Kilobase Pairs In Size Research Articles

Encapsulation of cells in agarose beads for use in the construction of human DNA libraries as yeast artificial chromosomes (YAC)

A simple and general method for the molecular cloning of fragments of over one hundred kilobase pairs of exogenous DNA, by the encapsulation of cells in agarose beads, was developed for the construction of a human genomic DNA library in a yeast artificial YAC chromosome vector (in situ YAC construction). The main advantages of this method for use in the construction of a human genome library are as follows. First, linear DNA molecules of up to several hundred kilobase pairs in size can easily be prepared by the partial restriction enzyme digestion of the DNA encapsulated in agarose beads in vitro. Second, less than 2 x 10(6) cells scraped from tissue culture plates are sufficient for preparation of the linear DNA molecule for construction of the genome library. The technical manipulations involved in construction of clones of very large segments of DNA, including encapsulation of cells in agarose beads, restriction enzyme digestion, ligation with the YAC vector, transformation into host yeast cells, and stable propagation are discussed.

Construction and characterization of a yeast artificial chromosome library containing seven haploid human genome equivalents.

Prior to constructing a library of yeast artificial chromosomes (YACs) containing very large human DNA fragments, we performed a series of preliminary experiments aimed at developing a suitable protocol. We found an inverse relationship between YAC insert size and transformation efficiency. Evidence of occasional rearrangement within YAC inserts was found resulting in clonally stable internal deletions or clonally unstable size variations. A protocol was developed for preparative electrophoretic enrichment of high molecular mass human DNA fragments from partial restriction digests and ligation with the YAC vector in agarose. A YAC library has been constructed from large fragments of DNA from an Epstein-Barr virus-transformed human lymphoblastoid cell line. The library presently contains 50,000 clones, 95% of which are greater than 250 kilobase pairs in size. The mean YAC size of the library, calculated from 132 randomly isolated clones, is 430 kilobase pairs. The library thus contains the equivalent of approximately seven haploid human genomes.

Identification and characterization of plasmids from the western aster yellows mycoplasmalike organism

Supercoiled double-stranded DNA molecules (plasmids) were isolated from plants infected with three laboratory strains of western aster yellows mycoplasma-like organism (AY-MLO) by using cesium chloride-ethidium bromide density gradients. Southern blot analysis, using plasmids from the severe strain of AY-MLO (SAY-MLO) as the probe, identified at least four plasmids in celery, aster, and periwinkle plants and in Macrosteles severini leafhopper vectors infected with either the dwarf AY-MLO, Tulelake AY-MLO, or SAY-MLO strain. Plasmids were also detected in two California field isolates of AY-MLO but not in plants infected with the beet leafhopper-transmitted virescence agent, western X, or elm yellows MLOs. SAY-MLO plasmids were 5.2, 4.9, 3.4, and 1.7 kilobase pairs in size. Plasmids isolated from dwarf AY- and Tulelake AY-MLOs were 7.4, 5.1, 3.5, and 1.7 kilobase pairs in size. No evidence was obtained for integration of SAY-MLO plasmids into the MLO chromosome.

Size of the Streptococcus mutans GS-5 chromosome as determined by pulsed-field gel electrophoresis

Rare cutting restriction endonucleases were used to cut the Streptococcus mutans chromosome into large fragments. Restriction enzymes utilizing recognition sites containing 6-, 7-, or 8-base-pair sequences with only G and C nucleotides produced few fragments, most of which were greater than 100 kilobase pairs in size. Addition of the fragments from digests of SmaI, NotI, ApaI, RsrII, and EagI yielded a molecular size for the S. mutans GS-5 genome of 2,819 +/- 60 kilobase pairs.

The structure of the TATA-less rat tissue-type plasminogen activator gene. Species-specific sequence divergences in the promoter predict differences in regulation of gene expression.

The genomic region carrying the rat tissue-type plasminogen activator (tPA) gene including its 5'-flanking sequence has been isolated and characterized by restriction enzyme analysis, Southern blotting, and DNA sequencing of all coding parts and the promoter region. The gene is approximately 25 kilobase pairs in size and comprises 14 exons separated by 13 introns. All the exon/intron boundaries agree with the GT-AG rule. The organization of the rat tPA gene is very similar to its human counterpart, and the location of the introns in the protein structure is identical to the human tPA gene. To characterize the promoter region, the transcription initiation site was identified by S1 nuclease protection experiments. A DNA fragment carrying 621 nucleotides of the 5'-flanking sequence was found to confer basal promoter activity and hormone responsiveness to a reporter gene construct in primary cultures of rat granulosa cells. Analysis of the rat tPA promoter sequence and a comparison with the human and mouse counterparts reveal several species-specific differences: the rat and mouse tPA promoters lack typical TATA and CAAT sequences found in the human tPA gene. Furthermore, the rat tPA promoter contains a consensus cAMP-responsive element shown to be required for cAMP responsiveness in eucaryotic genes. At the same position as the cAMP-responsive element in the rat gene, the mouse and human tPA genes have a 12-O-tetradecanoylphorbol-13-acetate-responsive element known to mediate activation by phorbol esters. The differences in the promoter sequences of the rat, mouse, and human tPA genes may have implications for the regulation of the tPA gene in different species.

Isolation and sequence of ctaA, a gene required for cytochrome aa3 biosynthesis and sporulation in Bacillus subtilis

Cytochrome aa3 is one of two terminal oxidase complexes in the Bacillus subtilis electron transport chain. A novel genetic strategy was devised which permitted the isolation of B. subtilis mutants lacking cytochrome aa3 by selection for streptomycin-resistant clones which failed to oxidize the artificial electron donor N,N,N',N'-tetramethyl-p-phenylenediamine. Two mutations were studied intensively. Spectroscopic examination showed that each mutant lacked cytochrome aa3; they were also asporogenous and unable to grow on lactate as the sole carbon and energy source. These mutations were mapped to a locus designated ctaA, located at 127 degrees between pyrD and metC on the B. subtilis chromosome. Both ctaA mutations were closely linked by transformation to the pycA locus. The ctaA locus and a portion of the pycA locus were cloned from a B. subtilis integration library constructed in Escherichia coli. A recombinant plasmid containing a 4.0-kilobase insert of B. subtilis DNA could transform both ctaA mutants to CtaA+. Gene disruption and complementation experiments with subcloned fragments revealed that the ctaA locus consisted of a single transcriptional unit about 1.35 kilobase pairs in size. The nucleotide sequence of the ctaA transcriptional unit contains a single open reading frame capable of coding for a protein with a predicted molecular weight of 34,065. The predicted protein is extremely hydrophobic, with several probable membrane-spanning domains. No sequence similiarity was found between ctaA and the highly conserved procaryotic and mitochondrial oxidase polypeptides. Cloning and sequence analysis of two ctaA mutations revealed that one allele is a nonsense mutation in the carboxy terminus and the other is a missense mutation in the amino terminus; this indicates that the pleiotropic phenotype conferred by each mutation was caused by loss of CtaA or of its activity. Genetic evidence suggests that the ctaA gene product is required as an accessory protein in the genetic expression, posttranslational biogenesis, or both, of the cytochrome aa3 complex and during an early stage of sporogenesis.

Identification of cellular promoters by using a retrovirus promoter trap

A retrovirus vector has been developed that selects for instances in which the provirus integrates in close proximity to cellular promoters. Coding sequences for a selectable marker (histidinol dehydrogenase) were inserted into the 3' long terminal repeat (LTR) of an enhancerless Molony murine leukemia virus. Although 1.8 kilobase pairs in size, the elongated LTR did not appear to interfere with virus replication or integration. Thus, when the virus was passaged, the elongated LTRs efficiently duplicated, placing histidinol dehydrogenase-coding sequences in the 5' LTR just 30 nucleotides from the flanking cellular DNA. Selection for histidinol expression generated cell clones in which histidinol gene sequences in the 5' LTR were invariably expressed on transcripts initiating in nearby cellular sequences. The efficiency of transducing histidinol resistance was 2,500-fold lower than the efficiency of transducing neomycin resistance when the neomycin phosphotransferase gene was located within the retrovirus and expressed from an independent promoter. By tagging transcriptionally active sites, the vector provides a means to identify and isolate promoters active in different cell types. Furthermore, the virus may be useful as an insertional mutagen, since selection for cell populations containing proviruses only in expressed sites is expected to reduce the number of intergrants needed to screen for loss of gene function.

Construction of a lactococcal expression vector: expression of hen egg white lysozyme in Lactococcus lactis subsp. lactis

A pair of vectors for expression of heterologous genes in Lactococcus lactis was constructed. In addition to an origin of replication that has a broad host range, these vectors contain a multiple cloning site flanked by gene expression signals originating from L. lactis subsp. cremoris Wg2. The two vectors, about 3.7 kilobase pairs in size, differ only in the type of antibiotic resistance they confer to their hosts. pMG36 carries a kanamycin resistance marker, which was replaced by an erythromycin resistance marker in pMG36e. As an example of the use of these vectors, the hen egg white lysozyme-coding sequence was inserted. A fusion protein of the expected size was detected in a transformed L. lactis subsp. lactis strain by using Western blotting (immunoblotting).

Molecular cloning and expression in Escherichia coli of a cellodextrinase gene from Bacteroides succinogenes S85.

A DNA fragment coding for a cellodextrinase of Bacteroides succinogenes S85 was isolated by screening of a pBR322 gene library in Escherichia coli HB101. Of 100,000 colonies screened on a complex medium with methylumbelliferyl-beta-D-cellobioside as the indicator substrate, two cellodextrinase-positive clones (CB1 and CB2) were isolated. The DNA inserts from the two recombinant plasmids were 7.7 kilobase pairs in size and had similar restriction maps. After subcloning from pCB2, a 2.5-kilobase-pair insert which coded for cellodextrinase activity was isolated. The enzyme was located in the cytoplasm of the E. coli host. It exhibited no activity on carboxymethyl cellulose, Avicel microcrystalline cellulose, acid-swollen cellulose, or cellobiose but hydrolyzed p-nitrophenyl-beta-D-cellobioside and p-nitrophenyl-beta-D-lactoside. The Km (0.1 mM) for the hydrolysis of p-nitrophenyl-cellobioside by the enzyme expressed in E. coli was similar to that reported for the purified enzyme from B. succinogenes. Expression of the cellodextrinase gene was subjected to catabolite repression by glucose and was not induced by cellobiose. The origin of the DNA insert from B. succinogenes was confirmed by Southern blot analysis. Western blotting (immunoblotting) using antibodies raised against the purified B. succinogenes cellodextrinase revealed a protein with a molecular weight of approximately 50,000 in E. coli clones which comigrated with the native enzyme isolated from B. succinogenes. These data indicate that the cellodextrinase gene expressed in E. coli is fully functional and codes for an enzyme with properties similar to those of the native enzyme.

Complete nucleotide sequence of the chromosomal gene for human IL-4 and its expression.

We have isolated a chromosomal DNA segment of the human IL-4 gene based on homology with a human IL-4 cDNA sequence and determined its complete nucleotide sequence. The human IL-4 gene, which occurs as a single copy in the haploid genome, is mapped on chromosome 5. It is composed of four exons and three introns and is approximately 10 kilobase pairs in size. 5'-Flanking regions of human and mouse IL-4 genes share about 85% homology extending more than 500 base pairs upstream of a "TATA" like sequence. Several patches of sequences are found in the 5'-flanking region of the human IL-4 gene which are homologous to sequence in the 5'-flanking regions of the IL-2, IL-3, IL-5, and granulocyte-macrophage (GM)-CSF genes. The IL-4 gene is inducible after treatment of human T cell clone by phorbol-12-myristate-13-acetate (TPA) and calcium ionophore A23187. The 2.3-kb 5'-flanking region of the human IL-4 gene transiently transfected into Jurkat human T cell leukemia cells is activated efficiently in response to TPA and A23187 stimulation and, although less efficiently, by human T cell leukemia virus type I-encoded p40x or BPV-encoded E2 protein. Combination of TPA/A23187 and p40x or E2 protein further augmented the level of expression.

Plasmid DNA from the acetotrophic methanogen Methanosarcina acetivorans.

Nine acetotrophic and three methylotrophic strains of methane-producing bacteria were screened for the presence of plasmid DNA. Plasmids were detected in three marine isolates, including Methanosarcina acetivorans. All three plasmids appeared to be similar based on size and restriction site analyses. The plasmid from M. acetivorans, designated pC2A, was approximately 5.1 kilobase pairs in size and was estimated to be present in a low copy number of six plasmids per genome. Multimers were also observed. A restriction map was constructed. The function of this plasmid is cryptic.

Isolation and characterization of a cDNA that encodes the peptide core of the secretory granule proteoglycan of human promyelocytic leukemia HL-60 cells.

A cDNA that encodes the peptide core of the secretory granule proteoglycan of the human promyelocytic leukemic cell line, HL-60, has been isolated and analyzed. When human genomic DNA was digested and probed under conditions of low stringency with a rat cDNA that encodes a Mr = 18,600 serine/glycine-rich proteoglycan peptide core in L2 yolk sac tumor cells (Bourdon, M. A., Oldberg, A., Pierschbacher, M., and Ruoslahti, E. (1985) Proc. Natl. Acad. Sci. U.S.A. 82, 1321-1325) and basophilic leukemia-1 cells (Avraham, S., Stevens, R. L., Gartner, M. C., Austen, K. F., Lalley, P. A., and Weis, J. H. (1988) J. Biol. Chem. 263, 7292-7296), a number of DNA fragments were identified. A HL-60 cell-derived cDNA library was therefore screened under conditions of low stringency with the rat probe to identify and isolate a human homologue of this rat proteoglycan peptide core. Analysis of the resulting human cDNA clones indicated that the proteoglycan peptide core that is expressed in HL-60 cells is Mr = 17,600 and contains an 18-amino acid glycosaminoglycan attachment region that consists primarily of alternating serin and glycine. Northern blot analysis of total RNA probed with the human cDNA revealed that the major message for this proteoglycan peptide core in HL-60 cells is approximately 1.3 kilobase pairs in size. When a Southern blot of digested human genomic DNA was probed with the human cDNA, three bands of approximately 6, 9, and 12 kilobase pairs were detected. However, when the Southern blot was probed with the XmnI----3' fragment of this human cDNA, one prominent band was detected, indicating that a single gene encodes this protein in the human. Analysis of the DNA from human/mouse and human/hamster somatic cell hybrids probed with the human cDNA demonstrated that the gene that encodes this molecule resides on human chromosome 10. Because the proteoglycans that are present in the secretory granules of different types of rat and mouse mast cells possess small peptide cores that are rich in serine and glycine, we propose that this HL-60 cell-3 derived cDNA encodes the peptide core of the proteoglycan that is expressed in the secretory granules of this human promyelocytic cell.

Characterization of the promoter region of the Bacillus subtilis spoIIE operon.

Mutations that define the spoIIE locus of Bacillus subtilis block sporulation at an early stage and recently were shown to prevent the proteolytic processing of sigma E (sigma 29) into its active form, an event that is believed to control critical changes in gene expression during the second hour of development. By taking advantage of two Tn917-mediated insertional mutations in spoIIE, we have cloned DNA spanning the locus. Gene disruption experiments with subcloned fragments transferred to integrational vectors revealed that the locus consisted of a single transcription unit about 2.5 kilobase pairs in size. Transcriptional lacZ fusions were used to show that expression of this transcription unit initiated at 1.5 h after the end of log-phase growth and depended upon the products of all spo0 loci. Expression was directed by a single promoter whose position was determined by high-resolution S1 protection mapping. A deletion analysis of the promoter region was also carried out, with novel integrational vectors based on derivatives of coliphage M13. The results indicated that a region of DNA extending from 183 to 118 base pairs upstream from the start point of transcription was required for full activity of the spoIIE promoter. The presumptive RNA polymerase-binding region of the promoter exhibited striking similarity to the spoIIG promoter and featured perfect but unusually spaced -10 and -35 consensus sequences for sigma A (sigma 43)-associated RNA polymerase.

Complete primary structure of the rat cartilage proteoglycan core protein deduced from cDNA clones.

We have obtained overlapping cDNA clones for the entire coding sequence of the rat cartilage proteoglycan core protein from the Swarm rat chondrosarcoma. These cDNAs hybridize to two sizes of RNA transcripts of 8.2 and 8.9 kilobase pairs, which contain large 3'-untranslated sequences. The total contiguous cDNA is 6.55 kilobase pairs in size, and codes for a 2124-residue protein, including a 19-residue signal peptide. The sequence forms a series of eight structural domains including two globules, (Mr = 37,000 and 22,000) at the NH2 terminus of the molecule, one a complete and one a partial copy of the cartilage link protein. The major feature of the deduced protein sequence is a 1,104-residue segment containing 117 Ser-Gly sequences, the presumed chondroitin sulfate attachment sites. These are arranged in three domains of 428, 503, and 173 amino acids. The first domain contains 11 complete or partial repeats of a 40-residue unit, and the second domain is composed of six copies of a 100-residue repeating sequence. The first pattern is the more highly conserved, and may have given rise to the second. The carboxyl-terminal domain is a third globule which has homology with animal lectins.

Characterization of the human blood coagulation factor XII gene. Intron/exon gene organization and analysis of the 5'-flanking region.

A human genomic phage library was screened using a human factor XII cDNA as a hybridization probe. Two overlapping phage clones were isolated which contain the entire human factor XII gene. DNA sequence and restriction enzyme analysis of the clones indicate that the gene is approximately 12 kilobase pairs in size and is comprised of 13 introns and 14 exons. Exons 3-14 are contained in a genomic region of only 4.2 kilobase pairs with introns ranging in size from 80 to 554 base pairs. The coding sequence of factor XII consists of multiple putative domains that are homologous to putative domains found in fibronectin and tissue-type plasminogen activator. These regions were found as separate exons in the gene. The intron/exon gene organization is similar to the serine protease gene family of plasminogen activators and not to the clotting factor family. Analysis of the 5' end of the gene shows that it does not contain the typical TATA and CAAT sequences found in other genes. This is consistent with the finding that transcription of the gene is initiated at multiple sites.

Molecular cloning and characterization of the glucosyltransferase C gene (gtfC) from Streptococcus mutans LM7.

A glucosyltransferase (GTF) gene, designated gtfC, was cloned from Streptococcus mutans LM7. Its gene product was detected by screening a bacteriophage lambda library with rabbit antiserum raised against S. mutans LM7 extracellular proteins. DNA isolated from the immunopositive recombinant phage revealed two S. mutans chromosomal EcoRI fragment inserts, 8.1 and 4.7 kilobase pairs in size. Escherichia coli minicell analyses revealed the approximate position and direction of transcription of the gtfC gene. The gene product was determined to be a polypeptide of about 150 kilodaltons which synthesized a water-soluble glucan. Restriction endonuclease mapping and DNA hybridization indicated a repeated region of DNA corresponding to a portion of the coding region of gtfC immediately downstream from the intact gtfC locus on the chromosome. A 300-base-pair gtfC-specific probe showed that the gene and the putative duplicated sequence were present in S. mutans serotypes c, e, and f, but not in other related oral streptococci which had GTF activity. In addition, the gtfC determinant displayed homology to sequences corresponding to the carboxy-terminal coding region of a gene (gtfB) encoding a GTF activity that synthesized water-insoluble glucans. These data suggest that at least one class of GTF genes may be present in multiple copies in S. mutans and, further, that GTF genes may contain conserved sequences internal to their coding regions.

CDNA sequence and tissue distribution of the mRNA for bovine and murine p11, the S100-related light chain of the protein-tyrosine kinase substrate p36 (calpactin I).

We have isolated and sequenced cDNA clones of bovine and murine p11 mRNAs. The nonpolyadenylated mRNAs are predicted to be 614 and 600 nucleotides, respectively. The p11 mRNAs both contain a 291 nucleotide open reading frame, preceded by a 5'-untranslated region of 73 nucleotides in bovine p11 mRNA and of 68 nucleotides in murine p11 mRNA. The deduced bovine p11 amino acid sequence is identical to the previously published partial bovine and complete porcine p11 protein sequence except for an additional COOH-terminal lysine residue. The bovine and murine p11 proteins are 92% homologous, whereas at the nucleotide level the conservation is 89% in the coding region and 75% in the 3'-untranslated region. Southern analysis of murine genomic DNA detected a single p11 gene, less than 10 kilobase pairs in size, containing as many as three introns. The p11 gene has been assigned to mouse chromosome 3 by analysis of interspecific hybrid cell panels and recombinant inbred mouse strains. The p11 gene is closely linked to the Xmmv-65 endogenous leukemia virus env gene and the guanylate binding protein-1 gene. Northern analyses of RNAs from mouse tissues and cell lines indicated that p11 mRNA levels vary widely. They are very low in liver, heart, and testes, moderate in brain, spleen, and thymus, and high in kidney, intestine, and lung. Analysis of the same RNA samples for p36 mRNA levels showed that expression of p11 and p36 mRNAs is not always coordinated. Brain and the mouse embryonal carcinoma cell line F9 contain moderate to high levels of p11 mRNA with very low levels of p36 mRNA. Sequence homology between p11 and the S100 proteins, and the serum-induced 2A9 gene product, as well as possible functions of p11 are discussed.

High interindividual restriction fragment length and copy number of polymorphism of a tvri family in moderate human DNA repeats

The authors describe the selection of cloned human DNA sequences, with a copy number not exceeding 1000 copies per diploid genome, and their testing for interindividual restriction fragment lengths and copy number of polymorphism (RFLCP). As a result of the investigation a DNA clone was found (TVRI-6), about 2.8 kilobase-pairs in size, for which an unusually high level of interindividual RFLCP was discovered. The TVRI-6 sequence was obtained from a bank of Pst I restriction fragments of human placental nuclear DNA cloned in pBR 322. The bank was analyzed by hybridization of colonies with phosphorus 32-labelled human nuclear DNA.

Aerobactin biosynthesis and transport genes of plasmid ColV-K30 in Escherichia coli K-12.

The iron-regulated aerobactin operon, about 8 kilobase pairs in size, of the Escherichia coli plasmid ColV-K30 was shown by deletion and subcloning analyses to consist of at least five genes for synthesis (iuc, iron uptake chelate) and transport (iut, iron uptake transport) of the siderophore. The gene order iucABCD iutA was established. The genes were mapped within restriction nuclease fragments of a cloned 16.3-kilobase-pair HindIII fragment. Stepwise deletion and subsequent minicell analysis of the resulting plasmids allowed assignment of four of the five genes to polypeptides of molecular masses 63,000, 33,000 53,000, and 74,000 daltons, respectively. The 74-kilodalton protein, the product of gene iutA, is the outer membrane receptor for ferric aerobactin, whereas the remaining three proteins are involved in biosynthesis of aerobactin. The 33-kilodalton protein, the product of gene iucB, was identified as N epsilon-hydroxylysine:acetyl coenzyme A N epsilon-transacetylase (acetylase) by comparison of enzyme activity in extracts from various deletion mutants. The 53-kilodalton protein, the product of gene iucD, is required for oxygenation of lysine. The 63-kilodalton protein, the product of gene iucA, is assigned to the first step of the aerobactin synthetase reaction. The product of gene iucC, so far unidentified, performs the second and final step in this reaction. This is based on the chemical characterization of two precursor hydroxamic acids (N epsilon-acetyl-N epsilon-hydroxylysine and N alpha-citryl-N epsilon-acetyl-N epsilon-hydroxylysine) isolated from a strain carrying a 0.3-kilobase-pair deletion in the iucC gene. The results support the existence of a biosynthetic pathway in which aerobactin arises by oxygenation of lysine, acetylation of the N epsilon-hydroxy function, and condensation of 2 mol of the resulting aminohydroxamic acid with citric acid.

Identification of functional murine adenosine deaminase cDNA clones by complementation in Escherichia coli.

Total poly(A+) RNA derived from a mouse cell line with amplified adenosine deaminase genes was used as template to synthesize double-stranded cDNA. The cDNAs were inserted into the PstI site of the beta-lactamase gene in plasmid pBR322 following G-C tailing. After transformation into adenosine deaminase-deficient Escherichia coli hosts, recombinant plasmids containing functional murine adenosine deaminase cDNAs were identified by selecting for functional complementation. Analysis of plasmids containing functional adenosine deaminase cDNA sequences strongly suggested that adenosine deaminase expression resulted mainly from beta-lactamase/adenosine deaminase fusion proteins even when the adenosine deaminase codons were out-of-frame with respect to the beta-lactamase gene codons upstream. The nucleotide sequence of a 1.65-kilobase pair cDNA insert in one of the functional recombinant clones was determined and found to contain a 1056-nucleotide open reading frame. When this 1056-nucleotide open reading frame was inserted into a mammalian expression vector and introduced into monkey kidney cells, a high level of authentic mouse adenosine deaminase was produced. Nucleic acid blot analysis using a full-length adenosine deaminase cDNA clone as probe revealed that the mouse adenosine deaminase structural gene was at least 21 kilobase pairs in size and encoded three polyadenylated mRNAs. Analysis of the cDNA library from which the functional clones were isolated suggested that this approach of cloning functional mammalian adenosine deaminase cDNA clones by genetic complementation of enzyme-deficient bacteria could be accomplished even if the abundance of the adenosine deaminase mRNA sequences were as low as approximately 0.001%.