M E T H O D S T O A N A L Y S E T H E H U M A N G E N O M E

ROBERT WILLIAMSON AND SUSAN M. DARLING Department of Biochemistry, St Mary’s Hospital Medical School, University of London,

London W2 1PG

SUMMARY. There is much interest in the use of gene-specific probes for the study of dysfunction in human pathology. For the haemoglobinopathies, globin gene recombinants (either prepared from DNA sequences complementary to messenger RNA, or from genomic DNA) are used to determine whether globin genes are present, whether they are expressed in the nucleus, and whether they are correctly processed to give functional mRNAs. This has not only allowed a fuller understanding of the molecular aetiology of the thalassaemias, but also permitted antenatal diagnosis both by direct analysis of the gene lesion, and by linkage analysis using adjacent genes. Similar approaches are being applied to many other single gene defects.

There are, however, other possible ways to study human hereditary disease using recombinants. It is now feasible to use random human chromosome-specific sequences to establish a linkage map for the entire human genome. Such a map may then be used either to determine the chromosomal localisation of any “single gene” phenotype by linkage analysis, or to study the contribution of different genes to a complex phenotype determined by several genes, as in multifactorial disease.

GLOBIN GENE ANALYSIS

IT was extremely difficult to study human genetic disease at the level of the gene before the advent of recombinant DNA technology. Only a very few messenger RNAs for structural genes were available, such as those for the globins, and even these were never completely pure (Forget et al., 1974; Benz, Swerdlow and Forget, 1975; Tolstoshev et al., 1976). In spite of this, by 1974 it proved possible to analyse a-thalassaemia at the gene level using a partially purified cDNA for a-globin as a gene-specific probe, and in this way to demonstrate that the disease is due to a gene deletion (Ottolenghi et al., 1974; Taylor et al., 1974). However, each experiment required human reticulocyte RNA which is hard to obtain and must be purified extensively. By its very nature such a technique can only be applied to those genes where a protein product is present in large amounts in a human tissue which can be obtained ethically in the course of normal treatment of a disease.

The first human genes to be prepared as purified and characterised I93

194 ROBERT WILLIAMSON AND SUSAN M . DARLING

recombinants in bacterial plasmids were those for the human globin cDNA sequences complementary to globin mRNAs (Little et al., 1978; Wilson et al., 1978). Using these, the structure of the normal globin gene loci were determined, including the intergene distances and the presence of intervening sequences (Orkin, 1978; Bernards et al., 1979). Since cDNA clones could also provide probes for mRNA and nuclear precursors to mRNA, the processing of the globin gene transcripts was also studied (Courtney and Williamson, 1979). Finally, cDNA clones provide gene-specific probes for genomic DNA clones, which can be up to tens of thousands of base pairs in length (Lawn et al., 1980). These have been sequenced completely, and the entire base sequence surrounding the a-globin and P-globin genes, including related pseudogene sequences, have now been determined (Efstratiadis et al., 1980).

At the same time as the organisation and function of the normal globin genes were determined, groups also studied the reasons for the loss of normal function in thalassaemia. In most cases the cr-thalassaemias are indeed due to a gene deletion. However, there are a few cases of non-deletion forms of a- thalassaemia characterised by reduced production of or-globin, although both a-genes are intact (Higgs et al., 198 1).

The P-thalassaemias are more heterogeneous, both in aetiology and symptoms, and therefore of more interest to the molecular biologist. A small number (the GIP-thalassaemias) are usually due to a gene deletion, although if the deletion is large the lack of P-globin gene expression may be compensated by increased y-globin mRNA synthesis, indicating a possible involvement of non-coding sequences within the P-gene locus in the switching off of the y-globin genes (Bernards. and Flavell, 1980; Little, 198 1). However, the vast majority of P-thalassaemias seem to result from mutations that interfere with mRNA processing and stability (Spritz et al., 198 1 ; Westaway and Williamson, 198 l), or “nonsense mutations” (Chang and Kan, 1979; Orkin and Goff, 1981). The defect in sickle cell disease has been demonstrated to be due to a single nucleotide change at the codon specifying amino acid six of P-globin, as expected, by sequencing (Marotta et al., 1977).

ANTENATAL DIAGNOSIS OF HAEMOGLOBINOPATHIES

Antenatal diagnosis for known specific gene defects can be simple if the disease is due to a gene deletion-a laboratory can investigate whether the gene is present or absent in DNA isolated from fetal cells providing a DNA probe specific to the gene in question is available (Little, 1981). If the defect is due to a single base change, it is difficult to identify directly unless it occurs at a site specific for a restriction enzyme. Restriction enzymes recognise and cleave at short, specific sequences 4, 5 or 6 bases in length (Malcolm, 198 l), and generate fragments from several hundred to several thousand base pairs long. These fragments can be separated according to their size by electrophoresis on an agarose gel, and then “ blotted ” onto a sheet of nitrocellulose using a technique first described by Southern (1975). These fragments can then be hybridised

GENOME ANALYSIS 195

with radiolabelled gene-specific probes, and the sequences containing the gene of interest identified.

Where a gene-specific probe is available, and the nucleotide site of the lesion causing the genetic disease is known, it may be possible to perform a direct analysis, rather than relying on linkage. This is now possible for sickle cell disease, where the base change at codon six causes the loss of a site for the restriction enzyme DdeI that is completely predictive for the disease (Wilson et al., 1982).

Since not every base sequence is the same in different persons, corresponding sequences sometimes occur on fragments of differing length if a restriction enzyme recognition site is lost or gained through mutation. In this way, Kan and Dozy (1978) were able to show that the sickle /Is-globin gene is usually i n a 13,000 base pair fragment after DNA is digested with HpaI, while the normal gene occurs in a fragment of about 7000 base pairs in length. In this case the defective gene is not being examined directly, but by linkage to a restriction site dimorphism. In some cases of /I-thalassaemia, Little et al. (1980) showed that restriction site polymorphisms in the y-globin genes, a distance of some 15-20 kilobase pairs from the /I-globin gene lesion, could be used for antenatal diagnosis.

This approach is only useful if informative base polymorphisms occur in the P-globin genes of the affected family, allowing the defective parental chromosomes to be distinguished. Even where the polymorphisms are informative, a previously affected (or normal) child must in general be alive to allow accurate assignment of the alleles to the correct linkage groups, and paternity must be assumed. Also, any linkage analysis must by definition involve a possible error if crossover has occurred between the defect and the linked DNA site in the parental chromosomes. Because of this, it is likely that future diagnosis will use new techniques which can recognise single base changes accurately with synthetic oligonucleotide probes linked to fluorescent or enzyme probes (Malcolm, 1982). These techniques should shorten the time of analysis from 2 weeks to 2 days, and require much less DNA. With radioactive probe hybridisation, cells from 10 ml of amniotic fluid are required, while the new methods only require one-tenth this amount. This should prove particularly useful when antenatal diagnosis is carried out on trophoblast material obtained during the first trimester of pregnancy (Williamson et al., 1981; Old et al., 1982).

ANALYSIS OF OTHER MONOGENIC INHERITED DISEASES

The concept of linkage provides a clue to a method for approaching the problem of genetic diseases where a single gene is involved, but no biochemical defect is known. Obvious examples of such diseases are Duchenne muscular dystrophy (X-linked recessive), cystic fibrosis (autosomal recessive), and Huntington’s chorea (autosomal dominant). In each case, it may be possible to look for a linkage between the phenotype and a restriction site polymorphism within an informative family.

196 ROBERT WILLIAMSON AND SUSAN M . DARLING

The frequency of restriction site polymorphisms is not known with accuracy. However, in a careful study of the DNA sequences surrounding the human globin genes in different individuals, Jeffreys (1979) calculated that two chromosomes in a single individual will differ once for every sequence of approximately 100 base pairs in length. If this finding is correct throughout the genome, then there will certainly be many restriction site polymorphisms which will be potentially useful in the study of inheritance. Assuming that the average restriction enzyme cuts DNA at a site specified by five base pairs (most cut at a four-base or six-base site; Malcolm, 1981), and that the probe used recognises only one single-copy DNA sequence, there will be a 20 per cent. chance that the restriction fragments generated from the two homologous chromosomes will differ. Obviously, it will only be necessary to screen a few people with a given probe and a single enzyme to identify useful polymorphisms (Wyman and White, 1980).

Of course, one would have to be very lucky indeed to find a restriction site polymorphism linked to any given phenotype with a totally random DNA sequence-it is necessary to be systematic in the choice of probes. This can be achieved in several ways-it is, for instance, possible to use as probes only transcribed gene sequences from a cDNA library prepared from an affected tissue, such as liver for phenylketonuria. It is also possible to construct a total and random genomic DNA library and isolate a sufficient number of clones to “ cover ” the genome with a complete linkage map (Botstein et al., 1980).

What will determine how many clones are needed for such a linkage map (Solomon and Bodmer, 1979)? The critical factor is the rate at which linkage is lost during meiosis or, put differently, the distance between meiotic crossover sites during homologous chromosome exchange. This can be measured directly by studying the number of chiasmata which are visible during meiotic prophase, and it has been demonstrated that there are approximately two to three such chiasmata per “ average ” chromosome (Kurnit and Hoehn, 1979). Therefore, the X chromosome might cross over at three sites, giving four “ linkage groups ” during a single meiosis. Of course, the sites of crossover are not the same for any two meioses, and although there are thought to be “ho t spots” for crossovers, it is assumed that crossing over is random.

An analysis of the entire human genome is possible with a relatively small number of sequence-specific probes, between 250-500 in all (Botstein et ul.). However, for studying a disease such as Duchenne muscular dystrophy, where the defect is known to be on the X-chromosome, only about ten polymorphisms are required, provided these are all seen with X chromosome-specific probes. Chromosome-specific clone libraries can be prepared using rodent-human cell hybrids containing single human chromosomes (Gusella et ul., 1980), or by using the fluorescence activated cell sorter as a chromosome sorter to prepare populations of single purified chromosomes (Davies et al., 1981). Such chromosome-specific libraries simplify linkage assignment by allowing neighbouring DNA sequences to be isolated and tested independently.

The demonstration of linkage depends upon the unambiguous determination of which of the two parental sequences is received by the offspring. This relies

GENOME ANALYSIS 197

upon two related variables-the frequency of restriction site polymorphisms in the population which allow the two chromosomes carrying the assigned fragment to be distinguished, and the frequency with which these occur in an informative way in the families under study. The first frequency is known to be high in the DNA surrounding the human globin genes, but of course may differ markedly in other regions of the genome. The second problem is more difficult. For the X-chromosome, every mother-son inheritance for which a restriction- site polymorphism can be studied is by definition informative-the mother will have two alleles (seen as two bands in a Southern blot), the son will inherit one of the two. However, for an autosome the problem is more complex, as both father and mother each contribute one of two possible genes to a child.

CHROMOSOME WALKING

A chromosomal assignment is only the first step in finding the molecular basis of a disease of unknown biochemical origin. Distant linkage may be conclusive for research purposes but cannot be used for antenatal diagnosis; an unacceptable rate of false negatives could arise from meiotic crossover. Therefore it may be necessary to " walk the genome " from the linked DNA sequence towards the defect itself (Dahl, Flavell and Grosveld, 1981). This is carried out by successive isolation of long genomic DNA fragments, followed by sub-cloning and re-screening the genomic library. If cDNA clones are used as " landmarks " it should be possible to move from one unique expressed sequence to another, checking each for function in relation to the disorder en route. If very long genomic clones (in cosmids) are used, or chromosomal inversions, it may be possible to move more quickly along the DNA (" jogging the genome "> (Maniatis et al., 1978; Dahl et al.). Hopefully the defect itself will be identified in this way, and then a direct diagnosis will be possible.

ANALYSIS OF POLYGENIC INHERITED DISORDERS

Linkage data will be extremely useful, whether for monogenic or polygenic disorders. Obviously, using this technique any assessable phenotype can be regarded as the sum of the influence of a set of genes. To date all attempts at gene research have been directed to single genes thought to be involved in genetic diseases-the haemoglobinopathies, hemophilia, phenylketonuria, etc. However, a polygenetic disease, even with environmental factors, can also be studied in this way, provided that an accurate assessment of some phenotype is possible. Therefore the linkage technique described should also be applicable to hypercholesterolaemia or hypertension, as the blood cholesterol or blood pressure can be measured objectively, and genes which are known to be involved such as the LDL receptor gene can be isolated. It will be more difficult to study schizophrenia or juvenile onset diabetes by this technique, even though some studies indicate a genetic component (Kinney and Matthysse, 1978), because there is little agreement on objective and measurable criteria defining the conditions, and few genes known to be associated with the clinical pathology.

198 ROBERT WILLIAMSON AND SUSAN M . DARLING

At this time, recombinant DNA technology is seen as a technique providing gene-specific probes for the analysis of those genetic diseases where the defect is understood. This will certainly be possible, but in our view the use of gene libraries of unassigned gene sequences may be even more informative, as it will allow the assignment of linkage of chromosome-specific sequences with specific phenotypes, both for monogenic and polygenic conditions.

ACKNOWLEDGEMENTS The work of our Department is supported by grants from the Cystic Fibrosis Research Trust,

Muscular Dystrophy Group, U.S. Muscular Dystrophy Association, British Heart Foundation and Medical Research Council. Thanks are due to all of the members of the molecular biology research group, and in particular Julian Crampton, Kay Davies, Peter Little and Steve Humphries, for help in formulating these ideas. This paper is in part an updating of a previous talk given by Humphries and Williamson (1982).

REFERENCES BENZ, E. J. JR, SWERDLOW, P. S., AND FORGET, B. G. 1975. Absence of functional messenger

RNA activity for P-globin chain synthesis in /3”-thalassaemia. Blood, 45, 1-10. BERNARDS, R., AND FLAVELL, R. A. 1980. Physical mapping of the globin gene deletion in

hereditary persistence of fetal haemoglobin (HPFH). Nucleic Acids Res., 8, 1521-1534. BERNARDS, R., LITTLE, P. F. R., ANNISON, G., WILLIAMSON, R., AND FLAVELL, R. A. 1979.

Structure of the human $/I-globin gene locus. Proc. natn. Acad. Sci. U.S.A., 76, 4 8 2 7 4 3 3 1.

BOTSTEIN, D., WHITE, R. L., SKOLNICK, M., AND DAVIS, R. W. 1980. Construction of a genetic linkage map in man using restriction fragment length polymorphisms. Am. J. hum. Genet., 32.314-331.

CHANG, J. C., AND KAN, Y. W. 1979. $-Thalassaemia, a nonsense mutation in man. Proc. natn. Acad. Sci. U.S.A., 76,2886-2889.

COURTNEY, M., AND WILLIAMSON, R. 1979. A nuclear precursor to human y-globin messenger RNA. Nucleic Acids Res., 7, 1121-1 130.

DAHL, H. H., FLAVELL, R. A., AND GROSVELD, F. G. 1981. The use of genomic libraries for the isolation and study of eukaryotic genes. In Genetic engineering 11, edited by R. Williamson, Academic Press, London, pp. 49- 127.

DAVIES, K. E., YOUNG, B. D., ELLES, R. G., HILL, M. E., AND WILLIAMSON, R. 1981. Cloning of a representative genomic library of the human X chromosome after sorting by flow cytometry. Nature (London), 293,314-376.

EFSTRATIADIS, A., POSAKONY, J. W., MANIATIS, T., LAWN, R. M., OICONNELL, C. O., SPRITZ, R. A., DE RIEL, J. K., FORGET, B. G., WEISSMAN, S. M., SLIGHTHORN, J. L., BLECHL, A. E., SMITHIES, O., BARALLE, F. E., SHOULDERS, c. c . , AND PROUDFOOT, N. J. 1980. The structure and evolution of the human P-globin gene family. Cell, 21,653-668.

FORGET, B. G., BENZ, E. J. JR, SKOULTCHI, A., BAGLIONI, c . , AND HOUSMAN, D. 1974. Absence of messenger RNA for P-globin chain in $-thalassaemia. Nature (London), 247,

GUSELLA, J. F., KEYS, C., VARSANGI-BREIREI, A., KAO, F. T., JONES, C., PACK, T. T., AND HOUSMAN, D. 1980. Isolation and localisation of DNA segments from specific human chromosomes. Proc. natn. Acad. Sci. U.S.A., 77,2828-2833.

HIGGS, D. R., PRESSLEY, L., ALDRIDGE, B., CLEGG, J. B., WEATHERALL, D. J., CAO, A., HADJIMINAS, M. G., KATTAMIS, C., METAXATOU, A., RACHMILIEWITZ, E. A., AND SOPHOCLEOUS, T. 1981. Proc. natn. Acnd. Sci. U.S.A., 78,5833-5837.

Recombinant DNA and the study of the inheritance of human hereditary disease. In The future of prenatal diagnosis, edited by H. Galjaard, Churchill Livingstone, Edinburgh, pp. 62-68.

379-38 1.

HUMPHRIES, S. E., AND WILLIAMSON, R. 1982.

GENOME ANAL YSZS 199

JEFFREYS, A. J. 1979. DNA sequence variants in the P- and y-globin genes of man. Cell, 18, 1-10.

KAN, Y. W., AND DOZY, A. M. 1978. Polymorphism of DNA sequence adjacent to human P-globin structural gene: relationship to sickle mutation. Proc. natn. Acad. Sci. U.S.A., 75,

KINNEY, D., AND MATTHYSSE, S. 1978. Genetic transmission of schizophrenia. A . Rev. Med., 29,459473.

KURNIT, D. M., AND HOEHN, H. 1979. Prenatal diagnosis of human genome variation. A . Rev. Gent., 13,235-238.

LAWN, R. M., EFSTRATIADIS, A., O'CONNELL, C., AND MANIATIS, T. 1980. The nucleotide sequence of the human P-globin gene. Cell, 21,647-651.

LITTLE, P. F. R. 1981. DNA analysis and the antenatal diagnosis of haemoglobinopathies. In Genetic engineering I, edited by R. Williamson, Academic Press, London, pp. 6 1-102.

L I ~ L E , P. F. R., ANNISON, G., DARLING, s., WILLIAMSON, R., CAMBA, L., AND MODELL, B. 1980. Model for antenatal diagnosis of Po-thalassaemia and other monogenic disorders by molecular analysis of linked DNA polymorphisms. Nature (London), 285, 1 14-147.

LITTLE, P. F. R., CURTIS, P., COUTELLE, C., VAN DEN BERG, J., DALGLEISH, R., MALCOLM, s., COURTNEY, M., WESTAWAY, D., AND WILLIAMSON, R. 1978. Isolation and partial sequence of recombinant plasmids containing human E-, P- and y-globin cDNA fragments. Nature (London), 273,640-643.

MALCOLM, A. D. B. 1981. The use of restriction enzymes in genetic engineering. Zn Genetic engineering 11, edited by R. Williamson, Academic Press, London, pp. 129-173.

MALCOLM, S. 1982. Lighting up Time. Nature (London), 300,108-109. MANIATIS, T., HARDISON, R. C., LACY, E., LAUER, J., O'CONNELL, C., QUON, D., SIM, G. K.,

and EFSTRATIADIS, A. 1978. The isolation of structural genes from libraries of eukaryotic DNA. Cell, 15,687-701.

MAROTTA, C. A., WILSON, J. T., FORGET, B. G., AND WEISSMAN, S. M. 1977. Human P-globin messenger RNA. 111. Nucleotide sequences from complementary DNA. J. biol. Chem.,

OLD, J. M., WARD, R. H. T., KARNGROZHI, F., PETROV, M., MODELL, B., AND WEATHERALL, D. J. 1982. First trimester fetal diagnosis for haemoglobinopathies: 3 cases. Lancet, ii,

ORKIN, S . H., 1978. The duplicated human a-globin genes lie close together in cellular DNA. Proc. natn. Acad. Sci. U.S.A., 75,5950-5954.

ORKIN, S . H., AND GOFF, S. C. 1981. Nonsense and frameshift mutations in /lo-thalassaemia detected in cloned P-globin genes. J. biol. Chem., 256,9782-9784.

OTTOLENGHI, S., LANYON, W. G., PAUL, J., WILLIAMSON, R., WEATHERALL, D. J., CLEGG, J. B., PRITCHARD, J., POOTRAKUL, s., AND WONG, H. B. 1974. Gene deletion as the cause of a-thalassaemia. Nature (London), 251,389-392.

5631-5635.

252,5040-5053.

14 13-14 16.

SOLOMON, E., AND BODMER, W. F. 1979. Evolution of sickle variant gene. Lancet, i, 923. SOUTHERN, E. M. 1975. Detection of specific sequences among DNA fragments separated by

gel electrophoresis. J. molec. Biol., 98,503-5 17. SPRITZ, R. A., JAGADEESWARRAN, P., CHOUDARY, P. V., BIRO, P. A., ELDER, J. T., DE RIEL,

J. K., MANLEY, J. L., GEFTER, M. L., FORGET, B. G., AND WEISSMAN, S. M. 1981. Base substitution in an intervening sequence of a P+-thalassaemic human globin gene. Proc. natn. Acad. Sci. U.S.A., 78,2455-2459.

TAYLOR, J. M., DOZY, A., KAN, ,Y. W., VARMUS, H. E., LIE-INJO, L. E., GANESAN, J., AND TOOD, D. 1974. Genetic lesion in homozygous czthalassaemia (hydrops fetalis). Nature (London), 251,392-393.

TOLSTOSHEV, P., MITCHELL, J., LANYON, G., WILLIAMSON, R., OTTOLENGHI, S., COMI, P., GIGLIONI, B., MASERA, G., MODELL, B., WEATHERALL, D. J., AND CLEGG, J. B. 1976. Presence of gene for P-globin in homozygous P-thalassaemia. Nature (London), 259,

WESTAWAY, D., AND WILLIAMSON, R. 198 1. An intron nucleotide sequence variant in a cloned 95-98.

P+-thalassaemia globin gene. Nucleic Acids Res., 9, 1777-1738.

200 ROBERT WILLIAMSON AND SUSAN M . DARLING

WILLIAMSON, R., ESKDALE, J., COLEMAN, D. V., NIAZI, M., LOEFFLER, F. E., AND MODELL, B. 1981. Direct gene analysis of chorionic villi: a possible technique for first-trimester antenatal diagnosis of haemoglobinopathies. Lancet, ii, 1125-1 127.

WILSON, J. T., WILSON, L. B., DERIEL, J. K., VILLA-KAMAROFF, K., EFSTRATIADIS, A., FORGET, B. G., AND WEISSMAN, S. M. 1978. Insertion of synthetic copies of human globin genes into bacterial plasmids. Nucleic Acids Res., 5,563-58 1.

WILSON, J. T., MILNER, P. F., SUMNER, M. E., NALLASETH, F. S., FADEL, H. E., REINDOLLAR, R. H., MCDONOUGH, P. G., AND WILSON, L. B. 1982. Use of restriction endonucleases for mapping the allele for Ps-globin. Proc. natn. Acad. Sci. U.S.A., 79,3628-363 1.

WYMAN, A. R., AND WHITE, R. 1980. A highly polymorphic locus in human DNA. Proc. natn. Acad. Sci. U.S.A., 77,6154-6758.