WO1993017104A1 - Dna sequence of myotonic distrophy gene and uses thereof - Google Patents

Abstract

A method by which a nucleotide sequence, specifically a CTG triplet repeat, shown to be expanded in individuals affected with DM can be identified in a sample obtained from an individual. The present method can be used to identify individuals in whom the CTG triplet repeat is present in normal copy number and individuals in whom the CTG triplet repeat occurs in abnormally high copy number, as well as to further identify individuals likely to be minimally affected and individuals likely to be more severely affected. The work described herein also makes available a DM transcription unit or gene which is likely to have an important role in the pathophysiology of DM.

Description

DNA SEQUENCE OF MYOTONIC DYSTROPHY GENE AND USES THEREOF

Description

Background Myotonic dystrophy (DM) is an autosomal dominant neuromuscular disease with an estimated minimum incidence of 1 in 8000 (Harper, P.S., Mvotonic Dystrophy. 2nd ed. , .B. Saunders Co., London, 1989). It is the most common form of muscular dystrophy affecting adults. The clinical picture in DM is well established but exceptionally variable (Harper, P.S., Mvotonic Dystrophy. 2nd ed. , . B. Saunders Co. , London, 1989) . Although generally considered a disease of muscle, with myotonia, progressive weakness and wasting, DM is characterized by abnormalities in a variety of other systems. DM patients often suffer from cardiac conduction defects, smooth muscle involvement, hypersomnia, cataracts, abnormal glucose response, and, in males, premature balding and testicular atrophy (Harper, P.S., Myotonic Dystrophy. 2nd ed. , W.B. Saunders Co., London, 1989). One of the striking features of this disorder is the variability of phenotype, both within and between families. For clinical purposes, patients are often subdivided into three groups according to the clinical syndrome and age at onset of the disorder (Harper, P.S. and Dyken, P.R. , Lancet. 2.553-55 (1972)). The mildest form, which is occasionally difficult to diagnose, is seen in middle or old age and is characterized by cataract with little or of muscle involvement. The classical form, showing myotonia and muscle weakness, most frequently has onset in early adult life and in adolescence. The most severe form, which life and in adolescence. The most severe form, which occurs congenitally, is associated with generalized muscular hypoplasia, mental retardation, and high neonatal mortality. Those congenitally affected offspring 5 surviving the neonatal period invariably exhibit the classical form of the disease in late childhood or adolescence. The congenital form of DM is almost exclusively maternally transmitted. The phenomenon of anticipation (Howeler, C.J. et al.. Brain, 112;779-797

10 (1989)), in which the disease symptoms become more severe and age at onset earlier in successive generations, is often most strikingly manifested in a family producing a congenitally affected child.

Biochemical studies have failed to identify the

15 defective protein in myotonic dystrophy, although several have implicated defects in membrane structure and function (Harper, P.S., Mvotonic Dystrophy. 2nd ed., W.B. Saunders Co., London, 1989). Abnormalities in calcium transport (Seiler, D. and Kuhn, E., Schweitz Med. Wochenschr.

20100:1374-1376 (1970)), membrane fluidity (Butterfield, D.A. et al.. Biochemistry. 12:5078-5082 (1974)), sodium-potassion ATPase stoichiometry (Hull, K. L., Jr. and Roses, A.D., J. Phvsiol.. 254;l69-181 (1976)), and apamin receptor expression (Renaud, J.F. et al.. Nature

25319.676-680 (1986)) have all been reported for DM.

Evidence of reduced phosphorylation of membrane proteins in both red blood cells (Roses, A.D. and Appel, S.H., Proc. Natl. Acad. Sci. USA 70:1855-1859 (1973)) and sarcolemmal membranes from muscle biopsies of patients

30 (Roses, A.D. and Appel, S.H., Nature 250;245-247 (1974)) haε led to the suggestion by these authors that a defect in a protein kinase may be significant in DM.

A better understanding of the underlying mechanism of DM would be very valuable in diagnosing and, ultimately, treating or preventing DM.

Summary of the Invention

Applicants have identified a CTG triplet repeat, present on chromosome 19, which undergoes expansion in myotonic dystrophy (DM) patients. They have also shown that the normal population exhibits great variability in this sequence, which is present in unaffected individuals in 5-27 copies; that DM patients who are minimally af¬ fected have at least 50 CTG repeats; and that more severe¬ ly affected patients have expansion of the repeat- containing segment up to several kilobase (kb) pairs. In addition, Applicants have demonstrated that the CTG repeat is transcribed and is located in the 3' untranslated region of an RNA which encodes a polypeptide which is a member of the protein kinase family and is expressed in tissues affected by DM.

The work described herein makes available a method by which a nucleotide sequence, specifically a CTG triplet repeat, shown to be expanded in individuals affected with DM can be identified in a sample obtained from an indivi- dual. The present method can be used to identify indivi¬ duals in whom the CTG triplet repeat is present in normal copy number and individuals in whom the CTG triplet repeat occurs in abnormally high copy number, as well as to further identify individuals likely to be minimally affected and individuals likely to be more severely affected.

The work described herein also makes available a transcription unit or a DM gene, whose full genomic sequence has been determined by the Applicants, and which is likely to have an important role in the pathophysiology of DM. As shown herein, the RNA which includes in its 3' untranslated region the transcribed CTG triplet repeat, encodes a protein kinase. It is reasonable to expect that amplification of the CTG triplet repeat affects the function of the DM gene, such as by causing a loss of expression of the allele carrying the expanded repeat or by causing a gain of function in the DM gene (e.g., deletion or inactivation of a binding site for a negative control element) .

The protein kinase encoded by the DM gene can be used as the basis for a method of identifying individuals affected by DM, since the mRNA is expressed in tissues affected by DM. The presence or absence, as well as the level of expression, of the protein kinase can be determined in tissues using, for example, polyclonal or monoclonal antibodies which recognize (bind) the protein kinase. Alternatively, the DM gene or mRNA can be detected and/or quantitated using DNA probes described herein and artrecognized hybridization techniques.

The work described herein also makes it possible to- inhibit the protein kinase, particularly in tissues affected by MD and, thus, interfere with its effects on cells and its role in the pathophysiology of DM. Brief Description of the Drawings

Figure 1 is a restriction map of the 10 Kb EcoRI fragment that undergoes expansion in myotonic dystrophy patients. Restriction sites for BamHI(B), EcoRI(E), Hindi(C), HindΙII(H), PstΙ(P) and Saci(S) are indicated. The subcloned 1.4 Kb Ba HI fragment (pMIOM-6) is shown enlarged with the PstI and the Hindi sites, which flank the expanded region. The positions of PCR primers 96, 98, 101, 102 and 103 and the sequence between primers 101 and 102 (SEQ ID No. 1 and SEQ ID No. 2 respectively) are shown.

Figure 2 shows results of Southern blot analysis of the DNA fragment which is expanded in DM. DNA from a DM patient (DM) and an unaffected individual (N) digested with EcoRI, BamHI, PstI and Sacl and hybridized with pMIOM-6 shows an increased band in the DM lane with all enzymes.

Figure 3 is a graphic representation of the distri¬ bution of repeat lengths in the normal population (n=282) . Figure 4 is the pedigree of a myotonic dystrophy family with their genotypes at the DM locus.

Figure 5 is the nucleotide (SEQ ID NO: 3) and the deduced amino acid sequence (SEQ ID NO: 4) of cDNA C28.

Figure 6 is the nucleotide sequence of the DM gene, (SEQ ID NO: 13) including the deduced amino acid sequences of the three predicted reading frames (sequence a, upper line, SEQ ID NO:14; sequence b, middle line, SEQ ID NO:15; sequence c, lower line, SEQ ID NO:16) . The sequences of the intron/exon boundaries have been determined and are indicated by the XXX sites located within the sequence (the actual sequences are not shown) . Figure 7 is a graphic representation of the alternate splice sites of the DM gene and of the resulting cDNAs.

Detailed Description of the Invention

As described herein, the techniques of positional cloning have been used in order to identify the gene responsible for DM. This has resulted in two key find¬ ings, which are also described herein: first, a CTG triplet repeat whose copy number (number of repeats) is greater in affected individuals than in unaffected indi- viduals and correlates with phenotypic severity has been identified. Second, the CTG repeat has been shown to reside in RNA which encodes a polypeptide with strong amino acid homology to members of the protein kinase gene family and is expressed in many of the tissues affected in DM.

More specifically, the modal number of CTG triplet repeats in 282 unaffected individuals tested has been shown to be 5, with the largest number being 27. In contrast, individuals minimally affected by DM have been shown to have at least 50 CTG triplet repeats and more severely affected individuals have been shown to have even greater numbers of copies (e.g., expansion of the repeat- containing sequence by as much as several Kb pairs) . The protein kinase-like polypeptide has been shown to be highly expressed in heart, expressed to a lesser extent in muscle and also expressed in brain.

Based on the work described herein, a method of determining in DNA obtained from an individual, the copy number of the CTG triplet repeat which has been shown to be increased or expanded in DNA individuals affected with muscular dystrophy is available. Thus, a method of determining whether an individual is likely to be affected with myotonic dystrophy is also available. In the method, DNA is obtained from an individual to be assessed and the copy number of CTG triplet repeats on chromosome 19 DNA is determined. If the copy number of CTG triplet repeats is at least 50, the individual is likely to be affected with myotonic dystrophy. The presence of 50 CTG triplet repeats is an indication that the individual is or is likely to be minimally affected; the presence of a greater number of repeats is an indication that the individual is likely to be severely affected. In this method, DNA can be obtained from a variety of tissues (e.g., blood, muscle, skin) , either prenatally or postnatally. The present work also provides a method of deter¬ mining whether an individual is likely to be affected with myotonic dystrophy in which the expression of the protein kinase encoded by the DNA of Figure 5 or Figure 6 (or a protein kinase having a substantially equivalent sequence) is detected, particularly, in tissues affected by myotonic dystrophy (e.g., heart, brain and muscle). In this method a tissue sample to be analyzed is obtained from an individual to be assessed for the likelihood he or she will be or is affected with myotonic dystrophy. For example, expression of the protein kinase can be determined through the use of an antibody specific for (one which binds) the protein kinase described herein. The antibody used can be polyclonal or monoclonal and is contacted with the tissue to be assessed, after the tissue has been processed or treated to render the protein kinase (if present) available for binding by the antibody. Binding of the antibody to a component of the tissue sample is indicative of the presence of the protein kinase and, thus, of the likelihood the individual will be or is affected with myotonic dystrophy. The work described herein also makes available antibodies specific for (which bind to) the protein kinase encoded by the DNA sequence of Figure 5 or Figure 6 or an equivalent protein kinase (a protein kinase encoded by a substantially similar DNA sequence and/or having substantially the same amino acid sequence as that represented in Figure 5 or Figure 6) .

The work described herein also makes it possible to develop methods of treating or preventing myotonic dystro¬ phy. For example, it is now known that a protein kinase is expressed in tissues affected by myotonic dystrophy. The effects of the protein kinase can be reduced (totally or partially) by administering to an individual affected with or likely to be affected with myotonic dystrophy a drug which interferes with the kinase activity, either directly or indirectly. For example, a drug which inter¬ feres with expression of the protein kinase (e.g., a nucleotide sequence which binds to the kinase-encoding sequence and prevents it from being transcribed/expressed) can be used. As a result, less protein kinase is produced than would otherwise be the case and its effects are reduced. Alternatively, a drug which destroys or other¬ wise inactivates or interferes with the activity of the protein kinase can be administered. It is also reasonable to expect that the expansion of the CTG repeat in the 3' UTR of the protein kinase gene plays a role in the patho¬ physiology of myotonic dystrophy, perhaps through an effect on the gene. The expansion of the CTG repeat may lead to a gain or loss of function in the gene. In either case, it is possible to interfere with the effect of the expanded CTG repeats on the DM gene, such as by cleaving the expanded region from the gene or otherwise inactivat¬ ing it.

The following is a description of the mapping of the DM region of chromosome 19, which is increased in size in DM; identification of the CTG triplet repeats on chromosome 19 and assessment of the number of repeats in normal and affected individuals; identification and characterization of genomic clones spanning the 10 kb fragment which is increased in size in DM; the cDNAs which were isolated from various libraries by probing with one of the genomic clones; and the full length DNA sequence of the DM gene.

Over the past few years, both genetic maps (Johnson, K. et al.. Am. J. Hum. Genet. 46:1073-1082 (1990); Harley, H.G. et al.. Nature. 155:545-546 (1991); Tallfidis, C. et al. Am. J. Hum. Genet. 49_:961-965 (1991)) and physical maps (Korneluk, R.G. et al.. Genomics 5:596-604 (1989); Smeets, H. et al.. Am J. Hum. Genet. £6.:492-502 (1990) ; Brook, J.D. et al.. Hum. Genet. 87:65-72 (1991); Brook, J.D. et al.. J. Med. Genet. 2_6:84-88 (1991)) of the long arm of human chromosome 19 have been produced to localize the DM gene to band 19ql3.3 between DNA markers ERCC1 and D19S51. Construction and analysis of radiation-reduced hybrids, as described herein, and YAC, cosmid, and phage libraries (Buxton, J. et al.. Nature 355:547-548 (1992); Jansen, G. et al.. Nature 332:276-281 (1992)) has allowed saturation mapping of the interval between these two markers. The DM region could be narrowed down further through linkage disequilibrium studies. Markers D19S63 and D19S95 are in strong linkage disequilibrium with DM (Harley, H.G. et al.. Hum. Genet. 87:73-80 (1991), Harley H.G. et al.. Nature 355:545-546 (1992)). No disequilib¬ rium was observed for markers flanking these loci. Screening of phage libraries derived from the radiation reduced hybrid, 2F5 (produced as described in the Exemplification) produced a series of overlapping phage clones that spanned the interval between D19S63 and D19S95.

This intensive search resulted in the identification of DNA markers adjacent to D19S95 that detect patient specific bands on Southern blots (Harley, H.G. et al.. Nature 355:545-546 (1992); Buxton, J. et al.. Nature

355:547-548 (1992); Aslandis, C. et al.. Nature 355:548551 (1992)). Probes pBB0.7 (Harley, H.G. et al.. Nature 355:545-546 (1992) and cDNA 25 (Buxton, J. et al.. Nature 355:547-548 (1992) identify the same EcoRI restriction fragment length polymorphism in the normal population, with alleles of 9 kb or 10 kb. In 43 of 53 unrelated affected individuals reported in these two studies, only one of the normal-sized alleles is present, plus an additional, larger, disease-specific band. This restric- tion fragment varies in length between patients, even between siblings within the same family. Furthermore, the size of the variable fragment increases in successive generations and shows a correlation between increased severity and earlier onset of the disease. The largest fragment detected is 15 kb, an increase of 5 kb over the normal size (Harley, H.G. et al.. Nature 355:545-546 (1992) .

In many individuals with fragments larger than 11 kb, a diffuse hybridization signal corresponding to a DNA fragment size greater than 11 kb is observed in a gel that otherwise gives tight DNA banding patterns, indicating that somatic mosaicism with respect to the precise extent of the increase in size of the DNA sequence within the EcoRI fragment has occurred. This situation is strikingly similar to that reported recently for the fragile-X mental retardation syndrome, where variation in length of a CGG repeat results in genetic instability (Dietrich, A. et al.. Nucl. Acids Res. 19:2567-2572 (1991); Fu, Y.-M., et al.. Cell 67:1047-1058 (1991); Kremer, E.J. et al.. Science 252:1711-1714 (1991); Oberle, I. et al.. Science 252:1097-1102 (1991); Verkerk, A.J. et al.. Cell 65:905- 914 (1991); Yu, S. et al.. Science 252:1179-1181 (1991)).

In order to identify the mutation in DM, genomic clones spanning the 10 kb fragment that is increased in size in this disease were characterized. We have identified a CTG repeat sequence that is highly polymorphic in the normal population and that undergoes remarkable expansion in DM patients. The modal number of repeats found in 282 normal alleles surveyed is 5 (48%) , with the largest being 27. Minimally affected DM patients have at least 50 copies. The CTG repeat is transcribed and is found at a position 500 bp from the poly(a) tract of an mRNA expressed in many of the tissues affected in DM. The RNA in which the repeat resides encodes a polypeptide with strong amino acid homology to members of the protein kinase gene family. The work which resulted in the findings described above is described in detail in the following sections. The experimental procedures used are described in a subsequent section.

Fine Mapping of the Region Amplified in DM Patients A series of phage clones derived from libraries of radiation-reduced hybrid 2F5 (see the Exemplification) were used to span the interval between D19S63 and D19S95, the loci in linkage disequilibrium with DM. Intensive screening of this interval led to the identification of clones ΛM10M (detected by hybridization with cDNA 25

(Buxton et al.. Nature 155:547-548 (1992)) and kM8L and kSM2 (which contain clone pBBO.7 (Harley et al. , Hum. Genet. 87:73-80 (9992)). These clones span the lOkb EcoRI fragment that is increased in size in DM patients. Figure 1 shows a detailed restriction map of this interval. Sites for BamHI, Hindi, Hindlll, PstI, and Sad are marked. Single-copy probes mapping within this EcoRI fragment were hybridized to DNA from patient and normal individuals digested with a series of restriction enzymes. Figure 2 shows the result of one such hybridization. The normal control is heterozygous for an EcoRI RFLP. As previously reported, a patient-specific EcoRI band, larger in size than either normal allele, is observed in the DNA of the DM patient (Harley et al.. 1992; Buxton et al.. 1992). Digestion with BamHI, PstI, and Sad each revealed patient-specific bands. The smallest region containing the expanded sequence that could be established by hybri¬ dization was a 475 bp Pstl-HincII fragment contained within the 1.4 kb BamHI fragment (pM10M-6) shown in Figure 1. Part of this fragment was sequenced to make primers for polymerase chain reaction (PCR) analysis. Amplifi¬ cation by PCR between oligos 98 and 100 produced single bands that were identical in size in patient, normal and pM10M-6 lanes when visualized on ethidium-stained agarose gels. In contrast, PCR using oligos 96 and 103 produced two bands in a normal human sample, a single band in DM DNA and a single band in pM10M-6 that was smaller than any of the other bands. Analysis of the sequence derived from pM10M-6 between oligos 96 and 103 revealed tandem repeats of the trinucleotide CTG. Two other oligos, 101 and 102, which more closely flank this triplet repeat, were tested using PCR. These produced similar but more striking band size differences than with oligos 96 and 103 because of the smaller PCR product. The sequence of pM10M-6 between oligos 101 and 102 is shown in Figure 1.

Variability of the CTG Repeat in the Normal Population

In order to examine length variability of the PCR fragment produced with oligos 101 and 102 in the normal population, radio-labeled products obtained from a series of normal individuals were analyzed on sequencing gels. The majority of individuals are heterozygous at this locus. Interestingly, shadow bands occurred in positions indicating that they differed by three bases from the major bands. To confirm that this variability of length is due to different numbers of the triplet repeat, the PCR products from two normal heterozygous individuals were cloned and sequenced. From one individual, six clones were analyzed; three contained 12 copies of the CTG repeat, two contained 17 repeats, and one had 18 repeats (the variation between clones with 17 and 18 repeats may be an artifact of stuttering during PCR) . In the second individual, two clones were sequenced and contained 5 and 11 copies of the repeat. These repeat lengths are consistent with the size of bands from these individuals determined by analysis of labeled PCR products on sequencing gels. In all clones analyzed, the sequence flanking the repeat was identical to that derived from pM10-M6, with the exception of two point mutations, which differed between clones and which may be due to errors by Taq polymerase.

The variation in repeat length was analyzed in 282 individuals, and the distribution of repeats in the normal population is shown in Figure 3. Alleles were sized against sequenced standards and sequencing ladders. Over 40% of alleles analyzed had 5 repeats. No alleles were observed with 6 or 8 triplets, and only single examples of 7 and 9 were found. The majority of repeats were in the 10-16 interval, with the highest being 27. Normal Mendelian inheritance was demonstrated for 20 meioses in two CEPH pedigrees.

Variability of the CTG Repeat in DM Patients

Initial attempts to characterize the variable region in patients led to the observation that the DM sample tested by PCR assay produced a single band. To establish whether the single band was specific to this sample or whether it reflected something common to all patients, we analyzed by PCR 12 unrelated DM samples. PCR products were visualized on ethidium-stained gels. In all cases a single band from the normal size range (<200 bp) was present. In one sample (DMH9) a faint diffused band was observed above 360 bp on ethidium-stained gel. This patient was not severely affected and did not show signs of the disorder until age 45. In order to determine the nature of the larger fragment in patient DMH9, the PCR product from this individual was cloned and sequenced. In excess of 80 CTG repeats were present. It was not possible to read beyond the CTG triplet; however, sequencing from either side of the repeat revealed that the flanking DNA was intact and had the same sequence as pMlOM-6.

A single band was observed in PCR analysis of DNA from 11 of 12 patients, who had increases in Southern blot bands of 2 kb or more. The 12th patient, an individual who had later onset of the disorder, produced a faint upper band on PCR analysis. This suggested that in the other 11 patients it had been impossible to amplify their second allele because of the extent of the amplified CTG repeat. In order to identify additional individuals likely to contain expanded CTG repeats that could be successfully identified by PCR, we focused on other mildly affected individuals. Three pairs of grandparents were selected from DM pedigrees in which it was difficult to decide, on clinical grounds, which was the affected individual. In each case, one of the pair had cataracts. PCR products analyzed on ethidium-stained gels showed that one individual from each pair had two alleles in the normal range, whereas the other grandparent (in each case, the one with cataracts) had one band in the normal range in addition to a second diffuse band at about 250 bp. For accurate sizing of this larger band labeled PCR products were analyzed on sequencing gels. Although unrelated, these two individuals had very similar larger bands (corresponding to 50-55 repeats) . The larger allele was amplified using PCR and cloned and sequenced, in order to determine whether the increased fragment size in these patients is due entirely to the expansion of the CTG repeat. Six clones were analyzed from one patient (DMH6) : four contained 52 repeats, one had 54 repeats, and one had 57 repeats. Sequence from one such clone is shown in Figure 6. Analysis of five clones from patient DMH1 revealed two with 50 repeats, two with 52 repeats, and one with 61 repeats. In all clones the sequence of the DNA flanking the repeat was the same as that derived from pM10M-6. The only difference was in the length of the triplet repeat. The different CTG lengths observed in different clones from the same individual may reflect somatic mosaicism or stuttering during the PCR reaction.

A DM family that shows increased severity of disease in successive generations is shown in Figure 4. A band of approximately 300 bp is observed in the PCR product of four family members in the first two generations shown in the pedigree. This result demonstrates that a CTG repeat unit of approximately 60 repeats can be transmitted from one generation to the next without an obligatory expan¬ sion. Individuals in the first two generations had mild symptoms such as cataracts or were apparently asympto¬ matic. However, individuals in the third and fourth generations of the family showed severe symptoms of DM. These individuals showed only a single PCR band within the normal range. Southern blotting analysis confirmed that, for these individuals, as for most other severely affected DM patients, a dramatic expansion in allele size has occurred to increase the repeat size beyond that which can be detected by PCR analysis.

The CTG Repeat Is Transcribed

The triplet repeat sequences amplified in fragile-X syndrome and X-linked spinal and bulbar muscular atrophy (SBMA) are expressed in mRNAs. To test whether the CTG repeat amplified in DM might be within a gene, clone pM10M-6 was hybridized to a Southern blot of DNA from different species. Results showed that DNA sequences contained within this clone are strongly conserved, suggestive of a transcribed sequence. pM10M-6 was used to screen cDNA libraries derived from several different sources, including adult frontal cortex, substantia nigra, fetal muscle and fetal brain. A total of 10⁶ clones were screened from these libraries and 110 positives were identified. Twenty were purified and six, from different libraries, were selected for further analysis. These clones were designated C28, C31, C34, C35, C39 and C85. The insert size of each clone and the library from which it was derived are: C28 (frontal cortex), 2.5 kb; C31 (frontal cortex), 2.1 kb; C34 (substantia nigra), 1.7 kb; C35 (fetal brain), 1.7 kb; C39 (frontal cortex), 2.7 kb; C85 (fetal muscle), 2.8 kb. All six clones were hybri¬ dized to a panel of hybrid cell lines to confirm that they were from the expected region of chromosome 19. They were also mapped to filters of digested genomic phage clones ΛM10M and M8L (which span the 10 kb EcoRI fragment amplified in patients) to determine how much genomic DNA they cover. cDNAs C28 and C85 each span at least 10 kb of genomic DNA. Clone C39 was chimeric at the 5' end. whereas the others mapped as expected. Clones C28, C34 and C35 were completely sequenced and clone C85 was partially sequenced. All clones contained the CTG repeat, and this varied in length between clones. Clones C28, C34, C35, and C85 contained 11, 5, 12, and 13 triplets, respectively. Comparison of the cDNAs with the genomic clones indicates that the gene is transcribed in the orientation telomere-to-centromere and that the CTG triplet is on the coding strand. Results of Northern blot analysis with C28 showed that the full-length transcript is between 3.0 and 3.3 kb in length and is highly expressed in heart and to a lesser extent in muscle. Prolonged autoradiographic exposure indicates that this transcript is also expressed in brain, consistent with the identification of cDNAs from this tissue. The sequence of C28 is shown in Figure 5. This clone contains a complete 3' terminus, which includes the polyadenylation addition signal AAUAAA. The largest predicted open reading frame extends from the beginning of the sequence to position 1747 with a coding capacity of 582 amino acids. The C28 sequence was compared to the nonredundant sequence database, which combines all avail¬ able protein databases. This sequence search revealed homology to the cyclic AMP-dependent protein kinases. The highest score was to the protein kinase TKR-YKR from Saccharomyces cerevisiae. The 11 protein kinase domains are found within the first 300 residues of deduced amino acid sequence. Beyond the kinase domain, some slight homology to the chicken myosin heavy chain was observed. The disease, myotonic dystrophy shows a clear autosomal dominant pattern of inheritance. However, there are several aspects of this disorder that are particularly challenging to explain from a molecular genetic perspective. First, there is the considerable variability of phenotype between affected individuals, even within the same family. Second, there is an association of DM with specific haplotypes in the population (Harley, H.G. et al.. Hum. Genet. 82:73-80 (1991) indicating that most cases have resulted from a small number of genetic events. Third, there is the multi-systemic nature of the phenotype. Fourth, there is an apparent increase in severity of symptoms and reduction in age at onset that is observed during transmission of the gene within families. Several molecular genetic features of DM appear to be directly comparable to fragile-X syndrome. In fragile-X syndrome, increasing allele size at the FMR-1 locus, measured by Southern hybridization, is due to expansion of a CGG repeat at the 5' end of the FMR-1 gene. Increased allele size correlated with the severity of disease (Fu, Y.-M. et al. Cell 67:1047-1058 (1991). The extent of fragment size increase in DM also shows a clear correlation with increased severity and age at onset of the disease (Harley, H.G. et al.. Nature 355:545-546 (1992); Buxton, J. et a , Nature 355:547-548 (1992). The identification of a CTG repeat, which is highly variable in the normal population and which is greatly expanded in DM patients, extends the parallels between DM and fragile-X. Analysis of cloned PCR products reveals that the increase in size of the PCR products observed in mildly and minimally affected DM patients is due entirely to increased number of the CTG repeats. The DNA flanking the repeat is intact in all clones examined. Two minimally affected patients, DMH1 and DMH6, had repeat lengths of 50 and 62 CTGs, respectively, whereas a slightly more severe¬ ly affected patient, DMH9, had in excess of 80 repeats. PCR analysis of the most severely affected patients reveals only a single band, which is in the normal size range. In these individuals the expanded allele cannot easily be visualized by PCR assay. Thus, as with fragile- X, phenotypic severity correlates with the number of repeats. —

Other similarities and differences between DM and fragile-X are noteworthy. In fragile-X, individuals are categorized as normal premutation or full mutation on the basis of CGG repeat number (Fu, Y.-M et al.. Cell 57:1047- 1058 (1991)) . Transmissions of 46 repeats and below are within the normal range and are stable. The transition from stability to instability occurs within the 46-52 repeat range. Premutations showing no phenotypic effect range from 52 to 200 repeats and are meiotically unstable. For myotonic dystrophy there is, as yet, no clear classi¬ fication of premutation. However, it seems likely that minimally affected individuals could go undiagnosed, were it not for the appearance of a more severely affected individual in a subsequent generation. This point is further illustrated in Figure 4, in which two siblings appeared normal on clinical examination at ages 64 and 61, yet showed the same upper allele as their mother, who lived to age 89 and who did not show myotonia but did have cataractε. These minimally affected individuals may be comparable to the premutational state seen in fragile-X.

The molecular mechanisms that determine the stability during inheritance of DM alleles is unclear at present. There is considerable variability at this locus in the normal population: over 75% of normal individuals are heterozygous. The largest allele observed had 27 repeats. The most common allele in the population, 5 repeats, is found at a frequency of 0.48. Analysis of repeat length in 141 individuals revealed no alleles with fewer than 5 copies of the repeat. Furthermore, only two alleles were found in the 6-9 repeat range. Replication of the 5 repeat allele appears to be stable. Duplication or triplication with subsequent slippage during DNA repli- cation (Jeffreys, et al.. 1988)) may account for the generation of other alleles. Unequal crossing over would seem to be unlikely as the mechanism generating allelic diversity, in view of the shortage of alleles between 5 and 10 repeats. Clearly, further study of the mechanisms that underlie variation in repeat length at the DM locus will be of great interest. In addition, it will be important to determine the extent of tissue mosaicism in the expansion of the CTG repeat. Such mosaicism could have a significant impact.on clinical phenotype. One of the families documented by Fu, et al.. appears to represent a fragile-X premutation segregating in the normal population, which has yet to undergo expansion to the full mutation (Fu, Y.-M. et al.^' _r Cell 52:1047-1058 (1991)). It is unclear in fragile-X whether a new mutation is a frequent event. It seems quite possible, however, that transition from a large allele in the normal population to premutation allele in fragile-X could occur by the same mechanism that generates diversity amongst alleles in the normal size range (Fu, Y.-M et al.. Cell 57:1047-1058 (1991)). Thus, large alleles in the normal population and fragile-X premutations would represent a continuum across a stability threshold, with the fragile-X phenotype generated by multiple independent events. At the DM locus, however, two observations suggest that expansion of the CTG repeat to a clinically significant level is likely to occur only in a specific population subgroup.

First, there is clear evidence in heterogeneous populations of linkage disequilibrium between DM and polymorphisms at "nearby" loci (Harley, H.G. et al.. Hum. Genet. 87:73-80 (1991); Harley, H.G. et al.. Nature

355:545-546 (1992Ϊ) . This implies either that there are few mutations, possibly a single ancestral event, or that specific nearby polymorphisms predispose to the generation of DM mutations. It is difficult to envisage a mechanism by which multiple polymorphisms at distances of up to 70 kb (in the case of D19S63) from the CTG repeat "could predispose to DM. On the other hand, if there are very few, if any, new mutations, some mechanism must maintain the disease allele in the population, particularly in view of the genetic endpoint represented by severely affected individuals. Thus, there may be a large unrecognized pool of individuals in the population who carry and transmit the DM premutation with little, if any, phenotypic effect. Extensive studies of normal population (possible focusing on individuals with cataracts) will be necessary to test this possibility. Second, the distribution of CTG repeat alleles is quite distinct in the normal population and in DM pa¬ tients. In DM the smallest number of CTG repeats observed is 50, almost double the largest number of repeats seen in the normal population. It is possible that a doubling or tripling in repeat number is the ancestral event that predisposes an allele at the DM locus to further expansion into an allele associated with the complete disease phenotype. While it is certainly possible that amplification of the CTG affects the expression of several transcription units in the immediate vicinity of the repeat, it seems very likely that the transcription unit we have identified in this study plays an important role in the pathophysiology of DM. The mechanism through which expansion of the CTG repeat affects the function of the DM gene remains to be elucidated. Since DM is a dominantly inherited disorder, routant alleles must exert an effect in the presence of a normal allele. There are a number of possible ways in which an amplified sequence in the 3' UTR of a gene could exert an effect on the function of that gene. One possibility is that, analogous to fragile-X, the expansion of the CTG repeat causes a loss of expression of the allele carrying the expanded repeat. If this is the case, then the DM gene must indeed be extremely sensitive to gene dosage, since gene expression levels in the presence of a normal allele can range only between 50% and 100% of normal. Genomic imprinting cannot be invoked to increase this range too much further, since DM can be inherited from either the father or the mother, with quite severe symptoms. Alternatively, the expansion of the CTG repeat may lead to a gain of function in the DM gene. Gain-offunction mutations in the 3' UTR of the fem-3 and lin-14 genes of Caenorhabditis eleσans have recently been demonstrated (Ahringer, J. and J. Kimble, 5 Nature 349:346-348 (1991); .Wightman, B. et al.. Gene Dev. 1:1813-1824 (1991)). In both cases, deletion or inactivation of a binding site for a negative control element is thought to result in unregulated activity of these genes. Amplification of the CTG repeat in DM may be

10 producing a similar effect.

The similarity of the DM gene to members of the protein kinase family, in particular cAMP-dependent protein kinase (cAPK) , opens a broad range of physio¬ logical questions that should be directly tested. cAPKs

15 (Hunter, T., Meth. Enzymol. 200:3-37 (1991)) are known to modulate the activity of excitable cells by phosphoryl¬ ation of ion channels, exert control of glycogen and lipid metabolism through cascades of enzyme phosphorylation, and modify gene expression (Yamamoto, K.K. et al.. Nature

20 334:494-498 (1988); Foulkes, N.S. et al.. Cell 54:739-749 (1991)) . Abnormalities in function or regulation of such a molecule fit well with the diverse phenotypic effects exhibited by DM patients. It is intriguing that bio¬ chemical data produced nearly 20 years ago by Roses and

25 Appel, indicated that a defect in a protein kinase may be significant in DM. (Roses, A.D. and S.H. Appel, Proc. Natl. Acad. Sci. USA 70:1855-1860 (1973); Roses, A.D. and S.H. Appel, Nature 250:245-247 (1974)). A resolution of the underlying physiological bases of DM should prove to

30 be an important step both in developing appropriate therapeutic intervention and in gaining further insight into the multiple organ systems that exhibit physiological disruption in DM.

EXEMPLIFICATION

Experimental Procedures: PCR Analysis

The PCR analysis was performed as follows. Reactions (10 μl) were set up using standard PCR conditions (50 mM KC1, 1.5 mM MgCl₂, 10 mM Tris [pH 8.3], 200 //M dNTPs, 1 μVL of each primer, and 20 ng template DNA) . For the radio-labeled experiments, oligo 101 was incubated for 30 min at 37^βC with T4 polynucleotide kinase (3 μ) in 20 μl reaction with 50 mM Tris-HCl (pH 7.5). 10 mM MgCl₂, 5 mM dithiothreitol, 0.1 mM spermidine, and 1 μl of ( 32P) ATP (3000 ci/mmol) . The PCR reaction was spiked with the labeled primer in a ratio of 30:1 unlabeled to labeled. PCR reactions were carried out on a Stratagene 96 well thermocycler. Cycling conditions were as follows: 1 x (3 min, 94^*C) , 35 x (10s, 94'C; 30s, 62^°C; 30s, 72^*C) , 1 x (5 min, 72°C) , 15°C soak. Samples then were either loaded on agarose gels or were diluted 1:1 with sequencing buffer, denatured and loaded on 8% sequencing gels.

The sequences of oligos used in PCR reactions was as follows: 96, GGT GCG TGG AGG ATG GAA CAC GGA C (SEQ ID NO: 5) 98, GCG TGC GAG TGG ACT AAC AAC AGC TG (SEQ ID NO: 6)

100, CAC GCT CGG AGC GGT TGT GAA CTG G (SEQ ID NO: 7)

101, CTT CCC AGG CCT GCA GTT TGC CCA TC (SEQ ID NO: 8)

102, GAA CGG GGC TCG AAG GGT CCT TGT AGC (SEQ ID NO: 9) 103, CCA^'GTT CAC AAC CGC TCC GAG CGT G (SEQ ID NO: 10)

Genomic Digests and Southern Blots

Genomic DNAs were digested with restriction endonuc- leases PstI Sad, EcoRI or BamHI (New England Biolabs)* in 30 μl or 40 μl reactions with NE Buffer 10 x according to the manufacturer's instructions. Digested DNAs were run on 0.8% agarose gels (FMC), denatured in 0.5 M NaOH, 1.5 M NaCl, neutralized in 1 M Tris (ph 7.0), 1.5 M NaCl and transferred to zetabind (AMF) membranes in 10 x SSC.

Hybridizations

Hybridizations to both Northern and Southern blots were performed at 42^βC in 50% formamide with 5 x SSC, 1 x Denhardt's solution, 0.02 M NaP0₄, 100 μg/ml single- stranded DNA, 10% dextran sulphate. DNA probes were labeled by random priming (Feinberg, A.P. and B.

Vogelstein, Biochem. Biophvs. Res. Commun. 111:47-54 (1983)).

Northern Blots

Total RNA was extracted from baboon tissue using the method of Auffray and Rougeon (1980) with modifications from Buckler et al. (1991) . Poly(A) RNA was isolated from oligo(OT)-cellulose, and gels and Northern blots were set up as described in Buckler et al. fl991) .

DNA Sequencing Three parallel sequencing strategies were adopted. Much of the sequencing was carried out using a U.S. Biochemical sequence kit according to manufacturer's instructions on cDNA constructs. In general, vector oligonucleotides were used as primers, and in some cases, specific oligonucleotides were synthesized from deduced sequence.

cDNA Libraries

Four cDNA libraries were screened. Three, con¬ structed in ΛZAP from frontal cortex, substantia nigre and fetal brain were kindly supplied by Dr. Marcy MacDonald. A total muscle library in ΛGTIO was kindly supplied by Dr. L. Kunkel (Koenig, et aL , Cell 50:509-517 (1987).

DNA Database Searching

DNA databases were searched on a digital VAX computer using the GCC software package (Altschul, S.F. et al.. J. Mol. Biol. 215:403-410 (1990)) and the BLAST network service from the National Center for Biotechnology Information.

Radiation Reduced Hybrids for the Myotonic Dystrophy Locus:

The following methods were used for the construction and analysis of radiation-reduced hybrids for use in the identification of the myotonic dystrophy gene. One hybrid, 2F5, contains 2-3 megabases of human material, derived exclusively from human chromosome 19 and includes markers which flank DM. DNA from this hybrid was used to construct genomic phage libraries from which 230 phage containing human inserts have been identified. Two other hybrids produced provide breakpoints within the interval covered by 2F5 and are useful in subdividing the phage clones into three groups.

Cell Culture

Cell line 20XP3542-1-4 was used as the parental cell line in two different X-irradiation experiments.

1. X-irradiation of 5000 rads followed by fusion to DNA repair deficient hamster cell line UV20 and selection with mitomycin C and polio virus.

2. X-irradiation of 5000 rads followed by fusion to HPRT negative hamster cell line Wg3h and selection in medium containing HAT (Hypoxanthine, Aminopterin and Thymidine) for clones which retained the HPRT gene and other material from the parental cell line.

Routinely, cells were grown in minimal essential medium with 10% fetal calf serum, penicillin and strepto¬ mycin (Northumbria Biologicals Ltd. , Great Britain) . In each experimental group, three lots of 10⁷ cells were irradiated using a Cs¹³⁷ source (0.66Mev; 0.9 Gy/min) . Irradiated cells were fused with 10⁷UV20 cells in experi- mental group '1' and to 10⁷Wg3h cells in experiment '2'. Each fusion was split to ten 75cm dishes and exposed to selective media twenty four hours post irradiation. In group '1', cells were selected in a final concentration of 0.01 -M mitomycin C (Sigma Chemicals). For group '2', selection was performed with ixHAT (from a 10Ox concen¬ trate; Flow Laboratories, Great Britain) . Twelve to twenty days post-irradiation, two to three surviving clones were picked from each dish using metal cloning rings and transferred to 24 well tissue culture plates (Costar) . In experiment '1', cells surviving exposure to polio virus were grown up to 2x75cm² flasks. One flask was used for DNA extraction and the other was frozen down. In experiment '2', duplicate clones were grown in parallel and one flask exposed to polio virus, using the procedure described previously (Brook, J.D. et al.. Genomics 1:320- 328 (1987)). For those clones killed by polio virus the duplicates were grown up and DNA extracted for analysis and cells frozen.

DNA Techniques

Southern blotting, filter hybridizations and probe labelling were performed according to standard procedures (Sambrook, et al.. 1989) . Probes containing human repeat sequences were pre-annealed for five hours with a 1000- fold excess of sheared, unlabelled human DNA.

Cell line DNAs were analyzed with twenty-three DNA markers which were either positive in or derived from parental cell line 20XP3542-1-4. These were divided into two series.

Series 1 DNA markers included BCL3, AP0§c2, CKM, ERCC1, NE16, pD26 (D17S243) and pD48 (D8S42) .

Series 2 DNA markers include pD3 (D19S61) , pD8 (D19S62), pDIO (D19S63) , pD36 (D19S64) , pNE17, pD50

(D17S247), pD13 (D17S245) , pD38 (D8S81) , pD48 (D8S82) , pD47 (D8S83), pD51, pD78, pD55, pD32, pD67 and pD41. The phage for which D-S numbers have been assigned are des¬ cribed in Brook, et al. and Harley, et al. (Brook, J.D. et al.. J. Med. Genet. £8:84-88 (1991); Harley, H.G. et al.. Hum. Genet. 87:73-80 (1991)). The probe defining locus D19S51 (pl34C) was described by Johnson, et al. (Johnson, K. et al.. Am. J. Hum. Genet. 46:1073-1081 (1990)).

Library Construction and Screening A genomic DNA library was constructed from cell line 2F5 in vector Lambda DASH (Stratagene) by partial Mbol digestion of cell line DNA, size selection of 15-25 Kb fragments on low melting point agarose gels and cloning into the BamHI site of the vector. Recombinants were plated on bacterial strain NM542 and screened with total human DNA. Southern blots were prepared from DNA of 35 phage digested with restriction enzymes BamHI, EcoRI, Hindlll and Sail and hybridized with human DNA. Those bands not hybridizing well with human DNA were identified and excised from LMP agarose gel containing digests of the same phage and hybridized against mapping filters. These consisted of six lanes: Human, Hamster, 5B3, 3A3, GM89A99C7 and PK-87-19. Cell lines 5B3 and 3A3 are described in the results section. PK-87-19 contains a single chromosome 19 as its only human chromosome and GM89A99c7 contains the region 19ql3"3-19qter plus chromo¬ somes 3, 4, 7, 11, 18, 21, 22 and Xpter-Xq24.

Plus Field Gel Electrophoresis

For analytical and preparative pulsed-field gel . electrophoresis (PFGE) , a Biorad CHEF-DRII apparatus was used. Pulse times were ramped from 40 to 200 or from 50 to 300 seconds, and the gels were run at 160 volts for 42 or 46 hours, with a buffer temperature of 15°C. The gels were 1% agarose or 1% low melting-point agarose (Gibco- BRL) for preparative gels. Samples of DNA from human white blood cells (female) and-2F5 and 20XP3542-1-4 hybrid cell-lines were prepared in agarose blocks as described previously (Shaw, D.J. et al.. Hum. Genet. 83:71-74 (1989)) and digested with rare-cutter restriction enzymes. Phage lambda libraries were constructed from DNA fractionated by preparative PFGE. Approximately 100 μg (16 blocks) of 2F5 hybrid DNA was digested .with Notl and separated by PFGE. After electrophoresis, the outside lanes containing size markers were cut off and stained with ethidium bromide. The gel was re-assembled and the central section containing the fractionated hybrid cell line DNA, was cut into 2 mm slices at right angles to the direction of electrophoresis. These were melted at 65°C, cooled to 37°C and the agarose was removed by digestion with agarose followed by phenol and chloroform extraction. The DNA was recovered by ethanol precipitation, a small aliquot of each fraction was digested with PstI, the samples were separated by standard gel electrophoresis, blotted and hybridized with various probes to determine in which fractions the corresponding Notl fragments were present.

DNA from the chosen fractions was then partially digested with Mbol to 15-25 kb average size. Due to the small amount of DNA available, the partial digest con¬ ditions were established by electrophoresis of the trial samples in 0.6% agarose gels, followed by blotting and hybridization with labelled Chinese hamster DNA. The partial digests were cloned in two ways: firstly, using lambda EMBL3 cut with BamHI in order to obtain Mbol fragments internal to the original Notl fragment; and secondly, with a derivative of lambda EMBL3 in which one of the BamHI cloning sites was replaced with a Notl site. This allowed the ends of the Notl fragment to be obtained. The ligated DNAs were packaged in vitro and plated on IJ . coli strain ER1458. Phage with human inserts were identified by hybridization with labelled total human DNA.

In Situ Hybridization

DNA from 2F5 cells was prepared in agarose plugs for use as PCR template (van Oremen, G.J.B. and Verkerck, In: Human Genetic Diseases , A Practical Approach IRL Press, Oxford (1986)). PCR primers (Alu-1 and Alu-2) that specifically recognize human consensus sequences located at the 5' and 3' ends of Alu segments, were used together with 2F5 template to amplify human unique sequences (Liu, et al.. submitted) . Alu-1 and Alu-2 sequences were GGATTACAGGYRTGAGCCA (SEQ ID NO: 11)^" and

RCCAYTGCACTCCAGCCTG (SEQ ID NO: 12) respectively, where Y is either pyrimidine (T or C) and R is either purine (A or G). l μg of PCR product was labelled with biotin-7-dATP using a nick translation kit (BRL cat. no. 8160SB) . Free nucleotides were removed by passing the mixture through a Worthington Sephadex column. The procedure of Pinkel, et al. was followed for in situ hybridization with modifi- cations described in Doll, et al. (Pinkel, D. et al.. Proc. Natl. Acad. Sci. USA 83:2934-2938 (1986); Doll, G. et al.. Genes. Chromosomes and Cancer 3:48-54 (1991)). Slides were viewed with a Zeiss epi-illumination photo- scope with a filter combination 48 77 09 and photographed on Kodak Ektachrome™ 160 with exposure times between 30 and 50 seconds.

RESULTS:

Analysis of Cell Lines Cell lines from each of the radiation treatment groups were analyzed with two different sets of DNA markers. Three cell lines in particular appeared very useful and formed the basis of further analyses. Cell line 2F5 had lost all the non-chromosome 19 derived markers present in the parent cell line 20XP3542-1-4. Furthermore, it had also lost the four most proximal markers from chromosome 19; PVS, BCL2, AP0C2 and CKM, while retaining the other chromosome 19 markers including ERCC1 and pl34C (D19S51) which flank DM. Hybrid line 5B3 retained even fewer markers than 2F5, however, non-contiguous pieces of chromosome 19 were present in this case. CKM, which maps between APOC2 and ERCCl, was deleted from cell line 5B3 whereas these flanking loci were present. Marker pl34C (D19S51) , the closest marker flanking DM on the distal side, was also deleted from 5B3. Nevertheless, this cell line was useful for subdividing the region of chromosome 19 distal to ERCCl. Similarly, cell line 3A3 also provides a break¬ point within this interval. 3A3 had lost several of the distal chromosome 19 markers present in cell line 2F5, while retaining pl34C (D19S51) and other more proximal chromosome 19 markers, as well as several of the non- chromosome 19 markers from the parent cell line. Thus, cell lines 3A3 and 5B3 provided a means of assigning DNA clones derived from cell line 2F5 into three intervals. Phage clones present in both 3A3 and 5B3 were assigned to interval 'A'. Those present in 3A3, but absent from 5B3, were assigned to 'B' and those absent from both 3A3 and 5B3 were assigned to interval 'C.

Further Analysis of Cell Line 2F5

DNA from cell line 2F5 was labelled and used as probe on mitotic spreads of human chromosomes. This hybridized to a single region from the long arm of chromosome 19. The human DNA content was also characterized by PFGE. DNA from the hybrids 2F5 and 20XP3542-1-4 was digested with Notl, Mlul and BssHII and separated by PFGE. A blot of the gel was hybridized with total human DNA. 2F5 has a considerably reduced human DNA content compared to its parent cell line. In the Notl digest, fragments hybridizing with human DNA of approximately 50, 180, 200, 400, 500, 1000 and 1300 kb were present. The largest fragment was not present in the parental cell line and was probably due to a translocation between the end of the human DNA in 2F5 and a hamster chromosome. In situ hybridization with labelled human DNA onto chromosome spreads of cell line 2F5 indicate that two such fragments should be present. The 1300 kb fragment is probably mostly hamster DNA. Furthermore, hybridization with single-copy probes showed that some of the other larger fragments were due to partial digestion. Based on the Notl digestion, it was estimated that the human DNA content of 2F5 is approximately 2Mb. Construction and Screening of Libraries from 2F5

Three different libraries were made from cell line 2F5. The first was a total genomic library constructed in lambda Dash. 3.5 x 10^s recombinant phage were screened 5 with human cot 1 DNA and 230 phage containing human inserts identified (approx. 0.06%). Given a diploid cell content of 4 x 10⁹ base pairs, this should give a human DNA content of 2.5 megabases in general agreement with the estimates from PFGE.

10 Thirty-five clones were localized with a mini hybrid-panel and subdivided into three intervals; A, B and C. Eighteen clones, present in both hybrid cell lines 5B3 and 3A3, were assigned to interval A. Four clones mapped to cell line 3A3 but not 5B3 and were assigned to interval

15 B, and thirteen mapped to neither 3A3 nor 5B3 and were assigned to interval C.

The other two libraries were constructed from PFGE fractionated DNA as described above. The marker D19S63 showed marked linkage disequilibrium and no recombination

20 with the DM locus (Harley, H.G. et al.. Am. J. Hum. Genet. 49:68-75 (1991)). In order to obtain more cloned DNA and identify potential coding sequences in the vicinity of this marker, libraries were constructed from the 200 kb Notl fragment identified by D19S63. A total of

25 45 human clones were isolated, 5 of which were Notl end clones. These numbers were reduced to 24 and 2, respectively, when duplicate clones were eliminated. One of the Notl end clones (lambda 5) was used to extend the PFGE map as described below. All of the clones were

30 digested with SadI, an enzyme that generally cuts within HTF islands (Lindsay, S. and A.P. Bird, Nature 327:336-338 (1987)) . Six clones with SacII^" sites were identified. The phage DNAs were subcloned into plasmids and the fragments containing the Sacli sites were digested with Hpall. In all 6 cases, multiple Hpall sites were present, thus confirming that they represent genuine HTF islands.

Five of the HTF island subclones (p20.1, p36.1, 037.1, p42.3 and p56.1) gave unambiguous localizations on the PFGE map, and mapped to interval 'A' as defined above. Two of these clones (p20.1 and p36.1) were not on the same Notl fragment as D19S63. It is possible that the original Notl digest used in the library construction was .incom¬ plete, resulting in a contaminating 250 kb Notl fragment that was not completely resolved from the 200 kb fragment by the preparative PFGE. All of the HTF island subclones detected sequence conservation by zoo-blot analysis and were used to screen cDNA libraries.

Long Range Restriction Map of the DM Region

A number of single-copy probes from the libraries made from 2F5, together with some existing markers for this region, were used to complete the PFGE map of the DM region of chromosome 19. Notl and Mlul were.the sites principally used for the 2 enzymes. Probes containing or adjacent to Notl sites, obtained by selectively cloning the ends of Notl fragments,^• by chromosome walking or by screening phage clones by Notl digestion, were particu¬ larly useful in the construction of the map. Many of the sites identified showed partial digestion.

Part of the PFGE map has been previously reported. (Harley, H.G. et al.. Am. J. Hum. Genet. 49:68-75 (1991)) . In the data presented herein, the gap in the previous map by isolation of a Notl end clone (lambda #5) and the corresponding linking clone (lambda M23B) has been closed. A probe derived from the distal half of the latter identified the same 50 kb Notl fragment as does p36.1, which in turn identifies a 40 kb Mlul fragment and a 450 kb partial digest Mlul fragment. The 450 kb fragment was also identified by D19S51 (pl34C) . Since the latter marker was in interval 'B' as defined by X-ray hybrid mapping, and all of the former markers were in interval 'A', the breakpoint between intervals 'A' and 'B' must be within the 450 kb Notl fragment.

Sequence Conservation and cDNAs

Two of the clones identified in the initial genomic library screen, which map to interval A, lambda MW and lambda M2C showed hybridization to the rodent lane on southern blot analysis indicating sequence conservation. These clones were distinct from the HTF island clones described above. Fragments of each of these phage were screened against a muscle cDNA library and clones identi- fied. Each of these clones was purified, sub-cloned and hybridized back to the mapping filters. Both cDNAs localize to 19Q13^*3-19qter and map back to interval A of hybrid cell line 2F5. Human positive bands were present in cell lines 5B3, 3A3, GM89A99C7 and PK-87-19. Hamster bands were also present in 5B3, 3A3 and PK-87-19. The other bands in GM89A99C7 were derived from mouse.

In order to produce a cell line which will provide a source of DNA markers close to the DM locus, two tradi¬ tional approaches were used. As a starting point, cell line 20XP3542-1-4 (Stallings et al.. Am. J. Hum. Genet. 41:144-153 (1988)) which contains a single human element 20-30 megabases in size derived from at least four differ¬ ent chromosomes including a small part of 19q was used.

Of the two strategies adopted, group l, in which the parental cell line was lethally irradiated and fused to DNA repair deficient cell line UV20 followed by selection of clones in mitomycin C and polio virus, produced the most useful clones, in particular 2F5. Data from both DNA marker analysis and in situ hybridization indicated that the human material present in one of these clones (2F5) was derived exclusively from a small region of 19ql3. DNA from this cell line has been used for library construction and subsequent analysis. One other cell line from this group, 5B3, was also valuable as it provided a subdivision of the interval covered by 2F5.

A further useful cell line was produced in group 2. Hybrid 3A3 resulted from the exposure of the parental cell line to lethal dose irradiation followed by fusion to HPRT deficient hamster cells. As described by Cox et al. (1989) , no selection was employed for the region of interest. Cell line 3A3, like hybrid 5B3 from group 1, lacked some of the markers distal to ERCCl. Together these hybrids have been used to sub-divide this part of chromosome 19 into three intervals. The hybrid 2F5 provided a source of DNA specific for the region of chromosome 19 distal to ERCCl. By pulsed field gel electrophoresis, it was estimated that the size of this region was 2000kb, and a long range restriction map covering 1600kb was constructed. Two lines of evi- dence suggest that the DM gene was located within this interval. Firstly, crossovers have been reported between ERCC1 and DM, and between D19S51 and DM, indicating that the order of markers is ERCCl - DM - D19S51 (Johnson et al.. 1990; Smeets et al.. 1991) . Both of these markers flanking DM were present in the 2F5 hybrid cell line. Secondly, it had been shown that there is a strong linkage disequilibrium between DM and D19S63, but a lack of disequilibrium between DM and either ERCCl, CKM, D19S51 or D19S62 (Harley et al.. 1991b: Johnson et al.. 1990) . Of all the markers used in linkage analysis with DM, D19S63 was apparenting the closest to the DM gene. Thus, these results indicated that D19S63 was located between ERCCl and D19S51.

Because of the localization of the flanking markers ERCCl and D19S51 to intervals A and B respectively, it was reasonable to determine that the DM gene could be located within either or both of these intervals. Twenty seven phage clones were derived from 2F5 libraries mapped to intervals A and B. These clones were used to form a single contig across this interval by chromosomal walking from multiple points using the other human clones described herein, identified from the 2F5 libraries. With this information, it is was possible to screen for conserved sequences and to identify fetal muscle, fetal brain and adult brain cDNA clones, that were tested by DNA sequencing and utational analysis as candidates for the DM gene. The full-length DNA sequence of the DM gene was- determined and is shown in Figure 6 (SEQ ID NO:13), along with deduced amino acid sequences for three predicted reading frames (sequence a, upper line, SEQ ID NO:14; sequence b, middle line, SEQ ID NO:15; sequence c, lower line, SEQ ID NO:16). Preliminary results indicate that sequence c is the amino acid sequence resulting from the correct predicted open reading frame (the M indicated by the arrow on sequence c indicates the predicted start of the correct reading frame) . Also indicated on Figure 6 (lines and arrows) are the locations of two cDNAs comprising the gene sequence, cDNA 41 and cDNA 28.

As shown in Figure 7, DNA sequence analysis of multiple cNDA clones identified several DM gene variants, suggesting that the DM gene message undergoes alternative splicing. For example, cDNA 28 (isolated form an adult brain library) contains 4 bases at the 5' end of exon 14 which are not present in either cDNA 35 (from fetal brain) or 41 (from fetal muscle) . Alternatively, both cDNA 35 and 41 contain 15 bases at the 5' end of exon 9 which are not present in cDNA 28. Finally, cDNA 41 does not conatin exon 14. These data reasonably imply that structurally distinct forms of the DM kinase are expressed, possibly in a temporal, tissue-specific, or disease specific pattern. The clones reported here can also be used to test for the presence of coding sequences using the exon-amplification technique of Bucker et al.. PNAS 8^:4005-4009 (1991)).

Equivalents

Those skilled in the art will recognize, or be able to ascertain using no more than routine experimentation, many equivalents to the specific embodiments of the invention described herein. Such equivalents are intended to be encompassed by the following claims.

Claims

1. Isolated DNA having all or a portion of the nucleo¬ tide sequence of Figure 1.

2. A DNA probe which hybridizes to at least a portion of the nucleotide sequence of Figure 1.

3. Isolated DNA having the nucleotide sequence of Figure 5 or a portion thereof.

4. Isolated DNA of Claim 3 wherein the nucleotide sequence consists essentially of nucleotides 1 through nucleotide 1747, inclusive.

5. Isolated DNA of Claim 3 which encodes a protein kinase.

6. An isolated polypeptide comprising the amino acid sequence of Figure 5.

7. Isolated DNA having the nucleotide sequence corresponding to all or a portion of the nucleotide sequence of Figure 6.

8. Isolated DNA of Claim 7 which encodes a protein kinase.

9. Isolated DNA of Claim 7 which encodes a protein kinase comprising the amino acid sequence selected from the group consisting of sequence a, sequence b and sequence c, as shown in Figure 6.

10. A DNA probe which hybridizes to at least a portion of the nucleotide sequence of Figure 6.

11. An in vitro method of determining whether a CTG triplet repeat which is expanded in chromosome 19 DNA in individuals affected with myotonic dystrophy is present in an individual in a copy number comparable to the copy number of the CTG triplet repeat present in DNA from an individual affected with myotonic dystrophy, comprising the steps of: a) obtaining chromosome 19 DNA from an individual; and b) determining the copy number of the CTG triplet repeat in the chromosome 19 DNA wherein if the chromosome 19 DNA contains at least 50 CTG triplet repeats, the copy number is comparable to the copy number of the CTG triplet repeat in an individual affected with myotonic dystrophy.

12. An in vitro method of determining whether an individual is likely to be affected with myotonic dystrophy, comprising the steps of: a) obtaining DNA from the individual; and b) determining the copy number of CTG triplet repeats on chromosome 19 DNA present in the DNA obtained in (a) , wherein if the copy number of CTG triplet repeats is at least 50, the individual is likely to be affected with myotonic dystrophy, and if the copy number of CTG triplet repeats is significantly greater than 50, 5 the individual is likely to be severely affected with myotonic dystrophy.

13. The method of Claim 12 wherein chromosome 19 DNA is band 19ql3.3 DNA.

14. An in vitro method of determining whether an

10 individual is likely to be affected with myotonic dystrophy, comprising detecting expression, in a tissue affected by myotonic dystrophy, of a protein kinase having the amino acid sequence encoded by DNA of Figure 5 or Figure 6.

15 15. The method of Claim 14 wherein expression of the protein kinase is determined in heart tissue, muscle tissue or brain tissue.