) amplification with degenerate primers and reverse transcription PCR (RT-PCR). Initially, we sequenced the regions upstream and downstream from the known HOXD13 homeodomain (7 -8 ) using primers SPD1 and SPD2 as presented in Table 1 . Analysis of the obtained sequences identified an in-frame stop codon (TGA) that was in agreement with its previously published position (7 ) and a new dinucleotide repeat (CA)16 that is located 310 bp upstream from this stop codon. Next, we attempted to recover the sequences of the upstream exon by PCR amplification with degenerate primers. Using the cosmid G2 clone (9 ) as a template, a fragment of 1.3 kb was amplified with a forward degenerate primer [i.e. SPD3 designed according to the DGLRGDS amino acid motif which is encoded by the 5' end of the chicken hoxd13 gene (10 )] and a reverse gene specific primer (i.e. SPD2). Sequence analysis confirmed that the 3' end of this fragment was identical to the human HOXD13 sequence that is located upstream from primer SPD2. The sequence of its 5' end was then used as an initiation point for two consecutive upstream sequence runs (primers SPD4 and SPD5). Six-frame translations of the obtained sequences revealed an open reading frame spanning 993 bp. In order to test the possibility that this open reading frame may contain the HOXD13 coding sequences, primer SPD6 was designed to anneal within this open reading frame and was paired with primer SPD2 for PCR amplification of cDNA and genomic DNA (Fig. 1 ). The specific amplification of two different sized fragments was considered as evidence that the genomic region flanked by these primers has been subjected to transcription and splicing. Comparative sequence analysis of these two fragments identified the 3' end of the upstream exon, the 5' end of the downstream exon and the intron-exon boundaries of the HOXD13 gene. The established intron-exon boundaries (AGgta.....agGGA) conforms with the gt-ag rule. The size of the intron separating these two exons was estimated to be approximately 950 bp. An in-frame ATG initiation codon was identified 1005 bp upstream from the TGA stop codon as originally reported by Dolle et al. (7 ). Translation of this fragment produced a polypeptide of 335 amino acids (Fig. 2 ) that is very similar to that encoded by the chicken hoxd13 gene (Fig. 3 ). The DNA sequences obtained from the cosmid clone G2 was subsequently confirmed by additional sequencing of PCR fragments containing the entire gene that were generated from at least two different unrelated normal individuals (primer pairs SPD7/8 and SPD11/8). A complete list of primers used in this study is presented in Table 1 .
We first screened the homeodomains of both the HOXD12 and HOXD13 genes but observed no differences between the normal and affected subjects. During the characterization of the HOXD13 gene structure, we designed a series of oligos from the newly identified coding sequences and screened the homozygous affected individuals and their respective pedigree members for potential mutations. Furthermore, we sequenced one homozygous affected individual and compared it to the normal sequences. Direct comparison of DNA sequences at the 5' end of the HOXD13 gene in the normal and affected homozygous individuals identified a 27 bp duplication of the normal sequences that encodes for an additional nine alanine residues in the affected homozygous subjects. The normal individuals were identified to have a stretch of 15 polyalanine residues [poly(A)] which is located 145 bp downstream from the initiation codon, but the affected homozygotes had a total of 24 poly(A). Further inspection of DNA sequences in the affected homozygotes confirmed that this is due to a duplication of nine alanine residues that is inserted at position 187 and between the 14th and 15th normal alanine residues (Fig. 4 ). Three homozygous affected individuals from two different branches of the Derbent kindred were sequenced again which confirmed this poly(A) duplication. Further sequencing of two affected heterozygotes, one SPD normal subject and one additional unrelated normal individual confirmed this observation. We further designed a set of primers (i.e. SPD5 and SPD7) that allowed us to screen for this duplication by PCR and agarose gel electrophoresis.
As previously reported (4 ), during the mapping of the SPD locus, we identified a single recombination with the HOXD8 intragenic marker in an affected individual. Most recently during the characterization of the HOXD13 gene, we identified a CA-repeat polymorphism that also showed recombination with an additional affected subject of a previously unreported branch of the Derbent kindred. In order to substantiate the true nature of these two recombinants, the respective branches of these individuals were screened for the observed HOXD13 polyalanine duplication. As shown in Figure 6 , both of these affected individuals also carried the duplicated polyalanine residues and, therefore, themselves are heterozygous for the normal and poly(A) duplicated bands as previously observed in other branches of the same pedigree. Therefore, identification of a recombination within a region of ~1.5 kb and between the HOXD13 CA-repeat and HOXD13 poly(A) duplication is of considerable interest and may imply that this region of the genome is a hot spot for recombination. However, the recombination with the HOXD8 intragenic marker is comparatively more likely as location of the poly(A) duplication is at least 40 kb (8 ) from the site of this recombination.
Figure 6. Inheritance of the poly(A) duplication in two affected members of the Derbent kindred who were recombinant for HOXD13 CA-repeat (Family `A') and HOXD8 (Family `B') intragenic markers. Only the recombinant part of each respective branch is shown. The linkage map for the closely flanking markers are shown in the middle of the picture. In each pedigree, the cosegregation of the SPD phenotype with the affected haplotype is shown with a solid vertical bar. The haplotype in person 1 of family `A' is inferred, and it is based on the overall co-segregation of these markers in that kindred. As shown in the picture, all the affected individuals are genotypically heterozygous for the poly(A) and wild type bands. Letter `M' denotes the 1 kb DNA ladder, `N' is the unrelated normal, `H' is an homozygous individual and `C' is the negative control.
We have previously identified and reported one female individual who has inherited the entire affected chromosome from her affected father and subsequently passed it on to her affected son (4 ). However, both her physical and X-ray examinations of hands and feet revealed that she is phenotypically normal and, therefore, classified her as a proven example of incomplete penetrance for this phenotype. We have also identified four additional normal members of the Derbent kindred who have inherited the entire affected haplotype from their respective parents but themselves are phenotypically normal. However, none of these have any children as yet; and, therefore, their status as gene carriers remained unknown until now. As shown in Figure 7 , all of these phenotypically normal subjects carried the same size duplication and, therefore, themselves are gene carriers for the SPD phenotype. This also proved at the molecular level that the SPD phenotype is truly an incomplete penetrant condition. In our recent study of two Turkish families, haplotype analyses of a total of 169 individuals (105 affected) showed that 164 of them expressed the disorder phenotypically as predicted by their genotype. Thus, the gene expression is estimated as being 97% for this disorder and the remaining 3% are expected to be gene carriers as observed for the above-mentioned five individuals.
Figure 7. One incomplete penetrant (lane 2) and four gene-carrying (lanes 4-7) subjects of different branches of the Derbent kindred. Individuals in lanes 1 and 3 are SPD affecteds while individuals in lanes 2 and 4-7 are phenotypically normal. Individuals marked as `N' and `H' are unrelated normal and homozygous affecteds showing the wild type and poly(A) duplicated bands respectively. The affected father in lane 1 passed on the poly(A) band to his normal daughter in lane 2 who, in turn, passed it on to her affected son in lane 3. Individual in lane 4 is the half sib of person in lane 2 (different mother).
We identified the genomic structure of the human HOXD13 gene that is comprised of 1005 bp exonic sequences and encodes for a polypeptide of 335 amino acids (Fig. 2 ). This gene contains an upstream exon (757 bp) and a downstream exon (248 bp) that are separated by an intron of approximately 950 bp. The genomic organization of HOXD13, therefore, closely resembles a structure that is shared by other members of the HOXD cluster (11 ). The upstream exon at the 5' end of this gene encodes for 75% of the entire HOXD13 protein. This exon contains 15 polyalanine residues at its 5' end. Amino acid alignment of this gene in humans and chickens revealed a high degree of homology between the two species with most of the differences being located at the N-terminus of the respective polypeptide chains (Fig. 3 ). Pairwise comparison of the two polypeptides with the MAP program identified 37 new amino acids in the human sequences that have been inserted within the conserved sequences. However, the first five amino acids are totally identical between the two species.
The downstream exon of the HOXD13 gene contains the entire homeodomain sequences that are highly conserved. Sixty-three bp upstream of this exon lies a stretch of intronic CA-repeats that proved to be polymorphic in two different populations (Fig. 2 ). A total of six alleles were observed in this polymorphism which showed heterozygosity of 50% in the Turkish population and 30% in the Caucasian population.
Sequence comparison of the upstream exon in the homozygous affected members of the Derbent kindred identified nine additional polyalanine residues that resulted from a duplication starting at position 187 (Fig. 4 ). The direct transmission of the poly(A) duplicated band from the affected parents to their affected, but not the normal offspring, confirmed its segregation with the SPD phenotype (Figs 5 and 6 ). The nine polyalanine duplication also segregated in different branches of the Derbent kindred and one other unrelated family from the south-east of Turkey. It is also interesting to note that the size of this poly(A) duplication has remained constant for at least 150 years and over seven generations of the Derbent kindred. Detection of the same number of poly(A) duplications in two unrelated SPD families raised the possibility that either a common ancestral poly(A) duplication may exist in the Turkish population or that these two pedigrees have resulted from two independent mutations. As these two kindreds showed completely different haplotype composition for all the polymorphic DNA markers studied (see list in Fig. 6 ), it is less likely that a common ancestral mutation existed in this population. Furthermore, we did not observe any linkage disequilibrium with both HOXD8 and HOXD13-CA repeat intragenic markers that reside 40 kb and 1.5 kb from this poly(A) duplication respectively. Therefore, two independent poly(A) duplications have directly caused the SPD phenotype in these two unrelated pedigrees from different geographical areas and possibly with different genetic backgrounds. We have further tested a total of 150 chromosomes of unrelated healthy subjects of two different populations and did not observe the poly(A) duplication. Nevertheless, the same individuals were polymorphic for the HOXD13-CA repeats that are found 1.5 kb downstream of this duplication.
Taken altogether, this poly(A) duplication is within the coding sequences of the HOXD13 gene, it has not been observed in the general population, and it is exclusively segregated with the affected but not the normal subjects of two large unrelated families, which suggests its causative nature in these pedigrees.
The role of HOX genes in limb development has been the subject of intensive studies by a number of investigators. Although it is anticipated that the HOXD genes play a role in the control of limb morphogenesis, very little is known about the degrees or how these genes participate in the normal limb pattern formation (12 ,13 ). Although it is known that the 3' end of the HOX genes contain homeodomain sequences that are well conserved and have transcriptional activities that regulate various developmental processes (11 ,14 ), only limited knowledge is available on the functional activities of the 5' end of these genes. The presence of a number of short repeats known as homopolymeric amino acid stretches (HPAA) at the 5' end of the homeobox genes is a common feature, but their biological importance is not known (14 ,15 ). The identification of the poly(A) duplication at the 5' end of the HOXD13 gene in this study is the first example of an abnormality in a gene within this cluster that can directly be attributed to a clinical phenotype in man. The exact nature of this duplication and the mechanisms by which it leads to the clinical presentation of the SPD phenotype still needs to be determined. However, due to the hydrophobic properties of the polyalanine tracks, it has been suggested that they play a role in the determination of protein confirmation or protein-protein interaction rather than being involved in the DNA binding function (16 ). Further comparative studies on other species and biochemical characterization of the HOXD13 protein product can help to understand how this gene is involved in the limb development and the biological significance of the HPAA stretch sequences.
The clinical manifestation and pedigree structure of the Derbent kindred ascertained from the western part of Turkey has already been published (2 ). A detailed clinical observation in the seven homozygous subjects of this kindred (3 ) and the mapping of the SPD locus have also been reported previously (4 ). More recently, genotyping and haplotype construction of the two flanking markers (i.e., HOXD8 and D2S1238) in all of this kindred established that: (i) this pedigree has resulted from a founder effect which started at least 150 years ago; (ii) the kindred consists of one proven incomplete penetrant who has inherited an affected haplotype from her father and segregated it to her affected son; (iii) four phenotypically normal individuals are gene carriers as they have inherited the entire affected haplotype from their respective affected parents; and (iv) the seven homozygous individuals have inherited two affected haplotypes from their two affected parents while their heterozygous sibs have received one normal and one affected haplotype. Their only normal sib had inherited two copies of the normal haplotypes from her two affected parents. Additionally, X-rays have been obtained from the entire spine of two homozygous subjects. Anterior-posterior and lateral X-ray evaluation of the vertebral bones showed a total coccygeal agenesis in a 7 year old homozygous male but not in his 10 year old sister. Cervical, thoracal, lumbal and sacral vertebral bones were all normal. A second Turkish family consisting of five affected and six normal individuals in two generations have also been identified, sampled and proven to be linked to the same region.
For sequencing, the cosmid DNA was prepared with QIAGENTM Plasmid Midi Kit which utilizes the optimized alkaline lysis method of Birnboim and Doly (17 ). The amplified PCR fragments were purified directly or from agarose gels with WizardTM PCR Preps DNA Purification System (Promega). Sequence analysis was based on the dideoxy method originally developed by Sanger et al. (18 ). Dye terminator sequencing with Taq Polymerase was performed on an ABI-373 sequencer (Perkin Elmer). Manual sequencing was carried out with CircumVentTM Thermal Cycle Dideoxy DNA Sequencing Kit (New England Biolabs) according to the protocol supplied by the manufacturer. The gels were silver stained (19 ) and photographed.
Monolayers of human skin fibroblasts have been maintained at 37oC in CO2 incubator, in media MEM supplemented with 10% FBS and antibiotics. For total RNA isolation, preconfluent cells from a 60 mm dish were treated with 2 ml TRIzol reagent (GIBCO). Next, 400 [mu]l chloroform was added and the aqueous phase was recovered by centrifugation. RNA was precipitated with 1 ml isopropanol, washed with 75% ethanol and diluted in 50 [mu]l DEPC treated water. First strand cDNA synthesis was primed from 5 [mu]g total RNA with 50 ng random hexamers (GIBCO). Reaction was carried out in 20 mM Tris-HCl (pH 8.4), 50 mM KCl, 2.5 mM MgCl2, 0.5 mM from each dNTP, 0.01 M DTT, and 200 U SuperScript II RT in total volume of 20 [mu]l for 1 h at 42oC. Approximately 2 [mu]l from the cDNA sample were subjected to PCR amplification with 50 ng of primers SPD6 and SPD2 in a total volume of 25 [mu]l that consisted of 50 mM KCl, 10 mM Tris-HCl (pH 9.0), 1.0 mM MgCl2, 0.01 % Triton X-100, 0.2 mM from each dNTP and 0.5 U Taq Polymerase (Promega). PCR was carried out on the GeneAmp 9600 thermocycler (Perkin Elmer) for 35 cycles that consisted of 10 s denaturation at 94oC and 20 s annealing at 55oC.
Sequences of the forward degenerate primer SPD3 and the reverse gene specific primer SPD2 are given in Table 1 . In order to reduce the complexity of the PCR template, we amplified cosmid clone G2 DNA. Amplification reaction proceeded in 20 mM Tris-HCl (pH 8.8), 10 mM KCl, 10 mM (NH4)2SO4, 2.0 mM MgSO4, 0.1% Triton X-100, 0.2 mM from each dNTP, 100 ng from each primer, 0.03 ng cosmid DNA and 2.0 U 9oNm DNA Polymerase (New England Biolabs) in a total volume of 100 [mu]l. The PCR reaction was carried out in 35 cycles that consisted of 10 s at 94oC, and 30 s at 53oC and 30 s at 72oC. Under these conditions, we were able to amplify a single 1.3 kb fragment. Sequence analysis reveals that the 3' end of primer SPD3 anneals at nucleotide position 589 (Fig. 2 ).
Genomic DNA was amplified with PCR using SPD5 and SPD7 primers. We used a total of 100 [mu]l per reaction which included 1* PCR buffer (10 mM Tris-HCl pH 8.8, 50 mM KCl, 1.25 mM MgCl2 , 0.01% gelatin and 0.1% Triton) and 100 ng of each primer, 200 ng of template DNA, 0.2 mM of each dNTPs and 1 U Taq polymerase (AmpliTaq, Perkin Elmer). Amplification conditions consisted of 2 min initial denaturation at 94oC followed by 29 cycles of denaturation at 94oC for 15 s, annealing at 59.5oC for 30 s and extension at 72oC for 1 min. The amplified products were separated on 1.5% agarose gel, stained with ethidium bromide and photographed under UV light.
Two primers of SPD9 and SPD10 were used to amplify a 138 bp fragment that contains the HOXD13 CA-repeat polymorphism. The PCR amplification was carried out as described above but modified with 1.0 mM MgCl2 in a total of 25 [mu]l reaction mixture. The PCR conditions were as follows: initial denaturation at 94oC for 2 min and 32 cycles of 94oC for 10 s, 54oC annealing for 30 s, with no extension. The amplified product was separated on a 7% denaturant polyacrylamide gel, silver stained (19 ) and photographed.
The authors would like to express their sincere appreciation to the family members of the two SPD families for their participation in this study. We would like to thank Dr Edoardo Boncinelli for the gift of the G2 and P19 probes, Dr William Upholt for his help and comments and Ms Ming Li at the UCHC Molecular Core Facility for her help with automated sequencing. This work was supported by a Faculty Research Grant from the University of Connecticut Health Center.
1 Winter, R.M. and Tickle, C. (1993) Syndactylies and polydactylies: Embryological overview and suggested classifications. Eur. J. Hum. Genet., 1, 96-104.MEDLINE Abstract
2 Sayli, B.S., Akarsu, A.N., Sayli, U., Akhan, O., Ceylaner, S. & Sarfarazi, M. (1995) A large Turkish kindred with syndactyly type II (synpolydactyly). 1. Field investigation, clinical and pedigree data. J. Med. Genet.,32, 421-434.MEDLINE Abstract
3 Akarsu, A.N., Akhan, O., Sayli, B.S., Sayli, U., Baskaya, G. and Sarfarazi, M. (1995) A large Turkish kindred with syndactyly type II (synpolydactyly). 2. Homozygous phenotype? J. Med. Genet., 32, 435-441.MEDLINE Abstract
4 Sarfarazi, M., Akarsu, A.N. and Sayli, B.S. (1995) Localization of the syndactyly type II (synpolydactyly) locus to 2q31 region and identification of tight linkage to HOXD8 intragenic marker. Hum. Mol. Genet.,4, 1453-1458.MEDLINE Abstract
5 Dib, C. et al. (1996) The final version of the Genethon human genetic linkage map. Nature,380, 152-154.MEDLINE Abstract
6 Dolle, P. et al. (1993) Disruption of the Hoxd-13 gene induces localized heterocrony leading to mice with neotenic limbs. Cell,75, 431-441.MEDLINE Abstract
7 Dolle, P., Izpisua-Belmonte, J.C., Boncinelli, E. and Duboule, D. (1991) The Hox-4.8 gene is localized at the 5' extremity of the Hox-4 complex and is expressed in the most posterior parts of the body during development. Mech. Dev., 36, 3-13.MEDLINE Abstract
8 D'Esposito, M., Morelli, F., Acompora, D., Miggliaccio, E., Simeone, A. and Boncinelli, E. (1991) EVX2, a human homeobox gene homologous to the even-skipped segmentation gene, is localized at the 5'end of HOX4 locus on chromosome 2. Genomics,10, 43-50.MEDLINE Abstract
9 Acampora, D. et al. (1989) The human HOX gene family. Nucleic Acids Res., 17, 10385-10402.MEDLINE Abstract
10 Rogina, B. and Upholt, W.B. (1993) Cloning of full coding chicken cDNAs for the homeobox-containing gene Hoxd-13. Nucleic Acids Res., 21, 1316-1320.MEDLINE Abstract
11 Kappen, C., Schughart, K. and Ruddle, F. (1989) Organization and expression of homeobox genes in mouse and man. Ann. N.Y. Acad. Sci.,567, 243-252.MEDLINE Abstract
12 Johnson, R.L., Riddle, R.D. and Tabin, C.J. (1994) Mechanisms of limb patterning. Curr. Opinion Genet. & Dev., 4, 535-542.
14 Burglin, T.R. A comprehensive classification of homeobox genes. `Guidebook To The Homeobox Genes' Duboule, D., ed. (1994) Sambrook and Tooze publications at Oxford University Press, N.Y.
15 Magoulos, C., and Hickey, D. (1992) Isolation of three novel Drosophila melanogaster genes containing repetitive sequences rich in GCN triplets. Genome, 35, 133-139.
16 Pyrc, J.J., Moberg, K.H. and Hall, D.J. (1992) Isolation of novel cDNA encoding a zinc-finger protein that binds to two sites within the c-myc promoter. Biochem., 31, 4102-4110.
17 Birnboim, H.C. and Doly, J. (1979) A rapid alkaline extraction procedure for screening recombinant plasmid DNA. Nucleic Acids Res.,7, 1513-1522.MEDLINE Abstract
18 Sanger, F., Nicklen, S., and Coulson, A.R. (1977) DNA sequencing with chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA., 74, 5463-5467.MEDLINE Abstract
19 Bassam, B.S. and Caetano-Anneles, G. (1993) Silver staining of DNA in polyacrylamide gels. Protocols Biotechnol., 42, 181-188.
20 Xiaoqiu, H. (1994) n global sequence alignment. Comp Appl. Biosci., 10, 227-235.
*To whom correspondence should be addressed
This page is maintained by OUP admin. Last updated Thu Oct 31 15:25:15 GMT 1996. Part of the OUP Journals World Wide Web service.Copyright Oxford University Press, 1996