Cloning of Cdna Sequences of Human Adenosine Deaminase (Adenosine Deaminase Mutants/Leukemia Lymphoblasts/Differential Screening/A Phage) DAN A

Proc. Natl. Acad. Sci. USA Vol. 80, pp. 7481-7485, December 1983 Biochemistry

Cloning of cDNA sequences of human adenosine deaminase (adenosine deaminase mutants/leukemia lymphoblasts/differential screening/A phage) DAN A. WIGINTON*, GWENDOLYN S. ADRIAN*, RICHARD L. FRIEDMANt, D. PARKER SUTTLE*, AND JOHN J. HUTTON*: *Department of Medicine, University of Texas Health Science Center, and Audie L. Murphy Memorial Veterans Hospital, San Antonio, TX 78284; and tDepartment of Biochemistry, Stanford University School of Medicine, Stanford, CA 94305 Communicated by Eloise R. Giblett, August 22, 1983 ABSTRACT Cloned cDNA sequences of human adenosine de- amined by translation in vitro (7). Included was the interesting aminase (ADA; adenosine aminohydrolase, EC 3.5.4.4) have been finding that two cell lines with mutations in ADA contained isolated from a cDNA library constructed in bacteriophage AgtlO. higher levels of translatable ADA mRNA than corresponding The cDNA for the library was prepared from poly(A)+ RNA iso- normal lines. In the present report, we describe isolation of lated from a human T-lymphoblast cell line, CCRF-CEM. The cloned cDNA sequences for human ADA. The molecular size library was initially screened by differential plaque hybridization and quantity of ADA mRNA in a number of normal and mutant to labeled cDNA prepared from human T- and B-lymphoblast cell human lymphoblast cell lines have been determined directly by lines with a 21-fold difference in levels of translatable ADA mRNA. to cDNA Two recombinants containing cloned cDNA sequences for ADA hybridization the ADA-specific probes. were identified by hybridization-selected translation. Both recombinants contained approximately 1,600 base pairs of inserted METHODS human DNA. Restriction maps of the two inserts were not iden- RNA Isolation and Translation. Procedures for extraction tical. One contained approximately 40 base pairs of additional DNA and isolation of total cellular RNA, purification of poly(A)+ RNA, toward the center of the cDNA. The cloned cDNA specifically translation of RNA in a reticulocyte lysate system, and specific hybridized to five fragments generated by HindIll digestion of immunoprecipitation and quantitation of ADA have been de- human genomic DNA. It also hybridized to human lymphoblast scribed previously (7). RNA species 1.6 and 5.8 kilobases in length. The cDNA was used as a probe to estimate ADA mRNA levels in human lymphoblast Cell Lines. The lymphoblast cell lines GM-130, GM-131, cell lines. ADA mRNA levels correlate closely with levels of ADA GM-2184, GM-3043, and GM-2606 are transformed B-cell lines catalytic activity and ADA protein in cell lines containing struc- obtainedfrom the Human Genetic MutantCellRepository (Cam- turally normal ADA. A leukemic T-lymphoblast line produced 6 den, NJ). GM-130, GM-131, and GM-2184 were established to 9 times as much ADA protein and ADA mRNA as transformed from individuals with normal ADA activity, whereas GM-3043 B-lymphoblast lines. Two mutant B-lymphoblast lines from pa- and GM-2606 are from ADA-deficient individuals (5). The CCRF- tients with hereditary ADA deficiency contained unstable ADA CEM cells were obtained from the American Type Culture protein but had 3 to 4 times the normal level of ADA mRNA. Collection. The HPB-ALL cells were a gift from Jun Mino- wada, Roswell Park Memorial Institute (Buffalo, NY). Both the Adenosine deaminase (ADA; adenosine aminohydrolase, EC CCRF-CEM and HPB-ALL are human leukemia T-lympho- 3.5.4.4), a part of the purine catabolic pathway, catalyzes the blast lines and have structurally normal ADA. Cells were grown irreversible deamination of adenosine and deoxyadenosine. in RPMI 1640 medium supplemented with 10-20% fetal calf Deficiency of adenosine deaminase activity in humans is as- serum. sociated with an autosomal recessive form of severe combined Construction of a Human T-Lymphoblast cDNA Library. immunodeficiency disease (1). The underlying specific bio- The cDNA cloning vector, Agt1O, was kindly provided by Thanh chemical and metabolic abnormalities involved and their effects Huynh and R. Davis (Stanford University). After addition of on the immune system have been the subject of intensive in- EcoRI linkers, cDNA was inserted into the EcoRI site of Agtl0 vestigation (2, 3). and recombinants were selected on Escherichia coli BNN 102 A number of different mutations have occurred that can cause as described by Young and Davis (9). Technical details of the human ADA deficiency (4-8). Both the clinical and biochem- preparation of cDNA libraries in AgtlO will be presented else- ical effects of the mutations are strikingly different among in- where. The phage library then was grown on E. coli C600. dividuals. Some mutations are associated with an unstable ADA Screening of the cDNA Library. The cDNA library was protein, total deficiency of ADA activity in erythrocytes, partial screened by a variation of the Benton and Davis in situ plaque deficiency of ADA activity in peripheral lymphocytes, and ab- hybridization method (10, 11). Recombinant bacteriophage were sence of clinical disease (5). Other mutations have resulted in plated on 9 X 9 cm square plates and grown to produce plaques the virtual absence of ADA activity in erythrocytes and lym- approximately 1 mm in diameter. Duplicate replicas from each phocytes and cause severe combined immunodeficiency dis- plate were prepared on nitrocellulose filters, and these were ease (2-7). Within the second group, there are a number of dis- screened by differential hybridization to labeled cDNA pre- tinct mutations, as indicated by the amount of mutant ADA pared from mRNA preparations from lines HPB-ALL and GM- protein detectable by radioimmunoassay in extracts of cell lines 2184. These mRNA preparations contained 0.051% and 0.0025% carrying those mutations (4, 6). In a recent report, ADA mRNA ADA-specific sequences, respectively, as determined by in vi- in a number of these human lymphoblast cell lines was ex- tro translation and immunoprecipitation (7). The mRNA used to prepare cDNA for screening was first The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertise- Abbreviation: ADA, adenosine deaminase. ment" in accordance with 18 U.S.C. §1734 solely to indicate this fact. t To whom reprint requests should be addressed. 7481 Downloaded by guest on September 29, 2021 7482 Biochemistry: Wiginton et al. Proc. Natl. Acad. Sci. USA 80 (1983) partially purified by agarose/urea gel electrophoresis (12). Gel cDNAs were isolated from the A recombinants and recloned in slices were frozen, homogenized, and extracted with 0.04 M pBR325. Inserts were then isolated from the plasmids and di- Tris acetate buffer, pH 7.4. The RNA samples were extracted gested with restriction enzymes. Fragments were separated on with phenol and chloroform, precipitated, and assayed. The 5% polyacrylamide gels (14). To distinguish fragments at ends relative ADA mRNA contents of the HPB-ALL and GM-2184 from fragments located centrally, approximately 8,000 cpm of preparations were increased approximately 6-fold to 0.33% and insert DNA that had been 32P-labeled at both ends was in- 0.016%, respectively. Aliquots (0.5 Ag) of these partially pu- cluded in each digest. rified mRNAs were used for preparation of 32P-labeled cDNA Transfer of Denatured RNA to Nitrocellulose Paper and (8 x 108 cpm/,ug) by a variation of the method described by Hybridization with cDNA. Poly(A)+ RNA samples were glyoxy- Maniatis et al. (11), with oligo(dT) as primer for avian myelo- lated, electrophoresed in a 1.1% agarose gel, and transferred blastosis virus reverse transcriptase (J. Beard, Life Sciences) at to nitrocellulose paper (15). The filters were probed with nick- 100-150 units per Ag of RNA. translated (16) phage DNA (2 x 108 cpm/Ag). Each of the two labeled cDNA preparations (5 X 105 cpm Dot-Blot Hybridization. Relative amounts of ADA mRNA in per filter) was hybridized to one of the sets of replica filters poly(A)+ RNA preparations from lymphoblast cell lines were described earlier. The hybridization was carried out as de- estimated by hybridization of 32P-labeled cDNA to RNA spot- scribed by Maniatis et al. (11), with agitation at 420C for 72 hr. ted on nitrocellulose paper (15, 17). A 0.25-Ag sample of DNA The filters were autoradiographed for 18 hr at -70TC. Plaques from each of the two recombinant bacteriophages that con- producing a hybridization signal significantly stronger with the tained ADA sequences (2-11 and 3-3) was labeled by nick trans- HPB-ALL cDNA than with the GM-2184 cDNA were picked lation (16). A molar excess of the nick-translated recombinant with sterile toothpicks and replated. The resulting plaque ar- phage DNA (108 cpm/,ug) was hybridized with the bound RNA rays were transferred to duplicate nitrocellulose filters and re- at 420C for 20 hr (15). The filters were washed and autoradio- screened with the differential probes. graphed for 18 hr at -700C. Relative amounts of ADA mRNA The group of plaques that demonstrated a "strong" differ- were estimated by measuring the optical absorbance of auto- ential signal was screened further. Labeled cDNA was pre- radiographic spots, using a Dynatech micro ELISA minireader pared from mRNA derived from T-lymphoblast lines HPB-ALL MR590. and CCRF-CEM and B-lymphoblast lines GM-2606 and GM- Southern Blot Analysis. DNA (15 jig) was digested to com- 2184. The mRNA preparations from the paired T-cell lines and pletion with HindIII, separated in an 0.8% agarose gel, trans- paired B-cell lines differed 4-fold in ADA mRNA by in vitro ferred to a nitrocellulose filter, and hybridized to 3 P-labeled translation (HPB-ALL greater than CCRF-CEM; GM-2606 probe ADA 3-3 (11). greater than GM-2184). 32P-Labeled cDNAs prepared from these mRNAs were used as paired probes in rescreening the plaques RESULTS AND DISCUSSION by differential hybridization. In this manner plaques with a high Isolation of Cloned cDNA Sequences for Human ADA. We probability of containing ADA inserts were identified for fur- chose to construct the CCRF-CEM cDNA library in a phage A ther examination by hybridization-selected translation. vector because A vectors offer both increased cloning efficiency Isolation of A DNA. DNA from the AgtlO recombinants was and simplified screening (10) in comparison to conventional isolated by a modification of the procedure described by Man- plasmid vector techniques. The AgtlO constructed by Huynh iatis et al. (11). After bacterial lysis, removal of bacterial debris, and Davis was chosen because it offers increased packaging ef- precipitation with polyethylene glycol, and extraction with ficiency over other A vectors that will accept small inserts. Ad- chloroform, the A particles were isolated by centrifugation at ditionally, AgtlO permits a selection for recombinants based on 8,000 x g for 18 hr at 40C. Phage were then treated with pro- the hfl A150 mutants of Hoyt et al. (18) because insertion of teinase K in 0.5% sodium dodecyl sulfate and extracted with cDNA destroys cI repressor function. phenol and phenol/chloroform. Normal yield was 100-200 ,g Rapid screening of large numbers of recombinants was es- of DNA per 40-ml culture. sential because of the very low abundance of ADA mRNA. A Hybridization-Selected Translation. Recombinant phage that library containing approximately 900,000 recombinant bacte- contained ADA sequences were identified directly by hybrid- riophage was prepared by using cDNA to mRNA from CCRF- ization-selected translation. The filters were prepared as de- CEM lymphoblasts. The library was screened on the basis of scribed by Parnes et al. (13) and hybridized to RNA as described differential in situ hybridization to 32P-labeled cDNA prepared by Maniatis et al. (11). Because the ratio of vector to insert DNA from mRNAs with a 21-fold difference in translatable ADA is much greater for A than for plasmids, more DNA from a sin- mRNA. Phage containing ADA inserts was expected to give a gle vector was loaded onto each filter than in the cited refer- significantly greater hybridization signal with the T-cell cDNA ence. A 40-pg sample of DNA from a single recombinant phage (high ADA) than with the B-cell cDNA (low ADA). Because the was bound to a 7-mm nitrocellulose disc. Then, 50-300 Ag of recombinant cDNA library was prepared from mRNA from a HPB-ALL mRNA was hybridized with 20-25 discs in a total T-lymphoblast line (CCRF-CEM), many recombinant phage volume of 0.5 ml at 48°C for 4-5 hr with vigorous shaking. The were expected to contain cDNA sequences more prevalent in discs were washed and the bound RNA was eluted from each the T-cell probe than in the B-cell probe. Consequently, 1,072 disc in an individual tube. The RNA was precipitated along with plaques from an initial screening of 77,000 were selected for 30,ug of calf liver tRNA in 0.3 M sodium acetate and 70% (vol/ further examination. These phage were rescreened by differ- vol) ethanol. The precipitated RNA was dried, dissolved in 5 ential hybridization (Fig. 1). A group of 496 plaques with the Al of sterile water, and translated in vitro. The products were strongest differential hybridization signal was chosen for fur- examined for ADA by specific immunoprecipitation. ther study. In order to eliminate some plaques identified on the Restriction Endonuclease Map. The size of cDNA inserts basis of other inherent differences in T lymphoblasts and B was determined by cleavage of recombinant DNA with EcoRI lymphoblasts, these 496 recombinants were screened a third restriction endonuclease (2.5 units/pug of DNA, 37TC, 150 min) time with paired B-cell cDNA probes and a fourth time with and electrophoresis on both 1.2% and 2.0% agarose gels. Stan- paired T-cell cDNA probes. The paired mRNAs from which dard marker DNA fragments were as follows: Hae III-cut 4X174, these cDNA probes were prepared had a 3- to 4-fold difference HindIII-cut A, BstNI-cut pBR325, and Taq I-cut pBR325. The in translatable ADA mRNA. On the basis of the third and fourth Downloaded by guest on September 29, 2021 Biochemistry: Wiginton et al. Proc. Natl. Acad. Sci. USA 80 (1983) 7483 9 0 * 0 t * * *0

l111ii Is :io,.12 * 0. * 4k 0 *

1 lb

40.* ON& .S | 0 V. *p :.4 _Awm NW so~.*w _ /~~. 6 40 0 b /

}i;

7 a

42a 2b

FIG. 1. Hybridization of 32P-labeled cDNA to recombinant bacteriophage AgtlO DNA. DNA from recombinant Agt1O plaques was transferred to sets of duplicate nitrocellulose filters (a and b). The filter sets were hybridized to [32P]cDNA prepared from partially purified mRNA from B- lymphoblast line GM-2184 (a) and from T-lymphoblast line HPB-ALL (b). The mRNAs in a and b were low and high in ADA-specific mRNA, respectively. Plaques giving a significantly greater signal on the b filters than on the a filters were selected for further study. The arrows indicate recombinant bacteriophage 2-11 (1 a and b) and 3-3 (2 a and 6).

screens, 68 recombinants were chosen for examination by hy- sequences. Typical examples of these are shown in Fig. 2, lanes bridization-selected translation. a and c. Two recombinant bacteriophage contained cDNA se- Of the 68 recombinant bacteriophage examined by hybrid- quences that hybridized with ADA mRNA. The mRNA could ization-selected translation, 66 were negative for ADA cDNA then be eluted and translated (Fig. 2, lanes b, d, and j). The two recombinant strains of bacteriophage were identified as 2- 11 and 3-3. Because of the high plaque density in the initial plating, the possibility of plaque contamination seemed high. Therefore, both 2-11 and 3-3 were replated at low density (10-20 plaques per 9 x 9 cm plate) and 10 clones from each strain were isolated. All 20 of these clones contained cDNA with ADA sequences by hybridization-selected translation (Fig. 2, lanes e-i). In addition, electrophoresis of total translation products (without immunoprecipitation) from mRNAs selected by hybridization to DNA from clones 2-11 and 3-3 did not reveal any other selected proteins except ADA. Therefore, the cloned cDNAs do not appear to contain more than one insert that hybridizes with lymphoblast mRNA. Restriction Map. The size of the inserts in each of the two recombinant bacteriophage 2-11 and 3-3 was estimated by agar- Li c)C d t? f h j k 1 ose gel electrophoresis after cleavage with EcoRI. Both inserts contain between 1,525 and 1,640 base pairs. The map of re- FIG. 2. In vitro ADA synthesis from mRNA selected by hybridiza- striction sites is shown in Fig. 3. Inserts ADA 2-11 and ADA tion to recombinant bacteriophage DNA. A 40-,tg sample ofDNA from 3-3 appear identical except that an Hpa II fragment from ADA each AgtlG recombinant was bound to a 7-mm-diameter nitrocellulose 3-3 contains both an additional 40 base pairs of DNA and a novel disc, and 20-25 discs were hybridized with either 150 jig (lanes a-d) or 300 (lanes e-j) of HPB-ALL mRNA. Bound mRNA was eluted and jig HA B H P H H TH T P translated in vitro, and either one-half (a-d) or all (e-j) of the trans- ADA 211 JL'I lation products were immunoprecipitated with ADA-specific anti- A A A serum. Lanes a-j contain immunoprecipitated translation products from TH TP mRNA that hybridized to DNA from recombinant bacteriophage 4-14 ADA 33HTB8 H P H H (a), 3-3 (b, j), 2-17 (c), 2-11 (d), 2-11-8 (e), 2-11-9 (f), 2-11-10 (g), 3-3-9 AA A A (h), and 3-3-10 (i). The 2-11-8, 2-11-9, 2-11-10, 3-3-9, and 3-3-10 clones were isolated from 2-11 and 3-3 by replating atlow density. Lanes k and 100 bp 1 contain translation products precipitated with nonimmune serum (k) or anti-ADA serum (1) from translation of total HPB-ALL mRNA. The FIG. 3. Restriction maps of ADA cDNA inserts. Site locations are arrow indicates the migration of purified human ADA standard. Elec- ± 10 base pairs (bp) except the H site in the small A fragment has not trophoresis of total translation products (without immunoprecipita- been unambiguously located and lies 15-25 base pairs in one direction tion) from mRNA selected by hybridization to DNA from clones 2-11 or other of the center of the fragment. HindlIl, Pva A, Sma I, Xba I, and 3-3 did not reveal any hybridization-selected products other than and Xho I did not cut the insert. A, Ava i; B, BamHI; H, Hpa H; P, Pst ADA. I; T, Taq I. Downloaded by guest on September 29, 2021 7484 Biochemistry: Wiginton et al. Proc. Natl. Acad. Sci. USA 80 (1983)

Ava II site not present in ADA 2-11. The location of the Hpa molecular weight of 44,000 (19). A protein this size would re- II fragment containing the extra DNA in ADA 3-3 was inferred quire a coding sequence of approximately 1,100 bases, which from the location of the novel Ava II site on that fragment be- could easily be accommodated in a 1,600-base segment. cause it is almost certain that the novel site will be in the extra The quantitatively minor 5.8-kilobase RNA might be a pre- DNA. There are several possible explanations for the additional cursor of the mature 1.6-kilobase message. However, this mi- DNA toward the center of ADA 3-3. ADA 3-3 and 2-11 could nor species could also represent: a second processed transcript represent different ADA alleles present in the original CCRF- for ADA; another transcript related only by possession of a re- CEM lymphoblast line from which the cDNA library was de- peat sequence; an aberrantly processed ADA transcript; or a rived. Alternatively, (i) ADA 3-3 could be derived from a poly(A)+ second transcript that hybridizes to a non-ADA moiety inserted RNA not yet fully processed but containing a small intron or (ii) into the recombinant A in addition to ADA. It is not possible a short nucleotide sequence could have been deleted from ADA to choose definitively among these without additional infor- 2-11 or inserted into ADA 3-3 during the cloning procedure. mation, although no ADA-specific translatable mRNA of this Resolution of this issue will require determining the sequences size has been eluted from agarose/urea gels and no translation of the two cDNAs and possibly the protein. products other than ADA have been seen when RNA hybrid- Southern Blot Analysis of Genomic DNA. DNAs from cell izing to the ADA probes has been translated. lines CCRF-CEM, GM-2184, and GM-2606 were digested with Estimation of ADA mRNA Levels in Human Lymphoblast HindIII for Southern blot analysis. The three lines had quali- mRNA by Dot-Blot Hybridization. Levels of ADA-specific RNA tatively identical patterns. Two fragments (3.7 and 2.8 kilo- in total poly(A)+ RNA from several different human lympho- bases) hybridized strongly and three fragments (15.8, 8.5, and blast lines were estimated by dot-blot hybridization. This is a 1.0 kilobases) hybridized weakly to the ADA cDNA probe. The sensitive method for comparing levels of specific sequences, cDNA itself was not cleaved by HindIII, so it is likely that the even though absolute quantitative amounts cannot be deter- ADA gene contains introns with HindIII sites or, alternatively, mined. Relative levels of ADA mRNA were estimated by com- there may be ADA pseudogenes. paring autoradiographic spot intensities as determined by op- Blot Analysis of ADA Message in Human Lymphoblast Cell tical absorbance. Comparable results were obtained with probes Lines. RNA blots are shown in Fig. 4. Each of the cell lines prepared from bacteriophages 2-11 and 3-3. The results are contained a species of RNA approximately 1,600 bases in length shown in the last column of Table 1. that hybridized with the probe. This corresponds to the size of ADA mRNA and Protein in Normal, Leukemia, and Mu- the ADA mRNA estimated by in vitro translation of RNA frac- tant Lymphoblast Lines. When normal human peripheral lym- tions from agarose/urea gels. The fact that the lengths of the phocyte subpopulations have been examined, T cells have been cDNA inserts are similar in size to human ADA mRNA may found to contain significantly higher levels of ADA activity (4- mean that the inserts contain most or all of the sequence of to 5-fold greater) than B cells (20, 21). Differences in ADA have mature human ADA mRNA. Human lymphoblast ADA has a also been noted between human leukemia T-lymphoblast lines and transformed normal B-cell lines. ADA catalytic activity, ADA immunoreactive protein, and translatable ADA mRNA were found to be 6 to 8 times higher in the T-lymphoblast lines than in the B-lymphoblast lines (ref. 7 and Table 1). The difference

Table 1. Relative amounts of ADA catalytic activity, ADA jinmunoreactive protein, translatable ADA mRNA, and hybridizable ADA mRNA in normal and ADA- - 5.8 deficient human lymphoblast cell lines Relative ADA Immuno- mRNA mRNA Catalytic reactive content by content by Cell line activity protein translation hybridization Normal B line GM-2184 1.2 1.0 1.0 1.4 ..... 1.6 GM-130 0.9 ND 0.8 0.7 GM-131 0.9 ND 1.3 1.0 ADA-deficient B line GM-2606 0.01 0.12 4.1 2.7 a b c d e Partially ADA- deficient B line FIG. 4. Hybridization of recombinant A DNA containing ADA- GM-3043 0.38-1.9 0.45-1.07 3.6 4.7 specific cDNA inserts to poly(A)+ RNA from human lymphoblast lines. Leukemia.T line Glyoxylated poly(A)+ RNA was fractionated on a 1.1% agarose gel. Lanes CCRF-CEM 8.2 7.1 6.1 9.7 a-c contained 10 ,ug of RNA and lanes d and e contained 20 ,ug of RNA. Poly(A)+ RNA was isolated from: GM-130, a normal B line (a); CCRF- Relative.values were calculated in each case by setting the average CEM, a leukemia T line (b); GM-3043, a mutant B line (c); GM-131, a value of the normal B lines as 1.0. Catalytic activities were measured ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~...a T line The fractionated Immunoreactive ADA was measured ra- normal B line (d); and HPB-ALL, leukemia (e). as described (19). protein by RNA was transferred to nitrocellulose and probed with 0.5 zg of 32p_ dioimmunoassay with either goat or rabbit antiserum to pure human labeled DNA from recombinant bacteriophage 3-3. Hybridization was ADA (4). For a discussion of the properties of these sera when used to for20 hr at42°C. Blots were washed and autoradiographed 4 hr at -700C. assay mutant proteins, see ref. 4. Relative translatable mRNA content Approximate lengths are indicated in kilobase units. Human 18S and is based on translation assays with 20 jtg of poly(A)+ RNA per ml (7). 28S rRNA, HindlII-cut A DNA, and Hae rn-cut 4X174 DNA served as Estimation ofADA-specific hybridizable mRNA levels is described un- size standards. der Dot-blot hybridization. ND, not determined. Downloaded by guest on September 29, 2021 Biochemistry: Wiginton et al. Proc. Natl. Acad. Sci. USA 80 (1983) 7485 in ADA activity and protein between T lymphoblasts and B synthetase, aspartate transcarbamoylase, and dihydroorotase lymphoblasts had been noted previously and ascribed, on the (CAD) (26) and dihydrofolate reductase (27) have shown a close basis of in vivo pulse labeling, in part to differences in rate of correlation between levels of enzyme protein and specific mRNA synthesis but principally to differences in the rate of degra- in a variety of cell types. The question of the mechanism of the dation of the ADA proteins (22). The relative ADA mRNA level observed increase in ADA mRNA arises. It might be caused by in the T-lymphoblast CCRF-CEM line compared to the normal more efficient transcription of DNA, by differences in pro- B-lymphoblast lines, as determined directly by hybridization to cessing and degradation of mRNA, by regulatory responses to the labeled cDNA probe, agrees closely with our previous ob- differences in concentration of metabolites of the purine path- servations based on in vitro translation (Table 1 and ref. 7). The ways, or perhaps even by changes in the number of copies of relative amounts of ADA mRNA might be sufficient to explain the ADA gene. Southern blot analyses of DNAs from three cell the observed difference between T and B lymphoblasts in lev- lines with different lines of ADA mRNAs were qualitatively els of ADA catalytic activity and protein. identical. The patterns were sufficiently similar quantitatively An unexpected observation from our previous study was that to rule out more than a 2- to 3-fold variation in gene copy num- the relative amounts of translatable ADA mRNA in two mutant ber. Therefore, the increase in ADA mRNA could not be en- ADA-deficient B-cell lines (GM-2606 and GM-3043) were 3- to tirely accounted for by gene amplification. 4-fold greater than the amount in the normal B-cell lines (7). Mr. John Blodgett provided valuable assistance by growing numer- This was especially surprising because both of these cell lines ous cell lines. Dr. George R. Stark and Dr. John J. Kanalas made many produce altered ADA proteins, and determination of translat- helpful suggestions. Dr. Stark also provided access to critically impor- able mRNA levels depends upon immunoreactivity of the ADA tant reagents and facilities. This work was supported by Research Grant product after in vitro translation. Cell line GM-2606 is derived HD-15036 from the National Institutes of Health and by the Veterans from an ADA-deficient, immunodeficient child and the cells Administration Research Service. have an extremely low amount of ADA catalytic activity (about 1. Giblett, E. R., Anderson, J. E., Cohen, F., Pollara, B. & Meu- 1% of normal) and approximately 10% of the normal amounts wissen, H. J. (1972) Lancet ii, 1067-1069. of ADA protein, as detected by radioimmunoassay (Table 1). 2. Martin, D. W., Jr., & Gelfand, E. W. (1981) Annu. Rev. Biochem. This altered protein in the mutant is more thermolabile than 50, 845-877. the normal ADA protein (unpublished observations). The level 3. Thompson, L. F. & Seegmiller, J. E. (1980) in Advances in En- zymology, ed. Meister, A. (Wiley, New York), pp. 167-210. of ADA mRNA in cell line GM-2606, as determined by in vitro 4. Wiginton, D. A. & Hutton, J. J. (1982) J. Biol. Chem. 257, 3211- translation and now by direct hybridization, is 3- to 4-fold higher 3217. than in the normal B-cell lines (Table 1). 5. Hirschhorn, R., Roegner, V., Jenkins, T., Seaman, C., Piomelli, Cell line GM-3043 is derived from an individual with only S. & Borkowsky, W. (1979)J. Clin. Invest. 64, 1130-1139. partial deficiency of ADA in peripheral lymphocytes and no 6. Daddona, P. E., Frohman, M. A. & Kelley, W. N. (1980)1. Biol. That the mutant cell line contains Chem. 255, 5681-5687. associated disease syndrome. 7. Adrian, G. S. & Hutton, J. J. (1983)J. Clin. Invest. 71, 1649-1660. an unstable ADA protein has been demonstrated by its in- 8. Hirschhorn, R., Martiniuk, F., Roegner-Maniscalco, V., Ellen- creased thermolability (5). As previously reported, the trans- bogen, A., Perignon, J. L. & Jenkins, T. (1983)J. Clin. Invest. 71, latable ADA mRNA level in this line was about 4 times normal 1887-1892. (7), and this finding is confirmed by our estimation of ADA 9. Young, R. A. & Davis, R. W. (1983) Proc. Natl. Acad. Sci. USA 80, mRNA by specific hybridization (Table 1). This cell line, like 1194-1198. mutation associated with 10. Benton, W. D. & Davis, R. W. (1977) Science 196, 180-182. the GM-2606 line, seems to contain a 11. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) Molecular both an unstable ADA protein and increased levels of ADA Cloning (Cold Spring Harbor Laboratory, Cold Spring Harbor, NY). mRNA. 12. Locker, J. (1979) Anal. Biochem. 98, 358-367. As noted earlier (7), these findings are not unprecedented. 13. Parnes, J. R., Velan, B., Felsenfeld, A., Ramanathan, L., Fer- A mutation in the human glucose-6-phosphate dehydrogenase rini, U., Apella, E. & Seidman, J. G. (1981) Proc. Natl. Acad. Sci. (G6PD) gene results in both a single amino acid substitution in USA 78, 2253-2257. of 14. Maniatis, T., Jeffrey, A. & van de Sande, H. (1975) Biochemistry the protein and a 4-fold increase in concentration and rate 14, 3787-3794. synthesis of the protein (23). Apparently, G6PD in the mutant 15. Thomas, P. S. (1980) Proc. Natl. Acad. Sci. USA 77, 5201-5205. was as stable as the normal protein. Whether the increased syn- 16. Rigby, P. W. J., Dieckmann, M., Rhodes, C. & Berg, P. (1977)J. thesis was due to increased amounts of G6PD mRNA or in- Mol. Biol. 113, 237-251. creased efficiency of translation was not determined. A mouse 17. Birnie, G. D., Bums, J. H., Wiedemann, L. M., Warnock, A. M., neuroblastoma revertant cell line has been reported that over- Tindle, R. W., Burnett, A. K., Tansey, P., Lucie, N. P. & Rob- ertson, M. R. (1983) Lancet i, 197-200. produces 25- to 50-fold the mRNA for a structurally defective, 18. Hoyt, M. A., Knight, D. M., Das, A., Miller, H. I. & Echols, H. thermolabile hypoxanthine phosphoribosyltransferase (HPRT) (1982) Cell 31, 565-573. (24). The increased mRNA levels were determined by in vitro 19. Wiginton, D. A., Coleman, M. S. & Hutton, J. J. (1981) Biochem. translation and the line was subsequently shown to contain J. 195, 389-397. multiple copies of the HPRT gene (25). Thus, one response a 20. MacDermott, R. P., Tritsch, G. L. & Formeister, J. F. (1980) Clin. cell can make to a defective protein is to increase the number Exp. Immunol. 42, 303-307. increased amounts of 21. Tung, R., Silber, R., Quagliata, F., Conklyn, M., Gottesman, J. of copies of the mutant gene and make & Hirschhorn, R. (1976) J. Clin. Invest 57, 756-761. the corresponding mRNA and protein. In this manner the ef- 22. Daddona, P. E. (1981)J. Biol. Chem. 256, 12496-12501. fects of instability of the altered protein are partially overcome 23. Yoshida, A. (1970) J. Mol. Biol. 52, 483-490. by making abnormally large amounts of it. Whether this is a 24. Melton, D. W., Konecki, D. S., Ledbetter, D. H., Hejtmancik, common mechanism of response to mutations causing protein J. F. & Caskey, C. T. (1981) Proc. Natl. Acad. Sci. USA 78, 6977- instability is not known. 6980. the 25. Brennand, J., Chinault, A. C., Konecki, D. S., Melton, D. W. & Variation in amounts of translatable ADA mRNA among Caskey, C. T. (1982) Proc. Natl. Acad. Sci. USA 79, 1950-1954. lymphoblast cell lines is caused by actual differences in quan- 26. Wahl, G. W., Padgett, R. A. & Stark, G. R. (1979)J. Biol. Chem. tities of mRNA,- rather than by different efficiencies of trans- 254 8679-8689. lation of the mRNA. Similarly, studies with the multifunctional 27. Alt, F. W., Kellems, R. E., Bertino, J. R. & Schimke, R. T. (1978) protein containing the enzyme activities carbamoyl-phosphate J. Biol. Chem. 253, 1357-1370. Downloaded by guest on September 29, 2021