J. Biochem. 110, 407-411 (1991)

Genetic Linkage of Lung Cancer-Associated Mspl Polymorphisms with Amino Acid Replacement in the Heme Binding Region of the Human Cytochrome P450IA1 Gene1 Shin-ichi Hayashi,* Junko Watanabe,* Kei Nakachi,** and Kaname Kawajiri* 'Department of Biochemistry and "Department of Epidemiology, Saitama Cancer Center Research Institute, Ina-machi, Kitaadachi-gun, Saitama 362 Received for publication, April 25, 1991

Individuals with high genetic risk of lung cancer had previously been identified by Mspl polymorphisms of the cytochrome P460IA1 gene. In the present study we analyzed the structures of individual P450IA1 genes by PCR direct sequencing of genomic DNA of each genotype raised by the Mspl polymorphisms, which were ascribed to a single point mutation in the 3'-flanking region. We then found a novel point mutation in the coding region of the gene which results in the substitution of He for Val at residue 462 in the heme binding region. We further analyzed the genetic association between this amino acid replacement and Mspl polymorphisms in the general population, using a new method to detect polymorphisms not recognized by restriction enzymes. The results showed that there are at least two forms of human P450IA1 protein with different primary structures and that one of the forms is closely linked with the lung cancer-susceptible genotype of Mspl polymorphisms. Thus Mspl polymorphisms, which are associated with increased risk of lung cancer, are linked to at least one amino acid substitution, which gives an important clue, at the molecular level, toward elucidation of increased susceptibility to lung cancer.

A large proportion of human cancers are known to be caused by synthetic or natural chemical compounds in the environment (1, 2). Many chemical carcinogens are metabolically activated to forms that have deleterious effects on organisms (3), and this metabolic activation is an obligatory initiation step in human chemical carcinogenesis. The microsomal electron transport system, including cytochrome P450s, plays the most important role in oxidation of chemical carcinogens (4, 5), and involves a variety of isozymes of P450. This oxidative activation shows genetic variation (4, 6), which may be responsible for individual differences in susceptibility to chemical carcinogenesis. Cytochrome P450IA1 is important in the initiation of lung cancer because it is responsible for the activation to mutagens of benzo f a] pyrene and other aromatic hydrocarbons in cigarette smoke (7, 8). We recently found a close correlation between development of lung cancer and Mspl polymorphisms in the 3'-flanking region of the P450IA1 gene, where three genotypes were determined the predominant homozygote (genotype A), the heterozygote (genotype B), and a homozygous rare allele (genotype C) (9). Namely, the incidence of genotype C of the P450IA1 gene was 21.2% in lung cancer patients, but only 10.6% in healthy controls. This genetic change was specifically correlated with increased risk of the

Kreyberg I type of lung cancer, which is closely associated with cigarette smoking, but not with the Kreyberg II type. Furthermore, our case-control study showed that the relative risk of genotype C was much higher (7.31-fold) than those of genotypes A and B at low levels of cigarette consumption (20). Therefore, it is indispensable to investigate whether this Mspl polymorphism is genetically associated with differences in protein structure or gene expression of P450IA1. On the basis of this investigation, the interindividual difference in the basal or induction response of P450IA1 can be elucidated. In the present study we analyzed individuals with genotypes A and C by PCR direct sequencing. We found that a novel point mutation in genotype C resulted in an amino acid replacement in the heme binding region of P450IA1 and also showed that at least two forms of P450IA1 protein with different primary structures exist among humans. We further determined the genetic association of this novel mutation with Mspl polymorphisms in the general population, developing a new detection method to identify the polymorphisms not recognized by restriction enzymes. It was found that these two loci, causing the observed DNA polymorphisms, were very closely linked.

1 This work was supported in part by Grants-in-Aid for Scientific Research from the Ministry of Education, Science and Culture of Japan, and a research grant from the Ministry of Health and Welfare of Japan. Abbreviations: P450IA1, previously called P450c; PCR, polymerase chain reaction; RFLPs, restriction fragment length polymorphisms; XRE, xenobiotic responsive element; BTE, basic transcription element.

Materials—Restriction endonucleases were purchased from Takara Shuzo (Kyoto). [a- 3B S]dATP (1,000 Ci/ mmol) and [y-"P]ATP (5,000 Ci/mmol) were obtained from Amersham (Arlington Heights, IL). All other reagents were of the highest quality commercially available. Isolation and Sequence Analysis of the 3' -Region of the P450IA1 Gene—A human gene library of Charon 4A

Vol. 110, No. 3, 1991 Downloaded from https://academic.oup.com/jb/article-abstract/110/3/407/819242 by University of Durham user on 10 January 2018

EXPERIMENTAL PROCEDURES

407

408

S. Hayashi et al.

bacteriophage was re-screened using an EcoBI-EcoBl 1-kbp fragment of the 3'end of the clone A.hP450mc-l as a probe (11). The clone AhP450mc-2 which contained the 3'-flanking region of the gene was isolated, and its EcoBlEcoBl 4-kbp fragment was subcloned in pUC18 (named as phP450mc-2). Uni-directionally deleted subclones generated by the Erase-a-Base system (Promega, Madison, WI) were used to obtain sequence information on the 3'-flanking region. Sequencing was carried out for both strands by the chain-termination method (12). Detection of Mspl Polymorphisms—Individual DNAs were isolated from peripheral lymphocytes in blood samples obtained from a cohort of 2,500 Japanese persons all over 40 years old. The genotypes of the P450IA1 gene ascribed to the Mspl site were identified as RFLPs by the polymerase chain reaction (PCR) (13) and were in complete agreement with our previous Southern blotting (9). The PCR-amplified DNA fragments including the polymorphic site were digested with Mspl and subjected to electrophoresis in 1.8% agarose gel. PCR Direct Sequencing of Individual Genomic DNAs— DNA fragments of 0.5 to 1.5 kbp were prepared by PCR-amplification and used as templates for direct sequencing. Synthesized 21-nucleotide oligomers were endlabeled with [y- 32 P]ATP by kination, and 4pmol of the labeled primers were used for sequencing reaction together with 200 ng of purified PCR template using a Sequenase sequencing kit (USB, Cleveland, OH). Oligo-nucleotides for sequencing primers and PCR primers (named as Cl to

C51) were synthesized on an Applied Biosystems Model 381 DNA sythesizer. Detection of Polymorphisms Not Recognized by RFLP— Two oligonucleotides of 20mer (primer 2a, primer 2b), both of which contained the polymorphic site at the 3' end, were synthesized and each of them was used as a primer for PCR-amplification together with another strand of 21 mer primer (primer 1) which is located about 200 bp upstream of a polymorphic site detected by sequencing. PCR was performed by 25 cycles under the following conditions: 1 min at 95°C for denaturation and 1 min at 70'C for primer annealing and primer extension. The other conditions were as described by Saiki et al. (13). The PCR products were then subjected to electrophoresis in 1.8% agarose gel. RESULTS AND DISCUSSION Structural Analysis of the 3'-Flanking Region of P450IA1—Previously we cloned the human P450LA1 gene and determined the nucleotide sequence (11), although the cloned gene did not include the 3'-flanking region where the polymorphic Mspl site is located. To identify the Mspl site and analyze the 3'-flanking region of the gene in this study, we re-screened a human genomic library and obtained clone AhP450mc-2 for analysis. A 4-kbp fragment including the 3'-flanking region was subcloned in a plasmid vector (phP450mc-2) and then sequenced. Based on the sequence information, primers were synthesized, and DNAs of types A and C were subjected to PCR direct sequencing. Figure 1

5836

ACAATCCTTCTATTCTAGCCTGCATTGAGCTTGCATGCTTGCATAAGAGC

5885

5886

TTAAGAAACCATTGATTTAATGTAATAGGGAAAATTCTAACCCAGGTATC

5935

59 36

CAAAAATGTGTAAGAACAACTACCTGAGCTA)ftATAA^GATATTGTTCAGA

5985

B

2

3

M

1

2 3

340 >•

[ 200 I 140

PolyA Signal

1

AAATCCTATAGGTGGAGATTTTTTGAATCATAAATGATTCATCACTCGTC

50

51

TAAATACTCACCCTGAACCCCATTCTGTGTTGGGTTTTACTGTAGGGAGG

100

101

AAGAAGAGGAGGTAGCAGTGAAGAGGTGTAGCCGCTGCACTTAAGCAGTC

150

151

TGTTTGAGGGACAAGACTCTATTTTTTGAGACAGGGTCCCCAGGTCATCC

200

201

AGGCTGGAGTGCACTGGTACCATTTTGTTTCACTGTAACCTCCACCTCCT

250

251

GGGCTCACACGATTCTCCCACCTCAGCCTCTGAGTAGTTGGGGCCGCCAG

C47

Msp\ISma\ 300

CB1

301

ACGCCACCACAGCTTTTTTTTTTTTTTTTTTTrrTTTTTGTAGAGATGGG

350

351

GTTTCACCATGTTGCCCAGGCTGGTCTCAAACTCCTGAGCTCAAGTGATC

400

401

CACCTGCCTCAGCCTCCCAAAGTGCTGGGATTACAGGCATGAGACAAGAC -*

4 50

451

TCCTAATCACTGTGCTGTCTTAGCGCCCTCTCTAACTTATCACAAATTGA

C44

500

Mspl

Smal

Fig. 1. Structural analysis of the 3 -region of the P450IA1 gene. (A) Nucleotide sequence of the 3'-flanking region of the gene. Clone /lhP450-mc2 was isolated, and aubclone phP450mc-2 was sequenced as described under "EXPERIMENTAL PROCEDURES." PCR direct sequencing was carried out using the primers indicated by broken lines, and the polymorphic site was identified. The nucleotides with numbers up to 5985 in the margins were transcribed in mRNA, and untranscribed nucleotides are numbered from 1. (B), RFLPs of PCR-amplified fragments by Mspl or Smal. The primers indicated by solid lines in (A) were used for PCR amplification, and the PCR products from type A, B, and C DNAs were digested with Mspl or Smal, and subjected to agarose gel electrophoresis. Lanes 1, 2, 3, and M are type A (ml/ml), B (ml/ m2), C (m2/m2), and size markers (SV40Hi/idlll), respectively. Allele ml and m2 are defined by the absence and presence, respectively, of the one Mspl site detected by RFLPs in our previous paper (9). J. Biochem.

Downloaded from https://academic.oup.com/jb/article-abstract/110/3/407/819242 by University of Durham user on 10 January 2018

409

Genetic Polymorphisms of Cytochrome P450IA1 phP45Omc-2 1 kbp

Fig. 2. Structural analysis of individual P450IA1 genes by PCR-dlrect sequencing. The DNA fragments amplified by PCR are shown by horizontal bars. Closed circles and squares show the primers used for the two opposite directions in PCR. These PCR products were sequenced directly by the chain-termination method using synthesized sequencing primers. Coding sequences are indicated by boxes, and leader and trailer sequences of mRNA are indicated by boxes of lower height. The solid arrow shows the site of Mspl polymorphisms, and the open arrow shows that of the novel mutation in the coding region. Solid triangles, the open circle, and the open triangle indicate the locations of the XRE, BTE, and TATA box, respectively.

B

Fig. 3. Novel polymorphic mutation in the coding Qhi region. DNAs from individuals were examined by PCR-direct sequencing. A, B, and C show the results for an Be/lie homozygote, De/Val heterozygote, and Val/Val homozygote, re- Ue,Val spectively. The point mutation is indicated by an arrow.

Ala C

-6

-4

-2

+2

+4

+6

T

References

Fish

F G M D K R R C I G E A

G R

(18)

Rabbit

F G L G K R K C I A E T

G R

(19)

Mouse

F G L G K R K C I G E T

G R

(20)

Sat

F G L G K R K C I G E T

G R

(21)

(lie)

F G M G K R K C I G E T

AR

(11)

Human (Val)

F G M G K R K C I G E T

AR

Human

A

Fig. 4. Comparison of the ami no acid sequences of the heme binding region in the P450IA1 subfamily. The heme-binding cysteine residue is designated as 0 (22). The amino acid residues at the polymorphic position are boxed. The sequences of all members of the LAI subfamily identified so far are shown (23).

summarizes the results. Replacement of thymine in genotype A by cytosine in genotype C was observed at the 264th base downstream from the poly A additional signal, forming an Mspl or Smal site in genotype C. For confirmation of this observation, PCR-amplification was carried out with DNAs of types A, B, and C as templates and the oligonucleotide primers indicated in Fig. 1A. The products were then digested with Mspl or Smal. The fragment amplified from type A DNA gave only a single undigested band at the position of 340 bp, the fragment from C type gave bands of digested DNA at 200 and 140 bp, while the fragment from Vol. 110, No. 3, 1991 Downloaded from https://academic.oup.com/jb/article-abstract/110/3/407/819242 by University of Durham user on 10 January 2018

G

C

T

A

G

C

T

A

G

B type gave three bands at 340, 200, and 140 bp (Fig. IB). This procedure was used for determining the genotypes of P450LA1 in a large number of individuals, and allowed identification of persons at genetically high risk of lung cancer. Amino Acid Replacement in the Heme Binding Region of P450IA1—For determination of whether there was any other mutation site in the P450IA1 gene linked with Mspl polymorphism, the regulatory regions in the 5'-flanking and coding regions of the P450IA1 gene were sequenced directly using PCR-amplified fragments from the DNAs of types A and C (Fig. 2). In the 5'-flanking region including the XRE (xenobiotics responsible element) (2 4), BTE (basic transcription element) {15), and TATA box, the DNA sequences of type A and C were identical. However, we found a difference of one base at position 4889 in the 7th exon. As shown in Fig. 3, adenine in type A was replaced by guanine in many individuals of type C. This novel point mutation resulted in replacement of He by Val at residue 462 in the HR2 region (16), which was well conserved among P450 families. The cysteine in this region has been shown to be the heme-binding thiolate ligand (17, 22), and thus this region is essential for the catalytic activity of this enzyme. As this polymorphic position is conserved as He among members of the IA1 subfamily (Fig. 4), the substitution of Val may affect the catalytic activity of this enzyme.

S. Hayashi et al.

410

Fig. 5. Detection of the point mutation of Ile-Val polymorphism. (A) Scheme 4890 4900 of PCR for the detection. Two 5'ATTGCCCGCTGGGAGGTCTT3' lie type primers (primer 1 and primer 3'TAACGGGCGACCCTCCAGAA5' 2a, or primer 1 and primer 2b) * were used for each DNA samprimer 2a ple, and PCR was performed according to the method de4900 scribed in "EXPERIMENTAL 4890 The point type PROCEDURES.' GTTGCCCGCTGGGAGGTCTT3' mutation site is indicated by CAACGGGCGACCCTCCAGAA5' asterisks. (B) Representative primer 2b results of PCR to identify IleVal polymorphisms. A, Ile/He homozygote; B, Ile/Val heterozygote; C, Val/Val homozygote. Primers 1 and 2a were used for lane 1 and primers 1 and 2b for lane 2; lane M, size markers U-.EcoRI-.ffj/idIII). 7th exon

6th tntron 4730 4740 5•GAACTGCCACTTCAGCTGTCT3'

primer 1

B 210 bp 1

2

M

TABLE I. Estimated frequencies of combined genotypes. Numbers of subjects examined are shown in parentheses. Genotype ml/ml ml/m2 m2/m2 Total lie/He 0.433 0.144 0.005 0.582 (42/43) (15/47) (2/45) De/Val 0.010 0.278 0.049 0.337 (1/43) (29/47) (21/45) 0.0 Val/Val 0.029 0.052 0.081 (0/43) (3/47) (22/45) Total1 0.443 0.451 0.106 1.000 (43) (47) (45) (135) "These frequencies were observed in our previous study (10).

The RFLP method cannot be used to detect this polymorphism because there is no suitable restriction site. Therefore, for use in screening, we developed the new detection method described in "EXPERIMENTAL PROCEDURES." The primers used in this method are shown in Fig. 5A. Figure 5B shows the clear profiles that identified the two homozygotes (He/He and Val/Val) and the heterozygote (He/ Val). The combination of primers 1 and 2a gave a PCR product of 210 bp (lane Al) when (lie/He) was used as a template. The combination of primers 1 and 2b gave no PCR product (lane A2): the product appeared only with the combination of primers 1 and 2b with (Val/Val) as template (lanes Cl, C2), and with both combinations with (lie/ Val) as template (lanes Bl, B2). The results by this new method were completely consistent with those obtained by direct sequencing. Linkage Analysis in the General Population—We found the DNA polymorphisms at two loci of the P450IA1 gene to be due to point mutations. Therefore, we next investigated the genetic association between these two loci in the general population. In our previous study on Mspl polymorphisms, we observed frequencies of 0.443, 0.451, and 0.106 for

genotypes A {ml/ml),

B (ml/m2), and C (m2/m2),

respectively, in 375 healthy controls (20). Our next problem was to determine the frequencies of the genotypes (He/ He), (He/Val), and (Val/Val) in the populations with genotypes A, B, and C. To do this, we randomly selected 43

and 47 individuals from among those with genotypes A and B, respectively, and all the 45 individuals with genotype C observed so far. The observed frequencies of (lie/He), (lie/ Val), and (Val/Val) in the groups of genotypes A, B, and C were then weighted by the frequencies of A, B, and C. The estimated frequencies are shown in Table I as a 3 x 3 genotype table. From the table we could calculate the respective frequencies f(ml; lie), f(m2; lie), f(ml; Val), and/(m2; Val) of the genes ml with He at residue 462, m2 with lie, ml with Val, and m2 with Val, assuming the Hardy-Weinberg equilibrium for f(ml; lie) and f(ml; Val) and also for f(m2; lie) and f{m2; Val) weighted by the observed gene frequencies /(ml) = 0.668 and f(m2) = 0.332. The gene frequencies were thus estimated as f(ml; lie) =0.661, f(m2; He) =0.092, f(ml; Val) = 0.007, f(m2; Val) =0.240, /(He) = 0.753, and /(Val) = 0.247. The frequencies expected from Table I agreed well with observed values within a range of 0.03. Linkage disequilibrium D (24) was then calculated to be 0.158, and the linkage coefficient R described by Hill and Robertson (25) to be 0.778, showing that these two loci are closely linked. To understand the genetic association more clearly, we calculated the occurrence of linkage, L e., p(ml; ne)=/(ml;Ue)//"(m2)=0.990,p(mi;Val) = 0.010,p(m2; lie) =0.272, p(m2; Val) =0.723, p(De; ml) = f(ml; Ee)/ Alle) = 0.878, p(Ile; m2) =0.122, p(Val; ml) = 0.028, and p(Val; m2) =0.972. These results clearly showed 99% linkage of ml with He and 97% linkage of Val with m2. We analyzed the gene structures of the individuals with different susceptibility to lung cancer identified by Mspl polymorphisms of P450IA1. As a result, a novel polymorphic site was then found in the DNA sequences of these individuals which results in an amino acid mutation in the heme binding region. The existence among human of at least two forms of P450IA1 protein in association with Mspl polymorphisms was thus revealed. To clarify this genetic association, we further assessed the frequencies of the genotypes with a combination of these two polymorphic loci among the general population using the new detection method and showed that these two loci were very closely linked. Recently the reasons for phenotypes of extensive and poor metabolism of debrisoquine have been explained by DNA polymorphisms of P450IID at plural polymorphic sites (26, 27). However, there is no previous report of J. Biochem.

Downloaded from https://academic.oup.com/jb/article-abstract/110/3/407/819242 by University of Durham user on 10 January 2018

Genetic Polymorphisms of Cytochrome P450IA1 analysis of the linkage between multiple polymorphic loci of human P450s in the general population. Such analyses are essential for estimating susceptibilities to drugs and chemical carcinogens, because the susceptibility or risk must be evaluated as the ratio of the genotype frequency in the susceptible subpopulation (or patients) to that in the general population (28). It is, of course, important to characterize the effect of this novel mutation and Mspl polymorphisms on phenotypic expression of P450IA1 at the molecular level. We are now investigating these successive questions, which involve biological assays of the activity of the enzyme with the mutation, influence of Afspl polymorphisms on expression, association with arylhydrocarbon hydroxylase activity together with epidemiological analysis of cancer risk. We are grateful to Dr. Yusaku Tagaahira for his support during this study, and Nahomi Shinoda for technical assistance. REFERENCES 1. Nagao.M. &Sugimura,T. (1978) Annu. Rev. Genet 12,117-159 2. Doll, R. &Peto, R. (1981) The Causes of Cancer, pp. 1256-1260, Oxford University Press, Oxford, New York 3. Heidelberger, C. (1975) Annu. Rev. Biochem. 44, 79-121 4. Guengerich, F.P. (1988) Cancer Res. 48, 2946-2954 5. Conney, A.H. (1982) Cancer Res. 42, 4875-4917 6. Gonzalez, F.J. (1988) PharmacoL Rev. 40, 243-288 7. Aoyama, T., Gonzalez, F., &Gelboin, H.V. (1989) Mol. Carcinog. 1, 253-259 8. McManus, M.E., Burgess, W.M., Veronese, M.E., Huggett, A., Quattrochi, L.C., & Tukey, R.H. (1990) Cancer Res. 50, 33673376 9. Kawajiri, K., Nakachi, K., Imai, K.( Yoshii, A., Shinoda, N., & Watanabe, J. (1990) FEBS Lett. 263, 131-133 10. Kawajiri, K., Nakachi, K., Imai, K., Hayashi, S., & Watanabe, J. (1991) in Proceedings of the 21th International Syjnposium of the Princes Takamatsu Cancer Fund (Emster, L., ed.) Japan Sci. Soc. Press, Tokyo/Taylor & Francis, London, in press 11. Kawajiri, K., Watanabe, J., Gotoh, O., Tagashira, Y., Sogawa,

Vol. 110, No. 3, 1991 Downloaded from https://academic.oup.com/jb/article-abstract/110/3/407/819242 by University of Durham user on 10 January 2018

411 K., & Fujii-Kuriyama, Y. (1986) Eur. J. Biochem. 159, 219-225 12. Sanger, F., Nicklen, S., & Coulson, A.R. (1977) Proc Nad. Acad Sci. U.S.A. 74, 5463-5467 13. Saiki, R.K., Gelfand, D.H., Stoffel, S., Scharf, S.J., Higuchi, R., Horn, G.T., Mullis, K.B., & Ehrlich, H.A. (1988) Science 239, 487-491 14. Fujisawa-Sehara, A., Sogawa, K., Yamane, M., & FujiiKuriyama, Y. (1987) Nucleic Adds Res. 15, 4179-4191 15. Yanagida, A., Sogawa, K., Yasumoto, K., & Fujii-Kuriyama, Y. (1990) Mol Cell. BioL 10, 1470-1475 16. Gotoh, 0., Tagashira, Y., Iizuka, T., & Fujii-Kuriyama, Y. (1983) J. Biochem. 93, 807-817 17. Poulos, T.L., Finzel, B.C., Gunsalus, I.C., Wagner, G.C., & Kraut, J. (1986) J. BioL Chan. 260, 16122-16130 18. Heilmann, L.J., Sheen, Y.Y., Bigelow, S.W., & Nebert, D.W. (1988) DNA 7, 379-387 19. Okino, S.T., Quattrochi, L.C., Barnes, H.J., Osanto, S., Griffin, K.J., Johnson, E.F., & Tukey, R.H. (1985) Proc. NatL Acad. Sci. U.S.A. 82, 5310-5314 20. Kimura, S., Gonzalez, F., & Nebert, D.W. (1984) J. BioL Chem. 259, 10705-10713 21. Sogawa, K., Gotoh, O., Kawajiri, K., & Fujii-Kuriyama, Y. (1984) Proc. Natl. Acad. Sci. U.S.A. 81, 5066-5070 22. Kawajiri, K., Gotoh, 0., Sogawa, K., Tagashira, Y., Muramatsu, M., & Fujii-Kuriyama,Y. (1984) Proc. NatL Acad. Sci. U.S.A. 81, 1649-1653 23. Nebert, D.W., Nelson, D.R., Adesnik, M., Coon, M.J., Estabrook, R.W., Gonzalez, F.J., Guengerich, F.P., Gunsalus, I.C., Johnson, E.F., Kemper, B., Levin, W., Phillips, I.R., Sato, R., & Waterman, M.R. (1989) DNA 8, 1-13 24. Vogel, F. & Motulsky, A.G. (1986) in Human Genetics, pp. 162164, Springer-Verlag, Berlin, Heidelberg, New York, London, Paris, Tokyo 25. Hill, W.G. & Robertson, A. (1968) Theor. Appl. Genet. 38, 226231 26. Kagimoto, M., Heim, M., Kagimoto, K., Zeugin, T., & Meyer, U.A. (1990) J. BioL Chem. 265, 17209-17214 27. Gough, A.C., Miles, J.S., Spurr, N.K., Moss, J.E., Gaedigk, A., Eichelbaum, M., & Wolf, C.R. (1990) Nature 347, 773-776 28. Khoury, M.J., Adams, M.J., Jr., & Flanders, W.D. (1988) Am. J. Hum. Genet 42, 89-95