A new method for isolation of S-adenosylmethionine (SAM)-accumulating yeast

S-Adenosylmethionine (SAM) is an important metabolite that participates in many reactions as a methyl group donor in all organisms, and has attracted much interest in clinical research because of its potential to improve many diseases, such as depression, liver disease, and osteoarthritis. Because of these potential applications, a more efficient means is needed to produce SAM. Accordingly, we developed a positive selection method to isolate SAM-accumulating yeast in this study. In Saccharomyces cerevisiae, one of the main reactions consuming SAM is thought to be the methylation reaction in the biosynthesis of ergosterol that is catalyzed by Erg6p. Mutants with deficiencies in ergosterol biosynthesis may accumulate SAM as a result of the reduction of SAM consumption in ergosterol biosynthesis. We have applied this method to isolate SAM-accumulating yeasts with nystatin, which has been used to select mutants with deficiencies in ergosterol biosynthesis. SAM-accumulating mutants from S. cerevisiae K-9 and X2180-1A were efficiently isolated through this method. These mutants accumulated 1.7–5.5 times more SAM than their parental strains. NMR and GC-MS analyses suggested that two mutants from K-9 have a mutation in the erg4 gene, and erg4 disruptants from laboratory strains also accumulated more SAM than their parental strains. These results indicate that mutants having mutations in the genes for enzymes that act downstream of Erg6p in ergosterol biosynthesis are effective in accumulating SAM.     

Biochemical studies on cell fusion. I. Lipid composition of fusion- resistant cells

A series of stable cell mutants of mouse fibroblasts were previously isolated (Roos, D. S. and R. L. Davidson, 1980, Somatic Cell Genet., 6:381-390) that exhibit varying degrees of resistance to the fusion-inducing effect of polyethylene glycol (PEG), but are morphologically similar to the parental cells from which they were derived. Biochemical analysis of these mutant cell lines has revealed differences in whole cell lipid composition which are directly correlated with their susceptibility to fusion. Fusion-resistant cells contain elevated levels of neutral lipids, particularly triglycerides and an unusual ether-linked lipid, O-alkyl, diacylglycerol. This ether lipid is increased approximately 35-fold over parental cells in the most highly PEG-resistant cell line. Fusion-resistant cells also contain more highly saturated fatty acyl chains (ratio of saturated to polyunsaturated fatty acids [S/P ratio] approximately 4:1) than the parental line (S/P ratio approximately 1:1). Cells which are intermediate in their resistance to PEG have ether lipid and fatty acid composition which is intermediate between the parental cells and the most fusion-resistant mutants. In a related communication (Roos, D. S. and P. W. Choppin, 1985, J. Cell. Biol., 100:1591-1598) evidence is presented that alteration of lipid content can predictably control the fusion response of these cells.     

Characterization of low potassium-resistant mutants of the Madin-Darby canine kidney cell line with defects in NaCl/KCl symport

Three independent mutants of the Madin-Darby canine kidney cell line (MDCK) have been isolated which were capable of growth in media containing low concentrations of potassium. All three mutants were deficient to varying extents in furosemide- and bumetanide-sensitive 22Na+, 86+b+, and 36Cl- uptake. The two mutants most resistant to low K+ media had lost essentially all of the 22Na+, 86Rb+, and 36Cl- uptake activities of this system. The third mutant was partially resistant to low K+ media and had reduced levels of bumetanide-sensitive uptake for all three ions. Extrapolated initial uptake rates for 22Na+, 86Rb+, and 36Cl- revealed that the partial mutant exhibited approximately 50% of the parental uptake rates for all three ions. The stoichiometries of bumetanide-sensitive uptake in both the parental cell line and the partial mutant approximated 1 Rb+:1 Na+:2 Cl-. The results of this study provide genetic evidence for a single tightly-coupled NaCl/KCl symporter in MDCK cells. The correlation between the ability to grow in low K+ media and decreased activity of the bumetanide-sensitive co-transport system suggests that the bumetanide-sensitive transport system catalyzes net K+ efflux from cells in low K+ media. The results of 86Rb+ efflux studies conducted on ouabain-pretreated mutant and parental cells are consistent with this interpretation. Cell volume measurements made on cells at different densities in media containing normal K+ concentrations showed that none of the mutants differed significantly in volume from the parental strain at a similar cell density. Furthermore, all three mutants were able to readjust their volume after suspension in hypotonic media. These results suggest that in the MDCK cell line, the bumetanide-sensitive NaCl/KCl symport system does not function in the regulation of cell volume under the conditions employed.     

The End2 mutation in CHO cells slows the exit of transferrin receptors from the recycling compartment but bulk membrane recycling is unaffected

We have characterized a new CHO cell line (12-4) derived from a parental line, TRVb-1, that expresses the human transferrin receptor. This mutant belongs to the end2 complementation group of endocytosis mutants. Like other end2 mutants, the endosomes in 12-4 cells show a partial acidification defect. These cells internalize LDL and transferrin at 70% of the rate of parental cells and externalize transferrin at 55% of the parental rate (Johnson, L. S., J. F. Presley, J. C. Park, and T. E. McGraw. J. Cell Physiol. 1993). In this report, we have used fluorescence microscopy to determine which step in receptor trafficking is affected in the mutants. Transferrin is sorted from LDL and is delivered to a peri-centriolar recycling compartment at rates similar to parental cells. However, the rate constant for exit of transferrin from the recycling compartment in mutant cells is 0.025 min- 1 vs 0.062 min-1 in the parental line. We also measured the trafficking of a bulk membrane marker, 6-[N-[7-nitrobenzo-2-oxa-1,3-diazol-4-yl]- amino]hexanoyl- sphingosylphosphorylcholine (C6-NBD-SM) that labels the exofacial side of the plasma membrane. C6-NBD-SM enters the same recycling compartment as transferrin, and it exits the recycling compartment at a rate of 0.060-0.065 min-1 in both parental and 12-4 cells. We conclude that bulk membrane flow in the recycling pathway of 12-4 cells is normal, but exit of transferrin from the recycling compartment is slowed due to retention in this compartment. Thus, in the mutant cell line the recycling compartment carries out a sorting function, retaining transferrin over bulk membrane.     

Mechanisms of cAMP-mediated gene induction: examination of renal epithelial cell mutants affected in the catalytic subunit of the cAMP-dependent protein kinase

The precise mechanistic role of the cAMP-dependent protein kinase (cAMP-PK) in cAMP-mediated gene induction remains unclear. Renal epithelial cell mutants were compared to the LLC-PK1 parental cell line for induction of the cAMP-responsive urokinase-type plasminogen activator (uPA) gene, as quantitated by the technique of mRNA solution hybridization. The FIB4 and FIB6 mutants, which possess less than 10% parental cAMP-PK catalytic (C) subunit activity, showed markedly diminished uPA mRNA induction in response to agents elevating intracellular cAMP such as the cAMP analogue 8-bromo-cAMP and the adenylate cyclase-stimulating hormones vasopressin and calcitonin. In contrast, the mutant cells responded to a similar or greater extent than the parental cells in terms of uPA mRNA induction following treatment with the Ca2+/phospholipid-dependent protein kinase activator phorbol 12-myristate 13-acetate (PMA). Elevation of intracellular cAMP was found to induce a translocation of the cAMP-PK C subunit from the perinuclear Golgi region to the nucleus in both parental and mutant cell lines, as shown by immunocytochemical techniques. Results argue for the role of the cAMP-PK C subunit activity and possibly nuclear translocation of the C subunit in cAMP-mediated uPA induction, which is mechanistically distinct from the PMA-stimulated response.     

Structure of amplified DNA in different Syrian hamster cell lines resistant to N-(phosphonacetyl)-L-aspartate.

Syrian hamster cell lines selected in multiple steps for resistance to high levels of N-(phosphonacetyl)-L-aspartate (PALA) contain many copies of the gene coding for the pyrimidine pathway enzyme CAD. Approximately 500 kilobases of additional DNA was coamplified with each copy of the CAD gene in several cell lines. To investigate its structure and organization, we cloned ca. 162 kilobases of coamplified DNA from cell line 165-28 and ca. 68 kilobases from cell line B5-4, using a screening method based solely on the greater abundance of amplified sequences in the resistant cells. Individual cloned fragments were then used to probe Southern transfers of genomic DNA from 12 different PALA-resistant mutants and the wild-type parents. A contiguous region of DNA ca. 44 kilobases long which included the CAD gene was amplified in all 12 mutants. However, the fragments cloned from 165-28 which were external to this region were not amplified in any other mutant, and the external fragments cloned from B5-4 were not amplified in two of the mutants. These results suggest that movement or major rearrangement of DNA may have accompanied some of the amplification events. We also found that different fragments were amplified to different degrees within a single mutant cell line. We conclude that the amplified DNA was not comprised of identical, tandemly arranged units. Its structure was much more complex and was different in different mutants. Several restriction fragments containing amplified sequences were found only in the DNA of the mutant cell line from which they were isolated and were not detected in DNA from wild-type cells or from any other mutant cells. These fragments contained novel joints created by rearrangement of the DNA during amplification. The cloned novel fragments hybridized only to normal fragments in every cell line examined, except for the line from which each novel fragment was isolated or the parental population for that line. This result argues that "hot spots" for forming novel joints are rare or nonexistent.     

Dependence of urokinase-type-plasminogen-activator induction on cyclic AMP-dependent protein kinase activation in LLC-PK1 cells.

The activation of cyclic AMP-dependent protein kinase (cAMP-PK) in vivo was studied in LLC-PK1 pig kidney cells and the mutant cell lines M18 and FIB5, which have total levels of cAMP-PK catalytic-subunit and regulatory-subunit activities comparable with those of parental cells. The extent of cAMP-PK activation (release of active catalytic subunit from the holoenzyme) was directly correlated with the cellular cyclic AMP concentration in LLC-PK1 cells. In LLC-PK1 cells, as well as in the mutants M18 and FIB5, the extent of the induction of urokinase-type plasminogen activator (uPA) by the cyclic AMP-mediated effectors calcitonin, vasopressin and forskolin was directly correlated with the levels of activated catalytic subunit. The 'receptorless' mutant M18, which is impaired in calcitonin- and vasopressin-receptor function, did not show any activation of cAMP-PK or uPA production in response to either hormone, whereas cAMP-PK and uPA responses to forskolin were about 35% higher than in parental cells. Analysis of the FIB5-cell line revealed a lesion affecting the regulation of adenylate cyclase activity, whereby basal and stimulated (both receptor- and non-receptor-mediated) adenylate cyclase levels were less than 36% of those in parental cells. The activation of cAMP-PK in response to cyclic AMP effectors was similarly reduced, and uPA induction was concomitantly lower than that in parental cells. The results demonstrate the dependence of uPA induction by cyclic AMP effectors on dissociation of the cAMP-PK holoenzyme, implying the importance of activated free cAMP-PK catalytic subunit in this process. Thus it is concluded that the mutations in the cellular cyclic AMP-generating apparatus of the M18 and FIB5 cell lines impair uPA induction by preventing cAMP-PK activation.     

Purification and characterization of adenosine deaminase from a genetically enriched mouse cell line

Mammalian adenosine deaminase has been shown by genetic and biochemical evidence to be essential for the development of the immune system. For the purpose of studying the function and structure of this enzyme, we have isolated by genetic selection a mouse cell line, B-1/50, in which adenosine deaminase levels were increased 4,300-fold over the parent cell line. The enzyme was purified from these cells in large quantity and high yield by a simple two-step purification scheme. The enzyme derived from the B-1/50 cells was indistinguishable from that of the parental cells as judged by several biochemical criteria. The Km (30 microM) and Ki (4 nM) values using adenosine as substrate and 2'-deoxycoformycin as inhibitor, respectively, were identical for the enzyme derived from the parental cells as well as the adenosine deaminase gene amplification mutants. The enzyme from both cell types exhibited multiple isoelectric focusing forms which co-purified using our purification protocol. Electrophoretic analysis using sodium dodecyl sulfate-polyacrylamide gels showed that adenosine deaminase migrated with an apparent molecular weight of 41,000 or 36,000 depending on whether the enzyme was reduced or oxidized, respectively. This shift was reversible, indicating that proteolysis was not responsible for the faster migrating form. Monospecific antibodies raised against purified adenosine deaminase cross-reacted with the enzyme derived from the parental cells and precipitated 37% of the total soluble protein in the B-1/50 cells. Continued genetic selection resulted in the isolation of cells in which adenosine deaminase was overproduced by 11,400-fold and accounted for over 75% of the soluble protein.     

Isolation and partial characterization of human cell mutants differing in sensitivity to killing and mutation by methylnitrosourea and N-methyl-N'-nitro-N-nitrosoguanidine

Isogenic variants resistant to alkylating agents have been isolated from the human lymphoblast cell line TK6. The cell lines may be divided into four classes on the basis of resistance to N-methyl-N'-nitro-N-nitrosoguanidine (MNNG). The sensitive TK6 parental line shows a 37% survival after 45-min exposure to 0.04 microM MNNG; the three classes of more resistant mutants show 37% survival after 45-min exposure to 2 microM (MF lines), 6 microM (MT lines), and greater than or equal to 10 microM (MX line) MNNG. A representative MF line, MF1, is resistant to both killing and mutation by MNNG or N-methyl-N-nitrosourea. An MT clone, MT1, is highly resistant to killing but hypermutable by MNNG. The MT1 line, like the parental TK6, does not remove O6-methylguanine adducts from the DNA. Our data are consistent with the hypothesis that the MT1 line possesses a nonexcision pathway of defense against killing by alkylating agents. Rather than preventing alkylation of DNA or removing alkylated adducts, the MT1 cells appear to be tolerant of the adducts that are not removed from the DNA.     

Low-level iron-dependent mutants of Listeria monocytogenes and their virulence in macrophages

Listeria monocytogenes is an opportunistic intracellular pathogen capable of growth that requires iron for growth within phagocytic cells and virulence expression. In the presence of an appropriate concentration tropolone, an iron-chelating agent, growth of L. monocytogenes is completely inhibited. However, this inhibition can be relieved by addition of dopamine, norepinephrine, or ferric citrate. By selection on streptonigrin medium supplemented with tropolone and norepinephrine, we have obtained two spontaneous mutants, Lm-8 and Lm-15, with the same iron dependence but lower iron dependence than the wild-type Lm-B38. The association between iron requirement and virulence of the two mutants and the wild type was studied in the J774 macrophage cell line. One hour after phagocytosis by the J774 macrophage cell line, the two mutants and the parental strain displayed no difference in the number of phagocytosed bacteria. Twenty-four hours after phagocytosis, the number of bacteria within the surviving macrophages was identical for the wild strain and the two clones. However, only 40% of macrophage cells infected with Lm-8 and 90% of those infected with Lm-15 were alive after 24 h in comparison with macrophage cells infected with the parental strain Lm-B38. These data demonstrate that there is no direct correlation between iron requirement and virulence of L. monocytogenes in the J774 macrophage cell line.     

Expression of recessive Aprt- mutations in mouse CAK cells resulting from chromosome loss and duplication

Karyotypes of recessive mutants at the autosomal adenine phosphoribosyltransferase (Aprt) locus in a clone of the near-diploid mouse CAK cell line have been analyzed. The Aprt located on chromosome 8. One copy of chromosome 8 was morphologically abnormal in the parental clone (CAK-B3-Toyr13) from which Aprt- mutants were isolated. Among 22 mutants, there were ten in which one copy of chromosome 8 had been lost. Four of these were monosomic, and in the others duplication of the remaining homolog had occurred. These findings indicate that newly induced recessive mutations in cultured mammalian cells can be expressed as the result of loss of one chromosome carrying a wild-type allele with or without duplication of the homolog carrying the mutant allele. Loss and duplication would not be detected in cell lines lacking morphologically marked chromosomes.     

Genetic mapping and physiological consequences of metE mutations of Bacillus subtilis.

Three metE mutations of Bacillus subtilis, which cause cells to have a 25- to 200-fold decrease in L-methionine S-adenosyltransferase (EC 2.5.1.6) activity, were mapped between bioB and thr. The corresponding three metE mutants contained three- to fourfold less intracellular S-adenosylmethionine (SAM) but at least sevenfold more methionine than the metE+ strain when grown in synthetic medium. This indicates a strong feedback control of SAM on its synthesis. However, only the metE2 strain, with the lowest SAM concentration, grew at a slightly lower rate than the parent, which showed that an intracellular concentration of about 25 microM SAM was critical for growth at the normal rate. Neither DNA methylation (measured by bacteriophage luminal diameter 105 restriction) nor sporulation was affected at this low SAM concentration. Addition of methionine to the growth medium caused an increase in the pool of SAM in some but not all metE mutants. Coaddition of adenine did not change this result. However, the extent of sporulation (induced by mycophenolic acid) was decreased 50-fold in all mutants by the addition of methionine and adenine. Therefore, the combination of methionine and adenine suppresses sporulation regardless of whether it causes an increase in the level of SAM.     

A human hybrid myeloma for production of human monoclonal antibodies

We produced somatic cell hybrids between human myeloma cells and a lymphoblastoid cell line that is hypoxanthine phosphoribosyl transferase-deficient and ouabain-resistant. These hybrids were phenotypically similar to the human myeloma parental cells and grew as well as the human lymphoblastoid parental cells. After counterselection in 6-thioguanine, mutants that were 6-thioguanine-and ouabain-resistant were obtained, one of which was used as a fusion partner with lymphoblastoid B cells that produce anti-tetanus toxoid (TT) antibodies. These hybrids secreted human anti-TT monoclonal antibodies in much larger amounts than the parental lymphoblastoid cells, and were stable for a period of over 10 mo until the present time. Thus, by hybridizing plasmacytomas with lymphoblastoid cells, we constructed a fusion partner that secretes large amounts of immunoglobulin (Ig), grows at a fast rate, has a high fusion frequency, and supports the production of monoclonal antibodies over long periods of time. Moreover, anti-TT antibody-producing hybrids have been grown as solid tumors in irradiated BALB/c nude mice and then adopted to ascites growth, producing 1 to 8 mg of human immunoglobulin per 1 ml of ascites fluid.     

A novel Arabidopsis gene TONSOKU is required for proper cell arrangement in root and shoot apical meristems

Root apical meristem (RAM) and shoot apical meristem (SAM) are vital for the correct development of the plant. The direction, frequency, and timing of cell division must be tightly controlled in meristems. Here, we isolated new Arabidopsis mutants with shorter roots and fasciated stems. In the tonsoku (tsk) mutant, disorganized RAM and SAM formation resulted from the frequent loss of proper alignment of the cell division plane. Irregular cell division also occurred in the tsk embryo, and the size of cells in meristems and embryo in tsk mutant was larger than in the wild type. In the enlarged SAM of the tsk mutant, multiple centers of cells expressing WUSCHEL (WUS) were observed. In addition, expression of SCARECROW (SCR) in the quiescent center (QC) disappeared in the disorganized RAM of tsk mutant. These results suggest that disorganized cell arrangements in the tsk mutants result in disturbed positional information required for the determination of cell identity. The TSK gene was found to encode a protein with 1311 amino acids that possesses two types of protein-protein interaction motif, leucine-glycine-asparagine (LGN) repeats and leucine-rich repeats (LRRs). LGN repeats are present in animal proteins involved in asymmetric cell division, suggesting the possible involvement of TSK in cytokinesis. On the other hand, the localization of the TSK-GFP (green fluorescent protein) fusion protein in nuclei of tobacco BY-2 cells and phenotypic similarity of tsk mutants to other fasciated mutants suggest that the tsk mutation may cause disorganized cell arrangements through defects in genome maintenance.     

mRNA from mutant Y1 adrenal cells directs the synthesis of altered regulatory subunits of type 1 cAMP-dependent protein kinase

The molecular basis for altered cyclic AMP-dependent protein kinase activity was examined in three different mutant clones (Kin-1, Kin-7, and Kin-8) derived from the Y1 mouse adrenocortical cell line. Parental Y1 cells and the Kin mutants were labeled with L-[35S] methionine and the regulatory subunit of the type 1 cAMP-dependent protein kinase isozyme (RI) was immunoprecipitated from each clone with a specific guinea pig antiserum. When analyzed by electrophoresis on isoelectric focusing gels, the immunoprecipitates from mutant clones exhibited parental forms of RI plus an additional acidic variant form which likely accounted for altered cAMP-dependent protein kinase activity. Poly(A+) RNA was isolated from Y1 and Kin mutant cells and was translated in a cell-free, reticulocyte lysate system in the presence of L-[35S]methionine. The RI synthesized from poly(A+) RNA was immunoprecipitated from the translation mixture and analyzed on isoelectric focusing gels. The poly(A+) RNA from the Kin mutant clones directed the synthesis of parental and acidic variant forms of RI. These results suggest that the altered electrophoretic forms of RI arise from mutations in one of two RI genes rather than from post-translational modifications of the protein. The coexistence of parental and variant forms of RI in the Kin mutants indicate that the mutations are codominant.     

Morphological characteristics and ganglioside composition of substratum adhesion sites in a hepatoma cell line (CMH5123 cells) during different phases of growth

This study compares the ganglioside composition of tissue culture substrate-attached material (SAM) with that of cell bodies in a line of transformed hepatocytes derived from the minimal deviation Morris hepatoma 5123 c (CMH5123 cells). We examined both confluent cultures (late-phase cultures) and cells which were allowed to attach for only 3 h (early-phase cultures). We also determined to what extent ganglioside compositions of SAM and cell bodies from early- and late-phase cultures of CMH5123 cells are affected by the block of complex ganglioside biosynthesis induced by treatment with chelating agents (EGTA + EDTA). The morphological characteristics of SAM were monitored by scanning electron microscopy during the different steps of this study. In early-phase cultures, SAM was composed of fragments of filopodia and small vesicles probably representing newly formed substratum adhesion sites. In contrast, SAM of late-phase cultures was made up of large pools of membranous material resulting from the breakage of thick retraction fibers connecting the cell body with broad, mature adhesion sites. SAM of early-phase cultures yielded ganglioside profiles with a higher content of GM1 and GD1 a than those of cell bodies, while in late-phase cultures there was no difference between SAM and cell body gangliosides. When cells were grown in the presence of chelating agents, SAM of early-phase cultures was composed of vesicles and filopodial fragments similar to those found in early-phase cultures grown in regular media; these morphological features also appeared in SAM of confluent cultures (in contrast to the membranous material characteristic of late-phase cultures grown in regular media). In early-phase cultures grown in the presence of chelating agents, gangliosides of SAM were enriched in complex homologs relative to their content in cell bodies. These ganglioside characteristics were also found in SAM of confluent cultures grown in the presence of chelating agents, reflecting the presence of newly formed adhesion sites. On the basis of these results, we may conclude that the molecular assembly of newly formed adhesion sites implies the preferential distribution of several surface components involved in cell adhesion, including complex gangliosides.     

Analyses of mutation in pSV2gpt-transformed CHO cells

We have developed a system to study mutations which affect expression of the E. coli xanthine-guanine phosphoribosyl transferase (XPRT) gene (gpt) in hypoxanthine-guanine phosphoribosyl transferase-deficient (HPRT-) Chinese hamster ovary (CHO) cells that have been transformed by the plasmid pSV2gpt. Several gpt-transformed cell lines have been isolated and characterized with respect to integrated pSV2gpt sequences, expression of the gpt gene, and cytotoxic and mutagenic responses to UV light. While the gpt-transformed CHO and wild-type CHO-K1-BH4 cell lines have similar cytotoxic responses to UV light, the gpt-transformed cell lines respond differently from the parental CHO-K1-BH4 cell line in terms of mutation induction. As with CHO-K1-BH4 HPRT mutants, spontaneous or induced XPRT mutants derived from the gpt+ cell lines can be selected for 6-thioguanine resistance (TGr). Analysis of cell-free extracts from a number of these TGr clones indicates that the mutant phenotype is due to the absence of XPRT activity. One transformant, designated AS52, has previously been described in limited detail. Here we describe additional characteristics of this cell line, as well as several related transformants.     

Azaguanine resistant hamster cell lines not deficient in hypoxanthine-guanine phosphoribosyl transferase

Using a serial selection technique in which Chinese hamster cells were treated first with 8-azaguanine and then subsequently with HAT medium it was found that approximately 15% of azaguanine resistant clones were also resistant to HAT. Several such clones were subcultured and found to be stably resistant to azaguanine, in some cases at a higher level than the usual azaguanine resistant mutants which are HAT sensitive. Measurements of hypoxanthine-guanine phosphoribosyl transferase levels were in some cases lower than the parental line but in three of the clonal lines were higher than the parental strain. The fact that azaguanine resistant lines constitute a biochemically heterogeneous population underscores the importance of careful characterization of mammalian cell culture variants.     

Gene dosage and complementation analysis of ataxia telangiectasia lymphoblastoid cell lines assayed by induced chromosome aberrations

An approach of general applicability to mammalian radiosensitive mutants has been used in the analysis of gene dosage and complementation in ataxia telangiectasia (A-T). Thymidine residues in DNA of one parental lymphoblastoid cell line were substituted with bromodeoxyuridine before fusion with a second parental cell line, to allow differential staining of the two sets of chromosomes. Following gamma-irradiation, induced chromosome aberrations were scored in diploid and homokaryon cells from each parental line as well as in heterokaryons. Four complementation groups were ascertained among 7 A-T cell lines. Analysis of heterokaryons formed between appropriate combinations of normal, A-T homozygote and A-T heterozygote cells, gave a quantitative measure of gene dosage and demonstrated increasing radiosensitivity with increasing numbers of A-T alleles.