Gene amplification accompanies low level increases in the activity of dihydrofolate reductase in antifolate-resistant Chinese hamster lung cells containing abnormally banding chromosomes.

Three independently-derived, antifolate-resistant Chinese hamster lung cell lines that exhibit low level increases in dihydrofolate reductase (DHFR) activity, i.e., three- to fivefold vs. controls, have been compared with drug-sensitive cells to determine relative DHFR gene content. With a solution hybridization technique that makes use of genomic DNA and a cloned double-stranded Chinese hamster DHFR cDNA probe, it has been found that the enzyme activity increases are associated with an approximately proportionate amplification of DHFR genes. Trypsin-Giemsa staining techniques and hybridizations in situ further show that the amplified DHFR genes are located within abnormally banding regions along chromosome 2q and also suggest that, in each subline, only one chromosome 2 homolog is initially involved in the amplification process.

Gene amplification-associated cytogenetic aberrations and protein changes in vincristine-resistant Chinese hamster, mouse, and human cells.

We carried out cytogenetic studies of four Chinese hamster, mouse, and human cell lines selected for high levels of resistance (500- to 4,000-fold) to vincristine (VCR) by a multistep selection procedure. All cells examined contained gene amplification-associated metaphase chromosome abnormalities, either homogeneously staining regions (HSRs), abnormally banding regions (ABRs), or double-minute chromosomes (DMs); control actinomycin D- and daunorubicin-resistant hamster lines did not exhibit this type of chromosomal abnormality. VCR-resistant Chinese hamster sublines exhibited both increased synthesis of the protein V19 (Mr 19,000; pl = 5.7) and increased concentrations of V19 polysomal mRNA. When VCR-resistant cells were grown in drug-free medium, level of resistance, synthesis of V19, and amount of V19 mRNA declined in parallel with mean length of the HSR or mean number of DMs per cell. Cross-resistance studies indicate that VCR-resistant cells have increased resistance both to antimitotic agents and to a wide variety of agents unrelated to VCR in chemical structure and/or mechanism of action. Our studies of tubulin synthesis in Chinese hamster cells indicate no overproduction of tubulin or presence of a mutant tubulin species. Comparison with antifolate-resistant Chinese hamster cells known to contain amplified dihydrofolate reductase genes localized to HSRs or ABRs strongly suggests that the HSRs, ABRs, or DMs of the Vinca alkaloid-resistant sublines likewise represent cytological manifestations of specifically amplified genes, possibly encoding V19, involved in development of resistance to VCR.

Amplification and expression of genes associated with multidrug resistance in mammalian cells.

In multidrug resistance, which is observed clinically and in tissue culture, cells that are challenged with certain cytotoxic drugs develop resistance not only to the selective agent but also to other, seemingly unrelated, agents. The multidrug-resistant phenotype is associated with DNA sequence amplification and with the overproduction of a number of cytosolic and membrane glycoproteins. The differential amplification and altered expression of at least two related genes, termed multidrug-resistant associated genes has been shown in multidrug-resistant Chinese hamster cells. In multidrug-resistant mouse and human cells, genes homologous to those in Chinese hamster cells are also amplified. The level of expression of these genes varied and did not correlate with their copy number. Furthermore, in Chinese hamster cells, the development of resistance to a single drug and multidrug resistance were closely related, but uncoupled, events. The overexpression of the multidrug-resistant genes was better correlated with the degree of resistance to the selective agent than it was with the extent of multidrug resistance.

Starvation phase of Physarum polycephalum: characterization of transfer ribonucleic acid.

We have begun a series of studies designed to characterize gene expression during differentiation in the slime mold Physarum polycephalum. This work concerns the starvation phase of the sporulation sequence and describes some of the quantitative changes which occur in plasmodial constituents during the 3-day starvation period and also describes alterations in the transfer ribonucleic acid (tRNA) population. The results show that whereas the plasmodial tRNA content decreased by 75% during starvation, concurrent de novo synthesis of tRNA also occurred, and they also show that overall amino acid acceptor activity of the starvation-phase tRNA population did not differ significantly from that found in the growth phase. Of the 19 starvation-phase tRNA families assayed, however, 6 were found to have consistently lower acceptor activities than did their growth-phase counterparts. Reverse-phase (RPC-5) chromatographic analysis of five of those families failed to reveal any major differences between growth- and starvation-phase isoacceptors. The data suggest that the depletion and resynthesis of tRNA during the starvation phase results in a quantitative alteration in the composition of the tRNA population and that the alteration is tRNA family and not tRNA isoacceptor specific.

Expression of the amplified domain in human neuroblastoma cells.

Screening of a partial cDNA library prepared from the human neuroblastoma cell line BE(2)-C with genomic DNA probes containing sequences representative of the amplified domain of that cell line allowed us to identify cloned transcripts from an active gene within the domain. The gene BE(2)-C-59 is amplified ca. 150-fold and encodes a 3.0- and a 1.5-kilobase RNA transcript, both of which are overproduced in BE(2)-C cells. A survey of a large variety of human tumor cell types indicated that this gene is amplified to varying degrees in all neuroblastoma cell lines and a retinoblastoma cell line that exhibit obvious cytological manifestations of DNA sequence amplification, i.e., homogeneously staining regions and double-minute chromosomes. The BE(2)-C-59 gene is not amplified, however, in other nonrelated tumor types, even those containing amplified DNA. Although the functional significance of this specific gene amplification in neuroblastoma cells remains unknown, an indication that it may relate to the malignant phenotype of these cells follows from the remainder of our data which show that the amplified BE(2)-C-59 gene shares partial homology with both the second and third exons, but not the first exon, of the human c-myc oncogene.

Phenotypic expression in Escherichia coli and nucleotide sequence of two Chinese hamster lung cell cDNAs encoding different dihydrofolate reductases.

Nucleotide sequence analysis of two cDNA clones, one shown to direct the synthesis in Escherichia coli of the pI 6.7 form of the 20,000-molecular-weight class of Chinese hamster lung cell dihydrofolate reductase, and the other shown to direct the synthesis of the pI 6.5 form of the 21,000-molecular-weight class of the enzyme, has revealed the following: (i) the differences in physical and enzymatic properties displayed by these two proteins are due to two variations in their respective amino acid sequences with the conversion of Leu to Phe at position 22 probably responsible for the differential sensitivity of these two enzymes to methotrexate and methasquin; (ii) the multiple mRNAs responsible for the synthesis of each of these proteins differ in size due, at least in part, to a length heterogeneity at their 3' ends; (iii) these two proteins are encoded by different genes; and (iv) the sequence AAATATA appears to be a major polyadenylation signal in one Chinese hamster lung cell dihydrofolate reductase gene and a minor signal in another.

Allelic variation in the dihydrofolate reductase gene at amino acid position 95 contributes to antifolate resistance in Chinese hamster cells.

The Chinese hamster lung cell line DC-3F contains two polymorphic dihydrofolate reductase (DHFR) alleles that are defined by an Asp-Asn amino acid sequence difference at position 95 in protein. Previously, we reported that the antifolate-resistant subline DC-3F/A3 overexpressed a Leu22-->Phe mutant of the Asp95 (21k) allele and that this was the basis of its resistance to methotrexate (MTX) and methasquin [P. W. Melera, J. P. Davide, C. A. Hession, and K. W. Scotto, Mol. Cell. Biol., 4: 38-48, 1984]. We now show that another independently selected antifolate-resistant subline of DC-3F, DC-3F8/A55, in addition to being severely compromised in its ability to accumulate MTX, overexpresses a Leu22-->Phe mutant form of the Asn95 (20k) allele. Characterization of purified DHFR from these cells showed that the enzyme displayed a 6-fold higher Kd for MTX (3.92 +/- 0.17 pM) than the wild type (0.58 +/- 0.10 pM), thus explaining its lowered sensitivity to drug. Unexpectedly, however, this value was 4-fold lower than that displayed by the DC-3F/A3 enzyme even though both contain the same (Leu22-->Phe) mutation and differ only at position 95. Indeed, we have also shown that the 21k and 20k wild type enzymes, both containing Leu at position 22, in fact differ by 3-fold (1.58 +/- 0.08 and 0.58 +/- 0.10 pM, respectively) in their Kd's for MTX. This demonstrates that the amino acid at position 95 has an effect on the ability of DHFR to bind MTX. On the other hand, these allelic variants are indistinguishable from each other in their catalytic properties and in their respective Kd's for dihydrofolate. Taken together, these characteristics are consistent with the observation that it is the wild type 21k allele which is preferentially overexpressed at a frequency of 3:1 in MTX-resistant Chinese hamster lung sublines derived by long-term selection in MTX. The results of these studies are novel in that they establish a role for allelic variation in the DHFR gene as a contributor to antifolate resistance in mammalian cells. Moreover, they implicate amino acid position 95 in the maintenance of the structure of the MTX binding pocket.

Metallothionein is overexpressed by hamster fibroblasts selected for growth in 15 pm folinic acid and provides a growth advantage in low folate.

DC-3F/FA3 (FA3) cells, selected for growth in folic acid-free medium containing dialyzed serum and 15 pM [6S]-folinic acid, and parental DC-3F cells were compared by mRNA differential display to identify genetic changes occurring during selection. One of the genes found to be overexpressed in FA3 cells was metallothionein II (MT-II). Northern blots using a full-length hamster MT-II cDNA probe that recognizes both MT-I and MT-II RNA showed that the steady-state level of MT mRNA was elevated at least 10-fold in FA3 cells and in two other selected clones, FA7 and FA14, as well. Southern blot analysis of HindIII-digested genomic DNA indicated that amplification of neither the MT-I nor MT-II gene had occurred, and measurements of MT mRNA decay rates in the presence of actinomycin D suggested that no changes in its half-life had taken place. Hence, overexpression was due to an increase in transcription from the normal gene complement. In FA3 cells, the MT mRNA expression level was found to be directly sensitive and inversely proportional to media folate concentrations, whereas in DC-3F cells it was not, suggesting that MT gene expression is differentially regulated in these two cell lines. Overexpression of MT-II in transfected DC-3F cells was unable to support growth in 15 pM folinic acid. However, when plated in 15 nM folinic acid, a growth rate similar to FA3 cells was observed, whereas sham-transfected controls and double transfectants expressing antisense MT-II RNA and control levels of MT-II protein ceased to grow. Hence, overexpression of MT-II provides a growth advantage in low folate.

Chromosomal organization of amplified genes in multidrug-resistant Chinese hamster cells.

Six independently derived multidrug-resistant Chinese hamster cell lines selected with vincristine, daunorubicin, actinomycin D, or colchicine were probed by in situ hybridization techniques with the cloned cDNA, p5L-18, to chromosomally localize known or presumed amplified P-glycoprotein genes. One or two clusters of amplified genes were demonstrable in all of the highly resistant sublines and were localized to homogeneously staining regions and/or abnormally banding regions, gene amplification-associated cytogenetic abnormalities, on eight different chromosomes. Analysis of trypsin-Giemsa banded karyotypes revealed additional, multiple chromosomal rearrangements that were apparently nonspecific. Mapping studies localized the native P-glycoprotein gene(s) to the region q23 to 31 (most probably band 26) on the long arm of chromosome 1 of normal Chinese hamster bone marrow fibroblasts and normal chromosome 1 homologues in resistant cells. Southern blot analysis of restriction endonuclease fragments indicated the amplification of one or both of (at least) two wild-type nonallelic genes in four of the lines and the presence in one line (DC-3F/DMM XX) of a unique 5.0-kilobase BamHI fragment resulting from a recombinational event during amplification. Comparison with the cytogenetic data indicated no correlation between restriction patterns generated by EcoRI, HindIII, PstI, or BamHI and chromosomal location of amplified genes. However, the only sublines in which the homogeneously staining region or abnormally banding region is positioned at 1q26 (at or near the site of the native gene) exhibit either alterations in gene structure (DC-3F/DM XX) or in regulation of gene expression (DC-3F/AD X), suggesting a process more complex than simply amplification of the gene in loco.

Isolation of amplified and overexpressed DNA sequences from adriamycin-resistant human breast cancer cells.

An MCF-7 human breast cancer cell line was selected which was 200-fold more resistant to Adriamycin than the wild type cell line. This Adriamycin-resistant (AdrR) cell line exhibited a multidrug-resistant phenotype and was cross-resistant to a wide range of antineoplastic agents including Vinca alkaloids, anthracyclines, and epipodophyllotoxins. Cytogenetic analysis of the AdrR cell line showed the presence of homogeneously staining regions on several chromosomes which were not present in the parental cell line. Using the technique of in-gel renaturation, DNA sequences which were amplified 50- to 100-fold in the AdrR cell line and which covered a total of over 140 kilobases were isolated. In addition, AdrR cells were found to contain amplified and overexpressed sequences which were homologous to hamster P-glycoprotein gene sequences. A hamster cDNA P-glycoprotein gene probe was used to screen a lambda gt10 cDNA library made from human AdrR cell line mRNA and human cDNA sequences homologous to the P-glycoprotein gene were isolated. Hybridization studies with the cloned human cDNA (pADR1) showed that the AdrR MCF-7 cell line contained a 60-fold amplification of this DNA sequence and that polyadenylated mRNA from the AdrR cell line contained a 4.8-kilobase transcript which was overexpressed 45-fold. There was a direct correlation between DNA and RNA copy number of this sequence and level of resistance among several MCF-7 Adriamycin-resistant cell lines. In situ hybridization studies demonstrated that the human P-glycoprotein gene sequence was found on chromosome 7q21.1 in normal human lymphocytes and that amplified DNA sequences isolated from the AdrR MCF-7 cells by the in-gel hybridization technique were linked to the human P-glycoprotein sequences in the homogeneously staining regions in the AdrR cells.

Investigation of human lymphocyte plasma membrane associated nucleic acid.

Plasma membranes were prepared from the human lymphocyte cell line WIL23A by hypotonic swelling, Dounce homogenization, differential and equilibrium centrifugation. The resulting vesiculated membrane fragments were found to have densities of 1.10 and 1.17 g/ml, and were defined by lactoperoxidase mediated whole cell iodination, L-[3H] fucose incorporation, 5'-nucleotidase activity (EC 3.1.3.5) and electron micrographic visualization. Recovery of plasma membrane from whole cell homogenates was estimated to be approximately 30-35% as judged by the recovery of 125I-labeled cell surface protein. When plasma membranes were prepared from cells which had been incubated for 18 h in the presence of 0.5 muCi/ml [3H] thymidine such that greater than 10(9) acid insoluble counts could be demonstrated in the whole cell homogenates, no [3H] thymidine label and presumably, therefore, no DNA, could be shown to be coincident with either the 1.10 or 1.17 density. Similar experiments with [3H] uridine suggested that 90% of the plasma membranes did not contain RNA, while 10% remained questionable.

The effects of differential polyadenylation on expression of the dihydrofolate reductase-encoding gene in Chinese hamster lung cells.

Three differently sized mRNAs are expressed from each of two DHFR (encoding dihydrofolate reductase) alleles present in the Chinese hamster lung (CHL) cell line, DC-3F. The relative abundancy of the transcripts produced from each allele differs dramatically as a result of differential utilization of the multiple poly(A) sites present in the DHFR DHFR gene and a genetic polymorphism located within the third poly(A) signal of one allele. We sought to determine whether such differences in polyadenylation affect the steady-state levels of DHFR and mRNAs expressed from either allele and, in a more general sense, to ask whether differences in 3' end RNA processing in a gene containing multiple poly(A) sites affects the final level of gene expression. An SV40 promoter-based transient expression system producing chimeric cat::DHFR transcripts was developed to regenerate the in vivo mRNA polyadenylation patterns associated with each of the two DHFR alleles. The results demonstrate that the total amount of polyadenylated RNA expressed from each of these constructs in vitro is the same regardless of the differential utilization of the poly(A) signals that occurs between them. Moreover, measurement of the individual turnover rates of the DHFR mRNAs expressed in vivo from each allele, as determined by pulse-chase labeling and actinomycin D inhibition studies, revealed no significant allele-specific differences in transcript half-lives. Finally, measuring the steady-state levels of DHFR poly(A)+ mRNA in parental DC-3F cells demonstrated that both alleles are expressed to the same extent during normal growth. Thus, even though dramatic allele-specific differences in 3' end processing of DHFR transcripts occur in vivo, such differences do not appear to influence the steady-state levels of DHFR gene expression.

Construction of a dominant selectable marker using a novel dihydrofolate reductase.

Simian virus 40 promoter-enhancer-based mammalian expression plasmids using dihydrofolate reductase (DHFR)-encoding cDNA sequences originally isolated from two methotrexate (MTX)-resistant, DHFR-overproducing Chinese hamster lung cell lines were constructed. One, designated pSVA75, contains a DHFR cDNA that encodes leucine (Leu22) and corresponds to the wild type (wt), MTX-sensitive form of the enzyme [Melera et al., J. Biol. Chem. 263 (1988) 1978-1990]. The other plasmid, pSVA3, contains a cDNA that encodes a novel mutant form of the enzyme in which Leu22 has been changed to Phe [Melera et al., Mol. Cell Biol. 4 (1984) 38-48]. The resulting DHFR displays a 20-fold-enhanced resistance to inhibition by MTX, but maintains the catalytic activity of the wt enzyme [Albrecht et al., Cancer Res. 32 (1972) 1539-1546]. Transfection of DHFR- Chinese hamster ovary cells with either plasmid demonstrated that both were able to reconstitute the DHFR+ phenotype with equal efficiency (i.e., greater than 2.5 x 10(-3), indicating that both the wt and mutant enzymes were catalytically active in transfected cells. In addition, the mutant form of the enzyme was found to act as a dominant selectable marker when transfected into diploid DHFR+ cells, and to allow selection of resistant clones at low MTX concentrations (125 nM MTX) with a frequency of greater than 8 x 10(-4). Moreover, transfected clones were found to amplify their exogenous DHFR sequences to reasonably high levels (42-fold) at relatively low (888 nM) MTX concentrations, suggesting that substantial amplification of DHFR DNA and cotransfected sequences as well, can be achieved with this vector.

Selection for the expression of one form of Chinese hamster dihydrofolate reductase over another during growth in methotrexate.

Two mammalian expression plasmids, each carrying a cDNA encoding a different allele of dihydrofolate reductase (DHFR) present in the Chinese hamster lung cell line DC-3F, were constructed. These simian virus 40 promoter-enhancer-based plasmids, designated pSVA75 and pSVMQ19, are identical except for a G-->A transition at nucleotide 286 of the DHFR coding sequence. Due to this change, the enzyme expressed by pSVA75 contains Asp95, while the enzyme expressed by pSVMQ19 has Asn95 [Melera et al., J. Biol. Chem. (1988) 1978-1990]. Both forms of the enzyme are catalytically equivalent and are produced to similar levels in DC-3F cells [Yu and Melera, Cancer Res. (1993) 6031-6035; H.Y., A.H. and P.W.M., in preparation]. Clonal cell lines expressing one or the other DHFR allele were obtained via transfection of DHFR- Chinese hamster ovary cells, and 74 clones of each type isolated. These were pooled and divided into 40 aliquots, each of which was then subjected to selection by growth in sequentially increasing concentrations of methotrexate (MTX). Analysis of the resulting drug-resistant populations revealed that cells producing Asp95 DHFR dominated with an overall frequency of 3:1, and therefore, under these growth conditions, display a selective advantage over those producing Asn95 DHFR. These data extend previous observations showing that independent selections of the heterozygous parent cell line DC-3F in MTX result in threefold more MTX-resistant lines overexpressing the Asp95-encoding DHFR allele than the Asn95-encoding DHFR allele.(ABSTRACT TRUNCATED AT 250 WORDS)

Application of the polymerase chain reaction to the ribonuclease protection assay.

We have developed a modified RNase protection assay in which the antisense RNA probe is prepared from a PCR-amplified DNA template rather than from a linearized plasmid DNA template. In this assay, an RNA polymerase promoter sequence is attached to the 5' end of the antisense PCR primer. Using this modified antisense primer in conjunction with the paired sense primer, PCR amplification generates a linear DNA template that includes an RNA polymerase promoter sequence. Transcription in vitro initiated by the incorporated promoter in the presence of RNA polymerase and ribonucleotide triphosphates produces a radiolabeled run-off antisense RNA transcript, which can then be used as probe for RNase protection analysis. Probes generated by this method obviate the need to subclone DNA sequences into transcription vectors for synthesis of antisense transcripts. Due to the simplicity of its design and the lack of need for subcloning, this strategy offers greater flexibility than conventional methods for the production of single-stranded RNA probes, and thus facilitates the implementation of the ribonuclease protection assay.

Functional studies with a full-length P-glycoprotein cDNA encoded by the hamster pgp1 gene.

Hamster cells grown in culture may, like human and mouse cells, develop multidrug resistance (MDR) when exposed to certain cytotoxic chemotherapeutic agents. Several phenotypic features that are characteristic of MDR have been described; these include (1) resistance to many structurally and functionally unrelated drugs that have different cellular targets and modes of action; (2) reversal of MDR by certain agents, including verapamil and cyclosporin A; and (3) reduced intracellular drug accumulation relative to that of drug-sensitive cells. In this report we show that the introduction and overexpression of the hamster pgp1 cDNA confers to otherwise drug-sensitive cells an MDR phenotype with these features. Moreover, pgp1 transfectants showed varying degrees of resistance to anthracycline analogues, indicating that structural analogues of commonly used anticancer agents are capable of circumventing drug resistance conferred by pgp.

Further characterization of the sixth transmembrane domain of Pgp1 by site-directed mutagenesis.

PURPOSE: Several studies have identified amino acid residues located on the hydrophobic side of the helix that forms transmembrane domain 6 (TM6) of the ABC transporter P-glycoprotein (Pgp) as being important for function. The purpose of this study was to determine if alterations to residues on the hydrophilic side could also affect function and to determine the extent to which altering the hydrophobic nature of residues on the hydrophobic side would impair the protein. METHODS: A full-length cDNA encoding wild-type Pgp1 from CHL cells was used as a template for site-directed mutagenesis. Eight different mutations, three on the hydrophilic side and five on the hydrophobic side, were prepared and transfected into drug-sensitive host cells. Wild-type transfectants served as controls. Drug resistance levels, RD50 values for cyclosporin A (CsA) and verapamil, iodoarylazidoprazosin (IAAP) photolabeling and verapamil-stimulated ATPase activity were evaluated. RESULTS: Substitution of any one of three amino acid residues on the hydrophilic side of TM6 disrupted function and led to alterations in drug resistance, CsA sensitivity, IAAP photoaffinity labeling, and in one case verapamil-stimulated ATPase activity. Replacement of a hydrophobic residue on the hydrophobic face of the helix with increasingly hydrophilic side-chains led to functional changes, the extent of which did not correlate with the degree of side-chain hydrophilicity. Finally, while the placement of a proline residue along either face of the helix had varying effects on function, in all cases its presence interfered with verapamil-stimulated ATPase activity. CONCLUSIONS: Taken together these results indicate that both faces of TM6 mediate Pgp1 function and that the expected conformational changes resulting from proline substitutions at a variety of locations within the helix can alter the protein's enzymatic activity.

In vitro translation of a 2.3-kb splicing variant of the hamster pgp1 gene whose presence in transfectants is associated with decreased drug resistance.

PURPOSE: P-glycoprotein (P-gp), a product of the Chinese hamster pgpl gene, confers multidrug resistance to mammalian cells in which it is overexpressed either by transfection or as a result of drug selection. It is encoded by a 4.3-kb mRNA and in its unglycosylated form has a predicted molecular weight of approximately 141 kDa. When a cDNA containing this sequence is transfected into drug-sensitive Chinese hamster lung cells and is expressed under the control of the beta-actin promoter, both the full-length 4.3-kb mRNA and a 2.3-kb transcript are produced. The latter results from a splicing event that utilizes near consensus 5' and 3' splicing signals resident in the full-length mRNA, and it has also been found to be present in cell lines that express the native gene. Therefore, it is a splicing product of pgpl per se. This report is concerned with the biological relevance of this transcript. METHODS: In vitro transcription and translation experiments were used to show that the putative open reading frame of the 2.3-kb transcript encodes a novel 57-kDa protein (p57pgp1) that contains transmembrane domains 9-12 and the C-terminal ATP binding fold of P-gp. To elucidate the function of p57pgp1, expression vectors containing cDNAs representing (1) the 2.3-kb transcript, (2) the full-length 4.3-kb mRNA, and (3) a splice-disabled 4.3-kb transcript in which production of the 2.3-kb transcript is eliminated by an in-frame mutation at the 3' splice site, were constructed and transfected into DC-3F cells. Additional expression vectors in which p57pgp1 represented the N-terminus of a green fluorescent protein fusion construct were also prepared and used for transient expression studies. RESULTS: Overexpression of the 2.3-kb transcript alone did not confer multidrug resistance. Transfectants in which both the 4.3-kb transcript and the 2.3-kb transcript were present, compared with transfectants in which no 2.3-kb transcript was expressed, but in which the level of expression of the 4.3-kb mRNA alone was the same, showed little change in cross-resistance pattern. However, the overall level of resistance in the latter cells was increased by approximately twofold. Hence the presence of the 2.3-kb transcript was associated with a decrease in drug resistance. In vitro transcription and translation experiments and transient expression studies indicate that p57pgp1 can be expressed both in vitro and in vivo. CONCLUSION: These results demonstrate that a splicing variant of pgp1 contains an open reading frame capable of translation in vitro and in vivo and suggest that alterations in splicing may contribute both directly and indirectly to the overall mechanism of pgp1-mediated multidrug resistance in CHL cells.