J. Biochem. 114, 15-20 (1993)

Cloning of a Human Acid Sphingomyelinase cDNA with a New Mutation That Renders the Enzyme Inactive1

Hiroyuki Ida,*,***,2 Owen M. Rennert,* Yoshikatsu Eto,*** and Wai-Yee Chan*"**

Departments of *Pediatrics, **Biochemistry and Molecular Biology, and Anatomy and Cell Biology, Georgetown University Medical Center, Washington, DC, U.S.A.; and ***Department of Pediatrics, Tokyo Jikei University School of Medicine, Minato-ku, Tokyo 105

Received for publication, November 13, 1992

A cDNA encoding human acid sphingomyelinase was initially obtained by screening a placental cDNA library in .i gtll with a synthetic oligonucleotide probe and subsequently with partial cDNA. The full-length cDNA, hPSM55, comprised 2,376 nucleotides, with a 5•L untranslated sequence of 122 nucleotides, an open reading frame of 1,884 nucleotides encoding a protein of 627 amino acids, and a 3•L untranslated region of 370 bases. hPSM55 was almost identical to pASM-IFL reported by Schuchman et al. (J. Biol. Chem. 266,8531 8539, 1991) except for a 6 base pair deletion in the signal peptide, which indicated the possible removal of valine and leucine residues between positions 36 and 37, and a 463T- to C-transition, which indicated a possible substitution of16Sarginine for cystine. This cDNA was expressed in both COS-7 cells and Chinese hamster ovary cells. There was no increase in acid sphingomyelinase activity in either cell line following transfection. However, the correction of a single base change, 463C to T, in hPSM55 caused increased acid sphin gomyelinase activity in transfectants. These results suggest that the mutation of nucleotide 463C to T plays an important role in the catalytic activity of acid sphingomyelinase.

Sphingomyelinase (sphingomyelin phosphodiesterase) [EC dation, ataxia, extrapyramidal signs, or psychiatric symp 3.1.4.12] is a hydrolase that catalyzes the cleavage of toms, patients with NPD type B exhibit marked phenotypic sphingomyelin to form phosphorylcholine and ceramide. At heterogeneity, particularly with respect to the nature and least two isozymes have been identified: neutral (NSM) and extent of the nervous system involvement (12). Biochemi acidic sphingomyelinase (ASM). ASM is located in lyso cal analyses cannot reliably distinguish the two phenotypes. somes, whereas NSM is membrane-bound. The two types Moreover, it is presently not possible to correlate the of sphingomyelinase isolated from human brain have been disease severity with the level of residual ASM activity. reported to share some physical and catalytic properties, To elucidate the heterogeneity of sphingomyelinase, and although certain differences were noted. Mg2+significantly the relationship between the phenotype and genotype of enhances the catalytic activity of NSM; the optimal pHs of NPD, cloning of cDNA encoding human ASM has been ASM and NSM are 4.5 and 7.0, respectively (1). Attempts attempted. Two types of partial cDNAs encoding human have been made to purify human ASM from brain (2, 3), ASM were isolated from a fibroblast cDNA library using synthetic oligonucleotide probes deduced from purified placenta (4-6), urine (7, 8), and seminal plasma (9). Human NSM has been purified from urine (10). Different human urinary ASM (13). The longest type 1, pASM1, and type 2, pASM2, were 1,879 and 1,382 base pairs (bp), purification procedures have led to the isolation of proteins with different molecular weights and specific activities (8, respectively. These cDNAs had identical sequences, with 11). Thus, although the heterogeneity of sphingomyelinase the exception of an internal region: type 1 cDNAs had 172 has clearly been documented, the cause of this phenomenon in-frame bp, which were replaced in type 2 cDNAs by a remains unknown. 40-bp in-frame sequence. Recently, the full-length type 1 ASM activity is deficient in patients with Niemann-Pick cDNA was isolated by Schuchman et al. (14). The 2,347-bp disease (NPD) type A and type B. NPD type A is character full-length type 1 cDNA comprised an 87-bp 5•L untranslat ized by massive hepatosplenomegaly, nervous system ed region (UTR), an 1,890-bp open reading frame (ORF) involvement in early infancy, and rapid progression with encoding a protein of 629 amino acids, and a 370-bp early death, usually prior to 5 years of age. Most patients 3•L-UTR. Kruth and Stoffel purified ASM from human with NPD type B have no neurological symptoms. Even if placenta and the peptide was cleaved with cyanogen bro they have neurological involvement, such as mental retar mide or proteolytically with trypsin, proteinase V8 or Lys C, and the resulting fragments were subjected to gas phase

1 This work was supported in part by a grant awarded to H. I. by the sequencing (15). They reported that the amino acid se Uehara Memorial Foundation and in part by an NIH grant, HD 21793, quence derived from recombinant human fibroblast cDNA awarded to W.Y.C. The nucleotide sequence data were submitted to clones reported by Schuchman contained none of the amino the EMBL Data Bank under the accession number x59960. acid sequence obtained from human placenta ASM. 2 To whom correspondence should be addressed at: Department of This communication describes the isolation of a full Pediatrics, Tokyo Jikei University School of Medicine, 3-25-8 length cDNA encoding human placenta ASM on the screen Nishi-Shinbashi, Minato-ku, Tokyo 105.

Vol. 114, No. 1, 1993 15 16 H . Ida et al.

ing a human cDNA placenta library with an oligonucleotide extension with an oligolabeling kit. probe, and its expression in both COS-7 cells and Chinese Site-Directed Mutagenesis-To evaluate the effect of the hamster ovary (CHO) cells. The relationship between the mutation of nucleotide 463C to T on the ASM activity, nucleotide sequence and the catalytic active site of ASM is transient expression was performed in COS-7 cells. Site discussed. directed mutagenesis was carried out by the method of Kunkel et al. (23) using a Mutan-K kit (Takara Shuzo, MATERIALS AND METHODS Kyoto). Briefly, hPSM55 was subeloned into the EcoRI site of M13mp18 in the 3•L-5•L orientation. Double-stranded General Materials-Restriction endonuclease, nucleic cDNA was transformed in BW313 (dut-, ung-) Escherichia acid modifying enzymes, and other chemicals were coli to obtain a uracil-containing single-stranded DNA. purchased from Bethesda Research Laboratory (BRL) Following isolation of the uracil-containing single-stranded (Gaithersburg, MD). A ƒÉ gtll human placenta cDNA DNA, a mutagenic primer (5•L-TCTGAGGCCTGTGGCCT library was obtained from Clontech Laboratory (Palo Alto, GC-3•L, identical with the nucleotide sequence from 460 to CA). Nylon filters, membranes, and radioisotopes were 478 of pASM-1FL) was annealed, and the second-strand purchased from Amersham (Arlington Heights, IL). Cul phage DNA was introduced into BMH71-18 mutS (ung+) E. ture medium and related supplements, such as fetal bovine coli to obtain thymine-containing DNA, and subsequently serum (FBS), non-essential amino acids, and G418, were single-stranded phage DNA was transfected into MV1184. obtained from GIBCO (Grand Island, NY). Ten plaques were picked at random and sequenced by the cDNA Library Screening and Isolation of the Full dideoxynucleotide chain termination method to confirm the - Length Human ASM cDNA-Initially, 1 x 106 independent correction of the single base change at 463C to T. recombinants derived from a human placenta cDNA library Expression of Cloned cDNA-For transient expression in X gtll were screened with a radio-labeled oligonu in COS-7 cells (ATCC), the full-length hPSM55 and cleotide probe (5•L-TGGCATCAGGTGGCGGCACAG-3•L) hPSM55 corrected by mutagenesis were subcloned in both which corresponded to the reverse complementary of bases orientations into the XbaI and KpnI sites of the eukaryotic 1434-1455 of the published partial cDNA sequence, expression vector, pcDE, which was kindly provided by Dr. pASM 1(13). 5•L-end labeling of the oligonucleotides with T4 J.A. Barranger (Pittsburg University, PA). DNA from polynucleotide kinase and [ƒÁ-32P] ATP (>5,000 Ci/mmol) sense and antisense constructs was transfected into COS-7 was performed by the standard method (16). The washing cells at 80-85% confluency in T-25 cm2 flasks using the conditions were 2 x SCC for 15 min twice, 0.1 x SCC and lipofectin reagent (BRL) according to the manufacturer's 0.1% SDS for 30 min, and 0.1 x SCC for 10 min at 60•Ž. instructions (24). The transfected cells were harvested Three hybridizing clones were isolated. The results indicat after 72 h, and then their ASM activities were determined. ed that these clones contained the partial cDNA of ASM. For stable expression in CHO-Kl cells, the full-length One cloned cDNA, hSM14, contained the sequence corre hPSM55 insert was subcloned in both orientations into the sponding to bases 443-1,870 of the published sequence of HindIII and XbaI sites of the eukaryotic expression vector, pASM1; it was used to rescreen the placental library. The pRC/CMV (Invitrogene, San Diego, CA). CHO-Kl cells cDNA probe was labeled by random primer extension using (105 viable cells/transfection) in T-75 cm2 flasks were oligolabeling kits (Pharmacia LKB, Piscataway, NJ) and transfected with 20ƒÊg of the linearized recombinant vector [ƒ¿-32P]dCTP (>300 Ci/mmol) (17). using 60ƒÊg of lipofectin in 5 ml of Opti MEM I medium Phage DNA preparation was carried out using DE52 resin supplemented with non-essential amino acids. After 24 h, (Bio-Rad, Richmond, CA) (18). The cDNA inserts of these 10 ml of F-12 medium supplemented with 10% FBS was clones were subcloned into pBluescript KSII (Strategene, added. After an additional 24 h, the medium was removed La Jolla, CA), and their nucleotide sequences were deter and 15 ml of F-12 supplemented with 10% FBS containing mined by the dideoxy sequencing method (19) using a T, 1 mg/ml G418 was added. Selection with G418 was con sequencing kit (Pharmacia LKB). tinued for 2 weeks (25). Cultures were maintained with The cDNA inserts, which were longer than 1.5 kilobases F-12 supplemented with 10% FBS and 0.5 mg/ml G418. (kb), were subcloned into pBluescript KSII and M13mp18 Three weeks after transfection, the enzyme activities were (Pharmacia LKB), and then sequenced with the T, sequenc determined and total RNA was extracted for analysis. . ing kit. Primers for sequencing ar.A screening were synthe Determination of ASM Activity-ASM activities were sized with an Applied Biosystems model 380B DNA syn determined using [methyl-14C]sphingomyelin (57 mCi/ thesizer using phosphoramidite chemistry (20). Computer mmol) as a substrate (26). Briefly, the substance and analyses were performed using the University of Wiscon Triton-X solutions were dried together, and then resus sin Genetics Computer Group (UWGCG) DNA sequence pended in acetate buffer (pH 4.5) by sonication, and the analysis software. sonicated cell suspensions were added to this mixture . Northern Blot Analyses-Total RNA was extracted from After incubation at 37•Ž for 1 h, the reactions were ter a T-75 cm2 flask of confluent CHO-K1 cells [American minated by the addition of chloroform-methanol (2: 1, v/ Type Culture Collection (ATCC), Rockville, MD] by the v). The upper phase was dried, and then the radioactivity acid guanidium thiocyanate method (21). Thirty micro was counted with a liquid scintillation counter. Protein grams of total RNA was separated by electrophoresis concentration determination was carried out by the Lowry through formaldehyde gel and transferred to a nylon method (27). The values were expressed as nmol/h/mg membrane by the capillary blot method in 10 x SCC. protein. Prehybridization and hybridization were carried out in 5•~ SSPE at 42•Ž (22). The probe used was the full-length cDNA, hPSM55, which was radiolabeled by random primer

J. Biochem. Cloning of Human Acid Sphingomyelinase 17

that the cDNA insert size ranged from 0.7 to 2.4 kb. Six

RESULTS cDNA inserts with size of longer than 1.5 kb were sub cloned into pBluescript KS II and M13mp18, and then Isolation and Characterization of Full-Length Human sequenced. All 6 clones contained type 1 cDNA. The clone ASM cDNA-Screening of 1•~106 independent recom that had the longest cDNA insert was hPSM55. The binants from a human placental cDNA library using a hPSM55 cDNA comprised 2,376-bp, a 5•L-UTR of 122-bp, partial cDNA resulted in the isolation of 58 putative human an ORF of 1,881-bp encoding a protein of 627 amino acids, ASM cDNA clones. Agarose gel electrophoresis revealed a stop codon, TAG, and a 370-bp 3•L-UTR (Fig. 1). The

details. The bold letters represent the changed nucleotides or the Fig. 1. Nucleotide and deduced amino acid sequences of the substituted amino acids. The underlined sequences represent the full-length ASM cDNA, hPSM55. This represents the composite tryptic peptides derived from purified human urinary ASM. The sequence of clone hPSM55, which spans base 122 to 2251, and clone underlined amino acid sequences with * represent potential N-glyco hPSM15, which spans base 774 to 2285. The 3•L end of hPSM55 cDNA sylation sites. The double underlined amino acid residues are putative and the 5•L end of hPSM15 are marked. Both clones were sequenced in signal peptides. The wavy line represents the polyadenylation signal. both orientations. #1 to #5 indicate differences between hPSM55 and pASM-1FL reported by Schuchman et al. See "DISCUSSION" for

Vol. 114, No, 1, 1993 18 H. Ida et al.

TABLE I. Transient expression of hPSM55 cDNA in COS-7 cells. Values are expressed as nmol/h/mg protein, and represent the average of two independent experiments.

TABLE II. Stable expression of hPSM55 cDNA in CHO cells. Values are expressed as nmol/h/mg protein, and represent the average of two independent experiments. Fig. 2. Northern blot analysis of total RNA isolated from CHO cells. Only CHO cells transfected with sense hPSM55 gave an intense band at 2.5 kb (lane 4). CHO cells (lane 1), CHO cells transfected with the pRC/CMV vector (lane 2), and CHO cells transfected with antisense hPSM55 (lane 3) gave no hybridized band against radio labeled full-length hPSM55 cDNA.

amino acid sequences of all 12 tryptic peptides derived cDNA. from urinary human ASM were present in the polypeptide Northern blot analysis of CHO cells was carried out (Fig. encoded by hPSM55. The 5•L-UTR was extremely G/C rich, 2). CHO cells, CHO cells transfected with the unmodified with 88 G/C out of the 122-bp. There were two 10-bp pRC/CMV vector and CHO cells transfected with the (CCCCGGCGCC) repeats in tandem starting from base antisense constructs showed no hybridized band. On the -81 to -61 , in which 8 of 10 of the nucleotides were other hand, the transfectant with sense hPSM55 gave a shared. There were two in-frame potential translation very intense 2.5 kb band. initiation codons, ATG at bases 1 and 97. Neither of these codons was flanked by sequences similar to the consensus DISCUSSION sequence proposed by Kozak (28). Examination of the amino acid sequence by means of UWGCG showed the In the present study, we isolated the full-length cDNA, characteristic structure of a signal peptide with a hydro hPSM55, encoding human ASM by screening a human phobic core rich in leucine and alanine residue. There were placenta cDNA library initially using an oligonucleotide and five leucine-alanine repeats starting at amino acid residue subsequently a partial cDNA as the probe. hPSM55 had the 35. The corresponding nucleotide sequence, CTGGCG, was characteristic features of a full-length cDNA, such as an also repeated 5 times in tandem. The encoded protein initial codon, a signal peptide, a stop codon, and a polyade contained 6 potential N-glycosylation sites of the form, nylation signal, and contained 12 tryptic peptides derived Asn-X-Thr/Ser. No poly(A) tail could be found in hPSM55. from human urinary ASM. However, it exhibited no However, a putative polyadenylation signal, AATAAA, catalytic activity in either transfected COS-7 cells or CHO was identified at base 2246 of the cDNA. In hPSM55 cDNA, cells. Northern blot analyses revealed the existence of a base 2251 was followed by three Cs. These nucleotides very intense 2.5 kb band, suggesting that the efficacy of were believed to be derived from the linker which was used transfectin was enough to express transcripts from in the generation of the library. An overlapping clone, hPSM55. hPSM15, contained a cDNA insert spanning base 774. It In order to determine why hPSM55 transfectants had no was identical to hPSM55 until base 2251. In hPSM15, the catalytic activity, we compared the nucleotide sequence of last three Cs of hPSM55 were replaced with 19 nucleotides hPSM55 with that of pASM-1FL, which had catalytic and a stretch of 15 As. activity (14). pASM-1FL had six additional nucleotides Expression of hPSM55 in Both COS-7 Cells and CHO (GTGCTG) encoding 2 amino acids, valine and leucine, Cells-The full-length cDNA, hPSM55, was inserted into between nucleotides 105 and 106 (Fig. 1-#1) in the signal the transient expression vector, pcDE, and then transfected peptide region of hPSM55. There were also four single base into COS-7 cells. As shown in Table I, the mean endogenous differences between hPSM55 and pASM-1FL (Fig. 1-#2 ASM activity in COS cells was 1.4 nmol/h/mg protein. The #5). A 463T to C-transition, resulting in a 155cystine (Cys: ASM activity in COS cells transfected with either the TGT) to arginine (Arg:CGT) substitution, a 959T- to antisense or sense construct was 1.1 nmol/h/mg protein. C-transition, resulting in a 320isoleucine (Ile:ATC) to threo The experiments demonstrated that hPSM55 transcripts nine (Thr:ACA) substitution, a 1510A- to G-transition, did not express the catalytically active enzyme in COS resulting in a 504arginine (Arg:AGG) to glycine (Gly:GGG) cells. substitution, and a 1812Gto A-transition, which is a silent hPSM55 was also inserted into the stable expression mutation that does not affect an amino acid, were recog vector, pRC/CMV, and transfected into CHO-K1 cells. As nized. The changes at nucleotides 959 and 1510 could be shown in Table II, the mean endogenous ASM activity in due to individual polymorphism since these two sites are CHO cells was 75 nmol/h/mg protein. The ASM activities known to be highly polymorphic (14). Therefore, hPSM55 in CHO cells transfected with the antisense and sense was mainly different from pASM-1FL with respect to a constructs were 87 and 84 nmol/h/mg protein, respective 6-bp deletion (Fig. 1-#1) and a single base change at ly. There were no significant differences between CHO nucleotide 463 of hPSM55 (Fig. 1-#2) (Fig. 3, a and b) . cells, and transfectants of the antisense and the sense Either the 6-bp deletion in the signal peptide or the single

J. Biochem. Cloning of Human Acid Sphingomyelinase 19

Fig. 3. Comparison of se quences of hPSM55 and pASM-1FL. Panel (a) repre sents a 6-bp deletion (GTGCTG) between nucleotides 105 and 106, resulting in the deletion of valine and leucine residues (Fig. 1-#1). Panel (b) demonstrates a 463T- to C-transition that pre dicts a cystine to arginine substi tution at codon 155 (Fig. 2-#2).

To date, three genetic mutations have been found in NPD patients, whose ASM enzyme activity was deficient. A G to T-transition of nucleotide 1487 and a G to A-transition of nucleotide 1729 in pASM-1FL have been determined in type A patients (31, 32). A three base pairs deletion, nucleotides 1821 to 1823 has been found in type B patients base change at 463 of hPSM55 could affect the ASM (33). That those mutations were responsible for the loss of enzyme activity. Since the 6-bp deletion in hPSM55 is ASM activity was confirmed by the expression of mutated located in the signal peptide it should not be responsible for cDNA. Schuchman et al. (29) noted that these mutations the enzyme activity, although an abnormality of a signal were in exon 6 and speculated that this region might encode peptide might affect the reaction between mRNA and a all or part of the ASM catalytic site. However, the mutation ribosome. Recently, the complete nucleotide sequence of described in this communication was in exon 2, and has not the gene encoding human ASM was determined (29). When been reported before, suggesting that this region also plays the nucleotide sequences of pASM-1FL and genomic se an important role in the ASM enzyme activity. The charac quences were compared, a 6-bp deletion (GTGCTG) near terization of these mutated enzyme proteins will be impor the 5•L end of the ASM genomic sequence, which was exactly tant for elucidation the properties of ASM, however, a the same as the deletion in hPSM55, was recognized. reliable antibody against ASM has not been prepared. Moreover, when the 5•L end of ASM cDNA, with the Through analysis of the relationship between mutations polymerase chain reaction using both a sense primer (5•L- and the catalytic activity, more detailed biochemical and AGTAGTGAATTCACGGGACAGACGAACCA-3•L), corre physical properties of ASM may be clarified. sponding to hPSM55 nucleotides -39 to -23 with an additional 12 bases that included an EcoRI site, and an REFERENCES antisense primer (5•L-AGTAGTCTGCAGAGGGAAGCTA TTGACAGG-3•L), corresponding to hPSM55 nucleotides 1. Rao, B.G. & Spence, M.W. (1976) J. Lipid Res. 17, 506-515 961 to 978 with an additional 12 bases containing a PstI 2. Gatt, S., Dinur, T., & Kopolovic, J. (1978) J. Neurochem. 31, 547-551 site, was amplified and sequenced, a 6-bp deletion was also 3. Yamanaka, T. & Suzuki, K. (1982) J. Neurochem. 38,1753-1754 found (data not shown). Therefore, it is most likely that the 4. Pentchev, P.G., Brady, R.O., Gal, A.E., & Hibbert, S.R. (1977) C to T-transition at nucleotide 463 is responsible for the Biochim. Biophys. Acta 488, 312-321 loss of ASM activity in hPSM55. To confirm this specula 5. Jones, C.S., Shankaran, P., & Callahan, J.W. (1981) Biochem. J. tion, we corrected nucleotide 463C- to T using the site 195,373-382 directed mutagenesis technique, and expressed the correct 6. Sakuragawa, N. (1982) J. Biochem. 92, 637-646 7. Weitz, G., Driessen, M., Brouwer-Kelder, E.M., Sandohoff, K., ed hPSM55 in COS-7 cells. The ASM activity of trans Barranger, J.A., Tager, J.M., & Schram, A.W. (1985) Biochim. fectants was 6.0 nmol/h/mg protein (about 5-fold the mean Biophys. Acta 838, 92-97 endogenous activity of COS-7 cells). These data suggest 8. Quintern, L.E., Weitz, G., Nebrkarn, H., Tager, J.M., Schram, that the single base change of nucleotide 4fi3C to T in A.W., & Sandohoff, K. (1987) Biochim. Biophys. Acta 922, 323 hPSM55 was responsible for the loss of catalytic activity. -336 The cause of this missense mutation could be a cloning 9. Vanha-Perttula, T. (1988) FEBS Lett. 233, 263-267 10. Chatterjee, S. & Ghosh, N. (1989) J. Biol. Chem. 264, 12554 artifact due to the aberrant synthesis of cDNA by reverse -12561 transcriptase during the ƒÉ gtl1 library construction or the 11. Jobb, E.A. & Callahan, J.W. (1987) J. Inhert. Metab. Dis. 10, occurrence of different species of mRNAs. However, the 326-328 significance of the present results is that a C to T-transi 12. Spence, M.W. & Callahan, J.W. (1989) in The Metabolic Basis of tion renders the human ASM inactive. The same phenome Inherited Diseases (Scriver C.R., Beaudet A.L., Sly, W.S., & non was reported in the case of cloning of human glucocere Valle D., eds.) Vol. 2, pp. 1655-1676, McGraw-Hill, New York brosidase by Tsuji et al. (30). 13. Quintern, L.E., Schuchman, E.H., Levran, O., Suchi, M.,

Vol. 114, No. 1, 1993 20 H. Ida et al.

Ferlinz, K., Reinke, H., Sandohoff, K., & Desnick, R.J. (1989) Enzymol. 154, 367-382 EMBO J. 8, 2469-2473 24. Felgner, P.L., Gadek, T.R., Holm, M., Roman, R., Chan, H.W., 14. Schuchman, E.H., Suchi, M., Takahashi, T., Sandohoff, K., & Wenz, M., Northrop, J.P., Ringald, G.M., & Danielsen, M. (1987) Desnick, R.J. (1991) J. Biol. Chem. 266, 8531-8539 Proc. Natl. Acad. Sci. USA 84, 7413-7417 15. Kruth, J. & Stoffel, W. (1991) Biol. Chem. Hoppe-Seyler 372, 25. Hefta, L.J.F., Schrewe, H., Thompson, J.A., Oikawa, S., 215-223 Nakazato, H., & Shively, J.E. (1990) Cancer Res. 50, 2397-2403 16. Maniatis, T., Fritsh, E.F., & Sambrook, J. (1989) Molecular 26. Suzuki, K. (1978) Methods Enzymol. 47, 465-468 Cloning: Laboratory Manual 2nd ed., Cold Spring Harbor Lab., 27. Lowry, O.H., Rosebrough, N.J., Farr, A.L., & Randall, R.J. Cold Spring Harbor, New York (1951) J. Biol. Chem. 193, 265-275 17. Feiberg, A.P. & Vogelstein, P. (1984) Anal. Biochem. 137, 266 28. Kozak, M. (1986) Cell 44, 283-292 -267 29. Schuchman, E.H., Levran, 0., Pereira, L.V., & Desnick, R.J. 18. White, B. & Rosenzweig, S. (1989) BioFeedback 7, 694-695 (1992) Genomics 12, 197-205 19. Sanger, F., Nicklen, J., & Coulson, A.R. (1977) Proc. Natl. Acad. 30. Tsuji, S., Choudary, P.V., Martin, B.M., Winfield, S., Barranger, Sci. USA 74, 5463-5467 J.A., & Ginns, E.I. (1986) J. Biol. Chem. 261, 50-53 20. Itakura, K., Rossi, J.J., & Wallance, R.B. (1984) Annu. Rev. 31. Levran, 0., Desnick, R.J., &Schuchman, E.H. (1991) Proc. Natl. Biochem. 53, 323-356 Acad. Sci. USA 88, 3748-3752 21. Chomczynski, P. & Sacchi, N. (1987) Anal. Biochem. 162,156 32. Ferlinz, K., Hurwitz, R., & Sandohoff, K. (1991) Biochem. -159 Biophys. Res. Commun. 179,1187-1191 22. Fourney, R.M., Miyakoshi, J., Day, R.S., III, & Peterson, M.C. 33. Levran, 0., Desnick, R.J., & Schuchman, E.H. (1991) J. Clin. (1988) Focus 10, 5-7 Invest. 88, 806-810 23. Kunkel, T.A., Roberts, J.D., & Zakour, R.A. (1987) Methods

J. Biochem.