Alu-Alu Recombination Deletes Splice Acceptor Sites and Produces

THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1987 by The American Society of Biological Chemists, Inc.

Vol. 262, No. 7, lssue of March 5 , pp. 3354-3361,1987 Prmted in U. SA.

Alu-Alu Recombination Deletes Splice Acceptor Sites and Produces Secreted Low Density Lipoprotein Receptor in a Subject with Familial Hypercholesterolemia*

(Received for publication, September 23, 1986)

Mark A. Lehrman, David W. Russell, Joseph L. Goldstein, and Michael S. Brown From the Deoartments of Molecular Genetics and Internal Medicine, University of Texas Health Science Center at Dallas, Southwestern Medical S&ool, Dallas, Texas 75235

A Japanese subject with homozygous familial hypercholesterolemia was found to have a 7.8-kilobase deletion in the gene for the low density lipoprotein receptor. The deletion joins intron 15 to the middle of exon 18, which encodes the 3’ untranslated region, thereby removing all 3’ splice acceptor sites distal to intron 15. By S1 nuclease mapping, we demonstrated that the 5’ splice donor site of intron 15 is no longer used, Instead a continuous transcript is produced in which exon 15 is followed by the remaining segments of intron 15 and exon 18. The translational reading frame of exon 15 continues for 165 nucleotides into intron 15 before a termination codon is reached. This mRNA should produce a truncated receptor that lacks the normal membrane-spanning region and cytoplasmic domain and that has 55 novel amino acids at its COOH terminus. A cDNA expression vector containing this sequence produced a receptor that behaved similarly to the truncated protein produced by the Japanese patient, i.e. >90% of the receptor was secreted from the cell, and the receptors remaining on the surface showed defective internalization. The deletion in this subject resulted from a recombination between two repetitive sequences of the Alu family, one in intron 15 and the other in exon 18. To date, Alu sequences have been observed at the deletion joints of all four gross deletions in the low density lipoprotein receptor gene that have been characterized. Within these Alu sequences, six out of the seven breakpoints have occurred in the left arm. These data suggest that recombination between Alu sequences may be a frequent cause of deletions in the human genome.

Mutations that abolish the function of the low density lipoprotein (LDL)’ receptor underlie familial hypercholesterolemia (FH), a disease that affects 1 out of 500 people in most countries (1). Study of these mutations has begun to reveal regions of the LDL receptor that are crucial for its specific functions, i.e. movement to the cell surface (2, 3), binding of

* This work was supported by Research Grants HL 20948 and HL 31346 from the National Institutes of Health, Research Career De- velopment Award HL 01287 from the National Institutes of Health (to D. W. R.), and a postdoctoral fellowship from the Jane Coffin Childs Memorial Fund for Medical Research (to M. A. L.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduer- tisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The abbreviations used are: LDL, low density lipoprotein; FH, familial hypercholesterolemia; kb, kilobase; Mops, 3-(N-mOrphO- 1ino)propanesulfonic acid; SDS, sodium dodecyl sulfate.

LDL (4), anchoring in the cell membrane (5), and internalization of LDL by receptor-mediated endocytosis in coated pits (5-7).

The human LDL receptor is a transmembrane glycoprotein of 839 amino acids (8,9). The NH2-terminal 767 amino acids, encoded by exons 2 through part of 16, face the extracellular environment and contain the LDL binding site. The remain- der of exon 16 and part of exon 17 encode a hydrophobic sequence of 22 amino acids that spans the plasma membrane. Portions of exons 17 and 18 encode a 50-amino acid cytoplasmic domain at the COOH terminus of the protein. Exon 18 also encodes a large 3’ untranslated region of 2.5 kilobases (kb) that includes several copies of a repetitive sequence of the Alu family.

Mutations that disrupt the cytoplasmic domain prevent the receptor from entering coated pits on the cell surface, thus creating the “internalization-defective” phenotype (5-7). The first such mutation to be elucidated at a molecular level occurred in a subject termed FH 274 (5). The defective gene had undergone a deletion that joined intron 15 to the noncoding region of exon 18, thus eliminating the exons encoding the membrane-spanning and cytoplasmic domains. The deletion resulted from a recombination between two repetitive sequences of the Alu family. Most of the receptors produced by this mutant gene were secreted from the cell, but a small fraction remained on the surface where they gave rise to the internalization-defective phenotype. We postulated that the internalization defect resulted from the lack of a normal cytoplasmic domain (5). However, the splicing pattern of the deleted mRNA was not explored, and the precise COOH terminus of the protein was not defined.

Subsequently, three other mutations all affecting the cytoplasmic domain of the receptor were found to prevent internalization of LDL (6, 7). These include a premature termination codon that leaves only 2 residues of the normal cytoplasmic domain, a substitution of a cysteine for a tyrosine residue a t position 807 in this domain, and a frameshift mutation that alters the sequence of the cytoplasmic domain.

In the current study, we report the basis of a new internalization-defective mutation that deletes the COOH-terminal end of the LDL receptor. This mutation occurred in a Japa- nese FH homozygote. Previous studies by Yamamoto and CO- workers (10, 11) have shown that this individual produces LDL receptors that can be visualized by electron microscopy on the cell surface, where they bind LDL but fail to migrate to coated pits. This individual appears to be a true FH homozygote, and not a genetic compound, since the parents were first cousins. Through use of genomic cloning and DNA sequencing techniques, we have found that the LDL receptor gene from this Japanese homozygote (designated MN in pre-

3354

Abnormally Spliced LDL Receptor 3355

vious studies (10, 11) and FH 781 in the current study) bears a deletion that joins intron 15 to exon 18. The Japanese mutation in FH 781 resembles the previously reported FH 274 mutation (5) in that it involves a recombination between two repetitive elements of the Alu family, one in intron 15 and the other in the noncoding region of exon 18. However, different Alu sequences and different mechanisms of recombination are involved in the two mutations.

The current data provide evidence that the deletion in both mutants leads to the production of an abnormally long mRNA in which the 5' end of intron 15 is contiguous with the 3' end of exon 18. The open reading frame of the receptor mRNA extends into intron 15, so that the resultant protein should have an abnormal sequence of 55 amino acids at the COOH terminus. This sequence includes a stretch of 17 amino acids that might form an inefficient membrane anchor, thus causing the retention of some of the receptors on the cell surface and explaining how a single mutation can give rise to the dual phenotype of secreted receptors and internalization-defective receptors.

EXPERIMENTAL PROCEDURES

h4aterials-3ZP-Labeled nucleotides were purchased from New England Nuclear. Unless stated otherwise, enzymes used in cloning, sequencing, and restriction digestion were purchased from New Eng- land Biolabs. Escherichia coli RNA polymerase was purchased from Sigma. Terminal deoxynucleotidyltransferase was from Life Sciences. S1 nuclease was purchased from Bethesda Research Laboratories. Bacteriophage X Charon 35 was obtained from Dr. Phil Tucker of the University of Texas Health Science Center at Dallas, Dallas, TX. E. coli host strain DB1255 (relevant genetic markers, recBC-, sbcB-) was provided by Dr. Arlene Wyman of the Massachusetts Institute of Technology, Boston, MA (12). X DNA packaging extracts were purchased from Amersham Corp. M13 replicative form DNAs were purchased from Pharmacia P-L Biochemicals. Skin fibroblasts from FH 781 (previously referred to as patient MN in Refs. 10 and 11) were kindly provided by Dr. Akira Yamamoto (National Cardiovas- cular Center Research Institute, Osaka, Japan).

Genomic Cloning-High molecular weight DNA (100 Gg) was isolated from the skin fibroblasts (5, 6) of FH homozygote 781 and restricted for 4 h at 37 "C with a 4-fold excess of XbaI. The restricted DNA was extracted with a mixture of phenol and chloroform and then chloroform, precipitated with 70% ethanol and 86 mM sodium acetate, and dissolved in 100 p1 of 10 mM Tris-chloride and 1 mM sodium EDTA at pH 7.5. The DNA was fractionated by centrifugation through a 10-40% sucrose gradient. Fractions from the gradient were subjected to Southern blotting with a uniformly 32P-labeled single- stranded cDNA probe (13) corresponding to exon 13 of the LDL receptor gene (8, 9). Aliquots (50 ng) of fractions containing the desired 10.5-kb XbaI fragment were mixed with 0.5 pg of XbaI- digested arms of X Charon 35 (14) and incubated (16 h, 14 "C) with 490 units of T4 DNA ligase. The ligated material was packaged into X phage particles in vitro to yield a total of 4.0 X lo4 plaque-forming units and screened in E. coli strain DB1255. One recombinant clone designated XFH 781-11 was identified in the library using a cDNA probe that hybridizes to the 3' end of exon 18 of the receptor gene (9). The insert was excised at the vector EcoRI sites and subcloned into EcoRI-digested pUC-18 (Pharmacia P-L Biochemicals). A restriction map of the cloned insert was generated by standard techniques (15) to verify that the size of the insert (10.5 kb) was in agreement with that measured by Southern blotting of FH 781 genomic DNA.

DNA Sequencing-The normal sequence of exon 18 has been previously reported (8). The sequence of intron 15 in the region of the deletion joint between the NsiI and Hind111 sites (see Fig. 2) was obtained with a combination of the following strategies. 1) A 0.45-kb PstI-Hind111 fragment (which contained the NsiI site) was subcloned into M13mp18 replicative form DNA and sequenced by the method of Sanger et al. (16). 2) A double-stranded plasmid p381-20 (17), which contained a 17-kb BamHI fragment of the normal gene extending from intron 10 to exon 18, was denatured with 0.4 N sodium hydroxide, annealed with a specific oligonucleotide (5' AAATAATGCATTAGA 3') that primed at the NsiI site (underlined), and sequenced by a modification of the method of

Sanger et al. (16). 3) p381-20 was digested with NsiI in the presence of E. coli RNA polymerase (18), end labeled with [a-32P]cordycepin and terminal deoxynucleotidyltransferase, and sequenced by the method of Maxam and Gilbert (19).

The sequence of the deletion joint in FH 781 between the NsiI and PvuII sites (Fig. 2) was obtained from plasmid p781-11 by a combination of the following strategies. 1) The specific oligonucleotide 5' AAATAATGCATTAGA 3' was used as a primer on plasmid p781-11 by a modification of the method of Sanger et al. (16) as described above. 2) The NsiI site was labeled for sequencing by the Maxam and Gilbert technique (19) as described above. 3) The PuuII site was labeled with [LU-~'P]~GTP and the Klenow fragment of E. coli DNA polymerase, followed by Maxam-Gilbert sequencing.

Immunoprecipitation of 35S-Labeled LDL Receptors from Cultured Cells-Human fibroblasts (20) and transfected hamster cells (21) were cultured for 4-6 days as described in the indicated reference. In the case of fibroblasts, the cells were induced for LDL receptor synthesis by incubation in the absence of lipoproteins as described (20). Cells were incubated with [35S]methionine as indicated in the figure legends. The cells and medium were then harvested and immunoprecipitated with an LDL receptor-specific monoclonal antibody (IgG-C7) (51, and the immunoprecipitates were subjected to SDS-polyacrylamide gel electrophoresis and autoradiography as previously described (5).

Blot Hybridization of RNA-Fibroblasts were cultured for 6 days in Dulbecco's modified Eagle's medium containing 10% fetal calf serum (20) and then induced for maximal LDL receptor expression by incubation for 14 h in medium containing 10% Cabosil-treated lipoprotein-deficient serum (22) and 10 p M compactin or suppressed for LDL receptor expression for 14 h in medium containing 10% fetal calf serum, 12 gg/ml cholesterol, and 3 pg/ml25-hydroxycholestero1. Cells were scraped from the dish in phosphate-buffered saline and pelleted by low speed centrifugation. Total RNA was isolated by a guanidinium HCl method (23). Briefly, fibroblasts from 30 dishes (100 mm) were resuspended in 10 ml of 6 M guanidinium HCl, 1% sodium lauryl sarcosine, 10 mM dithiothreitol, 0.1 M KOAc at pH 5.0 and subjected to homogenization with a Braun Polytron at medium speed for 15 s and high speed for 5 s. The resulting slurry was passed three times through a 22-gauge needle to shear any remaining high molecular weight DNA. One-half volume (5 ml) of cold ethanol was added, and nucleic acids were precipitated by incubation at -20 "C for a t least 6 h. The fluffy white precipitate was collected by centrifugation at 10,000 rpm for 10 min at -10 "C in a Sorvall RC-5B centrifuge with an HB4 rotor. The resulting pellet was resuspended in 4.75 ml of 6 M guanidinium HCl and 0.25 ml of 0.5 M EDTA, pH 7.0, by repeated vortexing. One-twentieth volume (0.25 ml) of 2 M KOAc, pH 5.0, was then added, and nucleic acids were precipitated with one-half volume of cold ethanol as described above. After centrifugation, the pellet was resuspended in 2 ml of 20 mM EDTA, pH 7.0. This solution was extracted with an equal volume (2 ml) of chloroformfiutanol in a 4:l ratio, respectively, by vortexing vigor- ously for 5 min. The aqueous and organic layers were separated by a 5-min centrifugation at 2,000 rpm in a tabletop Beckman centrifuge, and the upper aqueous phase was removed and stored in a sterile tube on ice. The organic layer was re-extracted with 2 ml of 20 mM EDTA, pH 7.0, and the aqueous and organic layers were separated as above. The aqueous layers were pooled (final volume, approximately 4 ml), and an additional 2 ml of 20 mM EDTA was then added. The resulting nucleic acids were precipitated by the addition of 0.3 ml of 2 M KOAc, pH 5.0, and 2% volumes (15 ml) of cold ethanol. After incubation at -20 "C for at least 6 h, the nucleic acids were collected by centrifugation. The resulting pellet was overlaid with 5 ml of 70% ethanol and centrifuged briefly, after which the 70% ethanol wash was decanted. The pellet was then dried in vacuo and resuspended in 1-2 ml of sterile HzO previously treated with diethylpyrocarbonate (15). The absolute yield of RNA recovered by this procedure varied depending on the fibroblast strain used for isolation but averaged around 0.5-1.0 mg/30 dishes (100 mm) of cells.

For blotting analysis, varying amounts of RNA were treated with glyoxal (23) and size fractionated by electrophoresis in sterile 1.5% agarose gels containing 40 mM Mops, pH 7.0. The RNA was transferred to Zeta Probe membranes (obtained from Bio-Rad) by capillary blotting in 20 X SSC for 16 h (1 X SSC = 0.15 M NaCl and 15 mM sodium citrate). Care was taken to ensure that an excess of blotting buffer was used and that the paper towels employed as a wick did not become completely saturated with buffer. Failure to meet these cri- teria resulted in the appearance of a high background on the autora- diogram. The filters were baked for 2 h at 80 "C in vacuo and then

3356 Abnormally Spliced LDL Receptor

boiled for 5 min in 20 mM Tris-HC1, pH 8.0. Prehybridization (22 h) and hybridization (216 h) were carried out as described (24), except that 1% SDS and single-stranded 32P-labeled probes were used. Three probes of about 200 nucleotides in length and complementary to different regions of the LDL receptor mRNA were synthesized by the methods of Church and Gilbert (13). The estimated specific activity of each probe was >lo9 cpmlpg. Hybridization reactions contained lo6 cpm/ml of each of the three probes. Following hybridization, filters were washed twice at room temperature for 10 min in 2 X SSC and 1% SDS and then once at 68 "C for 75-120 min in 0.1 X SSC and 1% SDS. The filters were air dried and used to expose Kodak XAR-5 film with Du Pont Cronex Lightning Plus intensifying screens at -70 "C.

SI Nuclease Analysis-Single-stranded hybridization probes were prepared as described by Reynolds et al. (25). Total cellular RNA was prepared from fibroblasts as described above. One fmol of probe, labeled to a specific activity of 2 X lo7 cpm/pg, was mixed with 25 pg of total cellular RNA, hybridized for 16 h at 37 "C in the presence of 50 ng of transfer RNA, and then digested with 300 units of S1 nuclease for 60 min at 37 "C (25). The hybrids were analyzed by electrophoresis on 6% polyacrylamide, 7 M urea gels.

Construction of Expressible cDNA Corresponding to mRNA of FH 781 and FH 274-A 385-base pair PstI-NdeI fragment of the normal LDL receptor gene containing 161 base pairs of exon 15 and 224 base pairs of intron 15 (9) was treated with the Klenow fragment of E. coli DNA polymerase in the presence of 1 mM dTTP and 1 mM dATP to generate a blunt end at the NdeI site and then subcloned into PstI- and SrnaI-digested M13mpll DNA (26). The insert was excised from the M13 template by hybridization with the universal primer, exten- sion with the Klenow fragment and four deoxynucleoside triphos- phates, digestion with EcoRI, blunt ending with Klenow fragment, and digestion with PstI. After purification by gel electrophoresis, the excised insert was mixed with two fragments: 1) a 0.45-kb BglII-PstI fragment of pDL-2 which contains exons 12-15 of the LDL receptor cDNA2; and 2) an 8-kb BglII-SmaI fragment of pDL-2 which contain exons 1-11 of the LDL receptor cDNA located between the SV40 early region promoter and the late region polyadenylation site. This plasmid also contains a cDNA encoding a mutant methotrexate- resistant dihydrofolate reductase (27).' These three fragments were ligated in an equimolar ratio for 16 h at 14 "C with T4 DNA ligase (15). Following transformation into E. coli DH5 cells (Bethesda Research Laboratories), colonies were screened with an appropriate probe, and plasmids were characterized by restriction endonuclease mapping. All ligation junctions were sequenced with specific oligonucleotide primers (16). The resultant plasmid was designated pIVS- E A .

Isolation of Transfected Cell Lines-Plasmid pIVS15-A was co- transfected with pSV3-Neo into ld&-7 cells, a mutant line of Chinese hamster ovary cells lacking functional LDL receptors (29), as previously described (30). Positive colonies were selected with G418 and identified by indirect immunofluorescence (30) after permeabilization with 0.1% (w/v) Triton X-100 at -20°C for 5 min. The positive colony selected for study was cloned, subcloned, and maintained as a permanent cell line as described (30). This cloned cell line is designated TR910-2. TR715-19 cells, which express the normal human LDL receptor cDNA encoding plasmid pLDLR-2 (8), were isolated as previously described (21).

LDL Receptor Assays-Binding, internalization, and degradation of Iz5I-LDL were measured in monolayers of intact cells as previously described (20).

RESULTS

Blot hybridization of genomic DNA from subject FH 781 revealed an abnormal restriction fragment when the DNA was restricted with XbaI and probed with a 32P-labeled cDNA probe derived from the 3' end of exon 18 (Fig. 1). In DNA isolated from a normal subject, a 6.6-kb band is detected (Lane C ) , whereas the FH 781 DNA showed a 10.5-kb band (Lane D). Digestion with SphI did not reduce the size of the XbaI fragment in the normal DNA (Lane A ) , but it did reduce the size of the fragment produced by the FH 781 DNA (Lane B ) . One possible interpretation of this data is that the FH 781 DNA contains a deletion that removes an XbaI site,

D. W. Russell, unpublished observations.

kb

23.0 - 9.4 - 6.6 - 4.4 -

I XbaI X b a I

A B C D FIG. 1. Blot hybridization of genomic DNA. Genomic DNA

(5 pg) isolated from cultured fibroblasts of the indicated subject (5) was digested with XbaI alone or in combination with SphI. Fragments were subjected to electrophoresis, transferred to nitrocellulose, and hybridized with a 32P-labeled probe that corresponds to the 3' end of exon 18 (nucleotides 4965-5065 of the normal cDNA) (8). Molecular size standards were generated by Hind111 cleavage of bacteriophage X DNA.

Exon NO. 1314 15 16 j7 I 8

LDL Receptor Gene """ . j.

FIG. 2. Comparison of restriction maps of the 3' end of the normal LDL receptor gene and the deletion-bearing gene from FH 781. Exons are indicated by solid uertical bars; introns are indicated by open segments; and known Alu repetitive elements in intron 15 and exon 18 are indicated by striped bars. Arrowheads indicate the relative orientation of each Alu sequence relative to the consensus structure and represent either complete (closed arrowhead) or half (open arrowhead) repeats. Restriction enzyme recognition sites used to define the deletion in FH 781 are shown; the StuI site shown was located by DNA sequencing. The map of the normal allele is from Sudhof et al. (9); the map of the FH 781 allele was generated from XFH 781-11 (see "Experimental Procedures"). Solid bars above or below the restriction maps denote segments that were subjected to DNA sequencing.

thereby joining two XbaI fragments. The resulting abnormally large XbaI fragment now includes an SphI site. Furthermore, the absence of fragments in the FH 781 digest that comigrate with the normal fragments suggests that this patient is a true FH homozygote.

To learn more about this putative deletion, we isolated the abnormal XbaI fragment by cloning in bacteriophage X. Fig. 2 compares restriction maps of the cloned FH 781 insert and the 3' end of the normal LDL receptor gene as determined from a previously cloned fragment (9). The results indicate that the FH 781 allele harbors a deletion of approximately 7.8 kb that begins in the middle of intron 15 and terminates in the middle of exon 18.

A DNA fragment encompassing the deletion joint from FH 781 and a fragment from intron 15 of the normal LDL receptor gene were isolated by subcloning and subjected to sequencing (Fig. 3). The sequences of normal intron 15 in the region of the deletion joint and exon 18 were found to be homologous. Computer-assisted analysis revealed that both DNA sequences correspond to the left tandem repeat (arm) of an Alu


-168 -117 -115 -63 I \ ( I

In t ron 15 5 ‘ N s i l c 1 2 n t + GGATTCtlGGGCAGGGCACA-GTGGCTCACACCTGlMTCCCAGCAClTTGGGAGGCTGAGGlG

Exon 18 5‘ Sstl 4 1 n t -w GAATT--TGGGCAG-ACACAGGTGGCTCACACACCTGTAATCCCAGCACT~lGGGAGGCTGAG~lGG----ATCACTlGAGTTCAGGAGTTGGAGACC~GAGC-MCAAAGCGAGA

FH 781 5‘ Nsi I 4-12 n t + GGATlCCTGGGCAGGGCACA-GlGGClCACACCTGl~TCCCAGCACTllGGGAGGCTGAGGlGG----ATCAClTGAGTTCAGGAGTTGGAGACC~GAGC-MCAMGCGA~

TGGATCACCTGAGGTCAGGAGTTTGAGACCAG-CCTG-GCCMCATGGlGPA4 0 .W . e 0 e e..

e .* .. Stul

FIG. 3. Nucleotide sequence across the deletion joint in the mutant allele of FH 781 and the corresponding regions of the normal gene, The normal sequence of intron 15 (top line), the sequence across the deletion joint in FH 781 (middle line), and the normal sequence of exon 18 (bottom line) are aligned. Gaps are indicated by dashes, and dots indicate positions at which the sequences differ. The vertical lines denote the postulated limits of the deletion joint. The horizontal Ziw above the sequence indicates the portion homologous to the consensus Alu sequence. This sequence is numbered according to the scheme used for the consensus Alu sequence (31). The StuI restriction site is denoted by the solid underline.

E m n NO

“-8”b”--r

1314 45 16 17 18

5’ -3 Ja LDL Receptor

Gene

I I .0.

-132nt -166nl

-168 -36 ‘1SO

I I N*l I

l ” ” I nlndm StVI Pwn

Normal Allele

Allele N r ~ 1 Stul Prvn c “. ”, ,”” 100 bp

H

FIG. 4. Localization of the deletion boundary in FH 781 to the left arm of an Alu repeat. The 3’ end of the LDL receptor gene and Alu repeats in intron 15 and exon 18 are shown schematically at the top, as described in the legend to Fig. 2. The deletion endpoints are shown in expanded form in the middle of the figure, and the resulting deletion joint is shown at the bottom. The solid and broken horizontal arrows denote the extent and orientation of se- quencingperformed with the methods of Sanger et al. (16) and Maxam and Gilbert (13), respectively (see “Experimental Procedures”). The structure of a consensus Alu repeat (32) with left (striped) and right (stippled) arms is shown in the box; the orientation and size of Alu repeats in the LDL receptor gene are indicated by the arrowheads, as described in Fig. 2.

element (Fig. 3). Within the homologous region, we have numbered the nucleotides according to the consensus num- bering system for Alu repeats (31). The sequence of the deletion joint in FH 781 shows a perfect match with the normal sequence of intron 15 from nucleotide position -168 t o -115, which is to the left of the vertical lines in Fig. 3. At position -114, the homology with intron 15 decreases. The sequence from exon 18 is completely homologous with the sequence in FH 781 to the right of the vertical lines, but the homology diminishes to the left of the vertical lines beginning with a base mismatch a t position -118. These results suggest that the deletion joint lies between nucleotides -117 and -115, the region enclosed by the vertical lines in Fig. 3.

Alu sequences consist of two tandem repeats referred to as the left arm (-132 nucleotides in length) and the right arm (-166 nucleotides in length), the latter being a duplication of the left arm with a 34-nucleotide insertion. The relation of the deletion end points in FH 781 DNA to the Alu sequence is diagrammed schematically in Fig. 4, from which it can be seen that the deletion occurs near the middle of the left arm of the Alu repeat.

Fig. 5 compares the deletion in FH 781 and the previously described deletion in another internalization-defective allele, FH 274 (5). Both deletions join an Alu repeat in intron 15 (designated A-E in Fig. 5) with an Alu repeat in exon 18

Exon No 4314 15 .(6 17

5’ 3’ Normal

I UGA

5 5 kb

5’ 3’ FH 274

< ,, UAA

7 0 hb

5’ 3’ FH 781

t & f UAA

FIG. 5. Comparison of the deletions in the FH 274 and FH 781 alleles and prediction of the mRNA structures. The stmc- tures of the 3’ end of the normal LDL receptor gene and the deleted genes in FH 274 and FH 781 are shown schematically with exons, introns, and Alu repeats represented by the closed, open, and striped segments, respectively. The sizes and positions of the two deletions are indicated by the open triangles. The portions of each gene that are retained in the mRNA after splicing are indicated below each gene map by the solid segments connected by lines. The vertical arrows indicate the earliest inframe termination codon for each mRNA. The open bars indicate the positions of probes used in various RNA blotting experiments.

(designated F-H). However, different repeats were involved in the two deletions. FH 274 joined Alu repeat E in intron 15 to repeat F in exon 18. FH 781 joined Alu repeat C in intron 15 to repeat H in exon 18.

Fig. 5 also shows the predicted effects that these deletions would have on mRNA structure. Both deletions remove all normal splice acceptor sites on the 3’ side of intron 15. Since no acceptor site is present, it is possible that the splice donor site at the 5‘ end of intron 15 would not be used and that these cells would produce a continuous transcript in this region extending from exon 15 through the remaining portions of intron 15 and exon 18 (Fig. 5). The predicted length for such a transcript in the FH 274 cells would be 7.7 kb and that for the FH 781 cells would be 6.1 kb.

We examined the size of the mRNA transcripts through use of agarose gel electrophoresis and blot hybridization (Fig. 6). The normal mRNA for the LDL receptor, which is 5.3 kb in length (8), migrates above the 28 S ribosomal RNA standard (Fig. 6). The FH 781 mRNA was slightly longer than normal, and the FH 274 mRNA was even longer by approximately 2 kb. These sizes are consistent with the mRNA structures predicted in Fig. 5.

To test directly for the presence of intron 15 in the mRNA transcripts, we performed an S1 nuclease analysis (Fig. 7). Total RNA from normal, FH 274, and FH 781 cells was


28 s- 23S- 16s-

IsterolsI-I+I-I-]

FIG. 6. Blotting analysis of RNA isolated from FH 274, FH 781, and normal fibroblasts. Total RNA was isolated from cells grown in the absence of sterols (induced for LDL receptor expression) or presence of sterols (suppressed for LDL receptor expression) and subjected to electrophoresis as described under “Experimental Pro- cedures.’’ Following transfer to Zeta Probe membranes, hybridization was carried out at 42 “C for 16 h with a mixture of three single- stranded uniformly 32P-labeled probes (1 x lo6 cpm/ml, >IO9 cpm/ pg for each probe) which together spanned exons 4-14 of the LDL receptor gene. After hybridization, the filter was washed as described under “Experimental Procedures” and exposed to XAR-5 film for 29 h. Horizontal lines indicate the positions to which bovine 28 S, 18 S, and E. coli 23 S ribosomal RNAs migrated in adjacent lanes.

JProbe, 418nt 7

385 nt

”161 nt

PROBE : 418 nt I Exon15 I Intron 15

nt - Normal

nt -1 and

Protected Fragments FH 274

FH 781 FIG. 7. S1 nuclease mapping of exon 15 in normal and

mutant RNAs. A 0.38-kb PstI-NdeI fragment covering 161 base pairs of the 3’ end of exon 15 plus 224 base pairs of the 5’ end of intron 15 was subcloned into PstI- and SmI-digested Ml3mpll RF DNA, and a single stranded uniformly labeled 32P-labeled probe was prepared as described under “Experimental Procedures.” The probe was hybridized with 25 pg of total cellular RNA from normal, FH 274, or FH 781 fibroblasts and then digested with S l nuclease. The relative orientations of the protected fragments with respect to the probe are indicated at the bottom of the figure. nt, nucleotides. Molecular size standards were generated by HaeIII digestion of 6x174 replicative form DNA.

hybridized with a single-stranded uniformly labeled probe that contained 385 nucleotides that spanned the junction between exon 15 (161 nucleotides) and intron 15 (224 nucleotides). The probe also contained 33 nucleotides of M13 sequence (Fig. 7). When RNA from normal cells was hybridized to this probe, S1 nuclease digestion gave rise to a protected

fragment of 161 nucleotides, consistent with the removal of intron 15 by splicing. In the FH 274 and FH 781 RNA, there was no evidence for splicing at the exon 15/intron 15 junction. The protected fragment of 385 nucleotides indicated that the probe had hybridized to mRNA species in which exon 15 and intron 15 remained in continuity. (The small proportion of 385-nucleotide species observed in normal cells probably corresponds to unspliced nuclear RNA as total cellular RNA was used in these experiments.) To test for the presence of other sequences in the mRNA downstream of exon 15, four probes whose positions are indicated by the open boxes in Fig. 5 (normal gene) were used in blot hybridization analysis of normal, FH 274, and FH 781 RNA (data not shown). The two probes from exon 18 were found to hybridize with all three mRNAs. However, the two probes from intron 15 hybridized only with FH 274 and FH 781 transcripts and not with the normal mRNA. These results are consistent with the predicted splicing pattern shown in Fig. 5.

We previously showed that the vast bulk of receptors produced by the FH 274 cells is secreted and that the remaining small fraction is found associated with the cell membrane where it binds lZ5I-LDL and gives rise to the internalization- defective phenotype (5). Furthermore, the secreted receptor on SDS-polyacrylamide gels was found to have a slightly lower molecular weight than the membrane-associated receptor (5). Fig. 8 shows that a similar phenomenon occurs in the FH 781 cells. When these cells are labeled with [35S]methionine, most of the immunoprecipitable receptors are found in the culture medium, and only a small fraction remain associated with the cell. These cell-bound receptors are presumably responsible for the previously described internalization-defective phenotype of these cells (10, 11).

How do the mutant mRNAs of Fig. 5 give rise to a protein that is partly cell associated and partly secreted? Translation of the FH 274 and FH 781 mRNAs shown in Fig. 5 should result in the same abnormal protein. In both cases the open reading frame of exon 15 continues across the unused intron splice junction and encodes an additional 55 amino acids before the first termination codon is reached. The predicted sequence of the additional amino acids is shown in Fig; 9. Within this sequence we note that there is a stretch of 17 uncharged residues (including one histidine) that is flanked on both sides by arginine residues. This sequence has many of the characteristics of a membrane-spanning sequence and

M x

160 155

M, X

160 155

FIG. 8. Recovery in the cells and culture medium of newly synthesized LDL receptor from FH 781 fibroblasts. Normal and FH 781 fibroblasts were induced for LDL receptor synthesis (20) and incubated with [35S]methionine (100 pCi/ml) for 18 h at 37 “C in medium containing 5 p~ unlabeled methionine. LDL receptors from the cells and medium of three 60-mm dishes were immunoprecipitated and subjected to SDS-polyacrylamide gel electrophoresis. Migration positions corresponding to molecular weights of 155 and 160 kDa were determined from plots of the migration distance of standard proteins, which were applied to adjacent lanes.

3359

FIG. 9. Amino acid sequence predicted by the open reading frame at the 5' end of intron 15. If this intron were not removed by splicing, these 55 residues would follow the first 749 amino acids of the normal receptor. Because of the hydrophobicity of 17 of these residues, it is likely that they would remain embedded in the membrane, forming a membrane anchor as illustrated. Amino acids having positively or negatively charged side chains are indicated by circles and squares, respectively. The underlines denote the sequences of two peptides used to generate antisera described in the text. The inset compares the orientations of the normal receptor and that postulated to be produced by the mRNA containing the unspliced intron in FH 274 and FH 781. The arrow (inset) indicates a possible proteolytic cleavage site a t two adjacent arginine residues in the mutant protein. The predicted 55-residue sequence was deduced by DNA sequencing of both strands by a combination of the methods of Sanger et al. (16) and Maxam and Gilbert (19). Plasmid pIVS15-A is predicted to express the complete protein sequence shown.

160 - 120 -

- 160" "

- 120"-". c -

FIG. 10. Expression in transfected hamster cells of the cDNA encoding the predicted FH 274FH 781 receptor. Trans- fected hamster cells were set up according to a standard format (30). Panel A, cells expressing the normal LDL receptor (TR-715-19 cells) or the predicted FH 274/FH 781 receptor (TR-910-2 cells) were labeled at 37 "C with 100 pCi/ml [35S]methionine for 90 min (pulse), and the cells and medium were either harvested immediately (Lanes 1, 3, 5, and 7) or after an additional incubation for 90 min in medium containing 200 p~ unlabeled methionine (chase) (Lanes 2, 4, 6, and 8). The cells (Lanes 1-4) and medium (Lanes 5-8) from one 100-mm dish were immunoprecipitated and subjected to SDS-polyacrylamide gel electrophoresis. The positions to which the precursor (120 kDa) and mature (160 kDa) forms of the normal LDL receptor migrated are indicated by horizontal lines. Panel B, cells were labeled at 37 "C with 75 pCi/ml [35S]methionine for 3 h (pulse), and the cells (Lanes 9 and 11) and medium (Lanes 10 and 12) were analyzed by immunoprecipitation and gel electrophoresis as described above.

could conceivably function as such in the FH 274 and 781 receptors (see below).

To ensure that the postulated amino acid sequence shown in Fig. 9 could give rise to a protein with the characteristics of the receptor observed in the FH 781 and FH 274 cells, we prepared a cDNA expression vector that contained the abnormal sequence. We excised a PstIINdeI fragment from a genomic clone that contained the 3' end of exon 15 and the contiguous 224 nucleotides of intron 15. This fragment was ligated into the corresponding PstI site of a vector containing a full-length expressible cDNA for the LDL receptor. The resultant cDNA contains the SV40 early region promoter followed by exons 1-15 of the LDL receptor, followed by a portion of intron 15 and SV40 late region polyadenylation signal. The mutated cDNA was introduced by transfection into ldlA-7 cells, a line of CHO cells that lacks LDL receptors (29). A permanent line of transfected cells, designated TR 910-2 cells, was selected for LDL receptor expression and then incubated with [35S]methionine. After 90 min the cells were harvested and the receptors were precipitated with a monoclonal anti-LDL receptor antibody. As a control we used TR-715 cells, a line of ldlA-7 cells that was transfected with the wild-type human LDL receptor cDNA (21). In the control TR 715-19 cells, after the 90-min pulse, the precursor form of the receptor (apparent molecular weight, 120,000) and the mature form (apparent molecular weight, 160,000) were seen (Fig. 10, lane I). When the labeled cells were chased for 90


Exon No 4 5 . 6 13 14 15 16 17 18

5 ' 3'

4 1 ? 4D I b I 4 7 7 9 , I , , I I 1 I , 8 ,

4-4

, I I '

-0.8 k b FH 626

I ' FH381 I

- 5.0 kb

I 1 , I I 1

1 ,

FH 274 I 1 , , I I*

I -5.5 kb I

I FH 781 e!

-8

I I I I

1 k b - - 7.8 kb

FIG. 11. Deletions involving Alu sequences in the LDL receptor gene. Portions of the LDL receptor gene are shown schematically with exons, introns, and Alu repeats represented by closed, open, and striped segments, respectively. The relative orientations of complete (-300 bp) and half Alu repeats are shown by the closed and open arrowheads, respectively. The horizontal arrows indicate the four deletions described in the text. The asterisk (*I indicates the site of a PuuII restriction fragment length polymorphism in the normal LDL receptor gene (17).

Left Arm

E -168 - 36 + 135

4 , b Globin Gene Deletions Q , D LDL Receptor Deletions

2 . 8 , p Globin I +5. FH 626

4. y . 8 . p Globin 3. FH 381

9. 8 , p Globin 6+7. FH 781

11. 6, p Globin 8 + 10. FH 274

FIG. 12. Clustering of deletion endpoints within the consensus Alu repeat for mutations in the human LDL receptor and y,b,&globin genes. A consensus Alu repeat (31) is drawn with left and right arms indicated. The locations of nine deletion joints are noted, and the direction of each deletion is indicated by the closed (globin deletions) or open (LDL receptor deletions) arrowheads. Re- gions of the Alu repeat that correspond to the A and B sequences of RNA polymerase 111 promoters (33) are shown by the solid bars. Deletions 2, 4, 9, and 11 are described in Refs. 34-37, respectively; deletions 1 , 3 , 5 , 8 , and 10 are described in Refs. 4 , 5 , and 32; deletions 6 and 7 are described in the current manuscript.

min in the presence of unlabeled methionine, all of the receptor was converted to the mature form (lane 2). None of the receptor produced by the wild-type cDNA was found in the culture medium (lanes 5 and 6).

A different result was seen in the TR 910-2 cells transfected with the mutated cDNA. Immediately after the 90-min pulse, only a small amount of mature receptor was present within the TR 910 cell (lane 3) , but a large amount was seen in the culture medium (lane 7). After the chase period, the amount of mature receptor in the TR 910 cells declined (lane 4) , whereas even more receptor accumulated in the culture medium (lune @.The receptor secreted into the culture medium appeared to be slightly smaller than the mature receptor associated with the transfected cells (compare lanes I I and 12).

DISCUSSION

The current results demonstrate that the defective LDL receptors in the FH 781 cells arose from a recombination between Alu elements in intron 15 and exon 18 of the LDL receptor gene, an event that deleted the exons encoding the transmembrane and cytoplasmic domains. In the RNA transcript produced from this gene, the reading frame of exon 15 continues into intron 15, thereby specifying 55 abnormal amino acids at the COOH terminus of the truncated receptor protein. Seventeen of these amino acids form a continuous nonpolar stretch that may have a tendency to remain associ-

ated with the membrane, serving as a rudimentary membrane anchor (Fig. 9).

Extensive prior studies have shown that at least some of the receptors produced by the FH 781 cells are attached to the surface membrane where they bind LDL but fail to carry it into coated pits (10, 11). This observation was documented by tracer studies with '251-labeled LDL (10) as well as by direct visualization of the binding of ferritin-labeled LDL (11). The latter studies showed that the LDL receptors in the FH 781 cells, like those of the other internalization-defective subjects, failed to cluster in coated pits. We have confirmed the presence of high affinity binding and lack of internalization of Iz5I-LDL in the FH 781 cells (data not shown). We also found that the receptor produced by the TR 910 cells, which were transfected with the cDNA encoding receptor that is suspected to be produced in the FH 781 cells, bound Iz5I- LDL but failed to carry it into the cell. After incubation with T - L D L for 5 h at 37 "C, the high affinity binding of lz5I- LDL in the cells transfected with the normal cDNA (T'R 715 cells) was 255 ng/mg as compared to 44 ng/mg in the TR 910 cells. The amount of internalization plus degradation was 8992 ng/mg in the TR 715 cells but only 155 ng/mg in the T R 910 cells. The internalization index (internalized plus degraded LDL divided by surface-bound LDL) was 35 in the T R 715 cells and only 3.5 in the TR 910 cells.

The mechanism by which some of the FH 781 receptors remain attached to the cell surface is not known. One possibility is that the relatively hydrophobic sequences contained in the additional COOH-terminal 55 amino acids remain attached to the membrane through hydrophobic interaction. Consistent with this interpretation is our previous finding that the surface receptors cannot be removed by treatments that break hydrophilic bonds, such as treatment with high salt, acidic or basic buffers, or EDTA (5).

If the FH 781 receptors do remain attached by virtue of a hydrophobic COOH terminus, this anchorage must be relatively weak since the bulk of the receptors in the FH 781 cells and the transfected T R 910 cells are secreted from the cell. Two potential explanations seem likely. 1) The new membrane anchor might be relatively weak in the sense that a high proportion of molecules might dissociate from the membrane after translation; the intact receptor would then be free in the lumen of the endoplasmic reticulum, from which it would be secreted. 2) Alternatively, the external domain might be released from the hydrophobic anchor by proteolytic cleavage (perhaps at the dibasic Arg-Arg sequence noted in Fig. 9), thus leading to secretion of a fragment lacking the hydrophobic sequence. The latter explanation is supported by the finding that the secreted form of the receptor migrates slightly


faster than the mature cell-associated form on SDS-polyacrylamide gels (see Figs. 8 and 10).

We attempted to resolve this issue by preparing antibodies against synthetic peptides corresponding to the predicted sequence above the hydrophobic region and at the extreme COOH terminus (Fig. 9, see underlining). Although the antibodies reacted with the synthetic peptides, they did not react with either the newly synthesized or the secreted receptor produced by the FH 781 or FH 274 cells. Immunoblotting as well as immunoprecipitation studies were negative. Because of this inconclusive result, we cannot determine whether the predicted sequence is present when the receptor is synthesized or if it is removed after the receptor is secreted.

At the DNA level, the deletion that brings about the FH 781 protein phenotype appears to have been caused by a recombination between two Alu sequences that are oriented in the same direction (Fig. 11). The most likely mechanism is one of homologous recombination in which unequal crossing over follows mispairing of homologous Alu sequences on different chromatids during meiosis. A similar homologous Alu- Alu recombination was postulated to explain a 0.8-kb deletion that occurred in the LDL receptor gene of subject FH 626 (4).

In the previously examined FH 274 gene (5), we observed recombination between Alu sequences that were oriented in opposite directions in the genome (Fig. 11). In addition, we have characterized another large deletion in the LDL receptor gene (FH 381 in Fig. 11) that involves the union of an Alu sequence in intron 15 with a non-Alu sequence in exon 13 (32). This latter sequence shows significant complementarity to the Alu sequence in an inverse orientation. In both the FH 274 and FH 381 cases, it was possible to draw complex double stem-loop structures involving the same chromosome in which the complementary sequences at the ends of the deletion joint are brought into proximity (5, 32). I t is likely that the ability to form these double stem-loop structures contributed to the formation of the observed deletions.

All together seven Alu sequences have been involved in deletions in the LDL receptor gene (Fig. 12). Six out of the seven breakpoints have occurred in the left arm, and all of these have occurred in the region between the A and B sequences that are homologous to the internal promoters that have been described in genes transcribed by RNA polymerase 111 (33). Alu sequences have also been reported to be involved in four deletions of the y,d,(l-globin gene complex (Fig. l l ) , and three of the four breakpoints also occur in the left arm, two of them between the A and B sequences (34-37).

From the data shown in Fig. 12, it is clear that the breakpoints in the Alu sequences are not located a t random. Rather, they tend to cluster in the left arm. In certain studies the left arm of the Alu sequence has been reported to be transcribed in uitro more efficiently than the right arm by RNA polymerase I11 (33). Most of the deletion breakpoints occur between the A and B sequences of the left arm. It seems likely that these sequences are unwound or bent at some point during transcription by RNA polymerase I11 and its associated transcription factors, and it is possible that this unwinding creates some predilection for recombination.

The data with the LDL receptor and y,6,(l-globin genes indicate that recombinations involving at least one Alu sequence are frequent causes of deletions in the human genome, but it is by no means universal. Several large deletions in the a- and y,d,(l-globin genes have been reported that do not involve Alu sequences a t either endpoint (38). In these cases there is no homology, either direct or inverted, between the sequences at the deletion joint. Thus, in these cases, other

mechanisms of deletion must have occurred. I t is interesting to note that the FH 274 and the FH 781

mutations both involve recombinations between Alu sequences in intron 15 and exon 18, although different Alu sequences are involved. A third deletion (FH 381) also involves an Alu sequence in intron 15 (Fig. 11). These findings raise the possibility that there is some intrinsic instability in the Alu sequences in intron 15. I t will be important to determine whether this region has been involved in other deletions in additional FH individuals. Further studies of the large number of mutations at the LDL receptor locus should help to clarify this point.

Acknowledgments-We thank Lavon Sanders and Edith Womack for able assistance in growing fibroblast cultures. Mark Woelfle, Gloria Brunschede, and Daphne Davis provided excellent technical assistance.

1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14. 15.

16.

17.

18. 19. 20.

21.

22.

23. 24.

25.

26. 27.

28. 29.

30.

31.

32.

33.

34. 35.

36.

37.

38.

REFERENCES Goldstein, J. L., Brown, M. S., Anderson, R. G. W., Russell, D. W., and

Yamamoto, T., Bishop, R. W., Brown, M. S., Goldstein, J. L., and Russell, Schneider, W. J. (1985) Annu. Reu. Cell Biol. 1, 1-39

Lehrman, M. A,, Schneider, W. J., Brown, M. S., Davis, C. G., Elhammer, D. W. (1986) Science 2 3 2 , 1230-1237

A,, Russell, D. W., and Goldstein, J. L. (1987) J. Biol. Chem. 2 6 2 , 401- 410

Hobbs, H. H., Brown, M. S., Goldstein, J. L., and Russell, D. W. (1986) J. Biol. Chem. 261,13114-13120

Lehrman, M. A., Schneider, W. J., Sudhof, T. C., Brown, M. S., Goldstein, J. L., and Russell, D. W. (1985) Science 227 , 140-146

Lehrman, M. A,, Goldstein, J. L., Brown, M. S., Russell, D. W., and Schneider, W. J. (1985) Cell 4 1 , 735-743

Davis, C. G., Lehrman, M. A., Russell, D. W., Anderson, R. G. W., Brown, M. S., and Goldstein, J. L. (1986) Cell 4 5 , 15-24

Yamamoto, T., Davis, C. G., Brown, M. S., Schneider, W. J., Casey, M. L., Goldstein, J. L., and Russell, D. W. (1984) Cell 3 9 , 27-38

Sudhof. T. C.. Goldstein. J. L.. Brown. M. S.. and Russell. D. W. (1985) Scieme 228,815-822

Natl. Acad. Sci. U. S. A. 78,5151-5155

. . , ,

Miyake, Y., Tajima, S., Yamamura, T., and Yamamoto, A. (1981) Proc.

Takaichi, S., Tajima, S., Miyake, Y., Yamamoto, A. (1985) Arteriosclerosis Fi 92R-9A2

Wyman, A. R., Wolfe, L. B., and Botstein, D. (1985) Proc. Natl. Acad. Sci.

Church, G. M., and Gilbert, W. (1984) Proc. Natl. Acad. Sci. U. S. A. 8 1 ,

-, "_ - - - U. S. A . 8 2 , 2880-2884

1991-1995 Loenen, W. A. M., and Blattner, F. R. (1983) Gene 2 6 , 171-179 Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982) Molecular Cloning:

-11- "1-

Laboratory Manual, pp. 1-545, Cold Spring Harbor Laboratory Press.

Sanger, F., Nicklen, S., and Coulson, A. R. (1977) Proc. Natl. Acad. Sci. U. Cold Spring Harbor, NY

Hobbs, H. H., Lehrman, M. A,, Yamamoto, T., and Russell, D. W. (1985) S. A. 74,5463-5467

Dombroski, D. F., and Morgan, A. R. (1985) J. Biol. Chem. 2 6 0 , 415-417 Proc. Natl. Acad. Sci. U. S. A. 8 2 , 7651-7655

Maxam, A. M., and Gilbert, W. (1980) Methods Enzymol. 6 5 , 499-560 Goldstein, J. L., Basu, S. K., and Brown, M. S. (1983) Methods Enzymol.

Davis, C. G., Elhammer, A., Russell, D. W., Schneider, W. J., Kornfeld, S., 9 8 , 241-260

Mosley, S. T., Goldstein, J. L., Brown, M. S., Falck, J. R., and Anderson, Brown, M. S., and Goldstein, J. L. (1986) J. Biol. Chem. 261,2828-2838

Thomas, P. S. (1983) Methods Enzymol. 100 , 255-266 R. G. W. (1981) Proc. Natl. Acad. Sci. U. S. A. 78,5717-5721

Russell, D. W., Yamamoto, T., Schneider, W. J., Slaughter, C. J., Brown, M. S., and Goldstein, J. L. (1983) Proc. Natl. Acad. Sci. U. S. A. 80, 7501 -750.5

Reynolds, G. A., Goldstein, J. L., and Brown, M. S. (1985) J. Biol. Chem.

Messing, J. (1983) Methods Enzymol. 101 , 20-78 Simonsen, C. C., and Levinson, A. D. (1983) Proc. Natl. Acad. Sci. U. S. A .

. - . - . I - -

260 , 10369-10377

80. 2495-2499 Deleted in proof. Kingsley, D. M., and Krieger, M. (1984) Proc. Natl. Acad. Sci. U. S. A. 8 1 ,

5A5A-5A.58 Davis, C. G., Lehrman, M. A., Russell D. W., Anderson, R. G. W., Brown,

Deininger, P. L., Jolly, D. J., Rubin, C. M., Friedmann, T., and Schmid, C.

Lehrman, M. A., Russell, D. W., Goldstein, J. L., and Brown, M. S. (1986)

Paolella, G., Lucero, M. A., Murphy, M. H., and Baralle F. E. (1983) EMBO

Ottolengbi, S., and Giglioni, B. (1982) Nature 3 0 0 , 770-771 Vanin, E. F., Henthorn, P. S., Kioussis, D., Grosveld, F., and Smithies, 0.

l _ l _ _ _ " M. S., and Goldstein, J. L. (1986) Ckll45, 15-24

W. (1981) J. Mol. Biol. 151 , 17-33

Proc. Natl. Acad. Sei. U. S. A . 83,3679-3683

J. 2,691-696

(1983) Cell 35.701-709 Jagadeeswaran,-P, Tuan, D., Forget, B. G., and Weissman, S. M. (1982)

Henthorn, P. S., Ma er, D L , Huisman, T. H. J., and Smithies, 0. (1986)

Orkin, S. H., and Kazazian, H. H., Jr. (1984) Annu. Reu. Genet. 18, 131-

Nature 296,469-470

Proc. Natl. Acad. Jci. U.'S.'A. 83, 5194-5198

171

Documents

Alu-Alu Recombination Deletes Splice Acceptor Sites and Produces