(Received for publication, November 22, 1996, and in revised form, February 24, 1997)
From the Department of Biochemistry and Molecular Pharmacology, Thomas Jefferson University, Philadelphia, Pennsylvania 19107
The antineoplastic alkaloid camptothecin interferes with the catalytic cycle of DNA topoisomerase I rendering it a cellular poison. Camptothecin stabilizes a covalent enzyme-DNA intermediate that is converted into lethal double strand DNA lesions during S phase of the cell cycle. Yeast SCT1 mutants were isolated in a screen for mutations in genes other than TOP1 that result in camptothecin resistance. Here we report SCT1 is allelic to PDR1 and that a Thr-879 to Met substitution in the PDR1-101 transcription factor confers multiple drug resistance. PDR1 regulates the expression of several gene products including the ATP-binding cassette transmembrane transport proteins PDR5, YOR1, and SNQ2. The PDR1 T879M mutant increased PDR5 transcription compared with wild-type PDR1 strains. Deletion of PDR1 or the downstream effector SNQ2 increased cell sensitivity to camptothecin, whereas deletion of YOR1 or PDR5 had little effect on camptothecin sensitivity. However, the camptothecin resistance accompanying GAL1-promoted overexpression of PDR5 suggests some substrate promiscuity among the ATP-binding cassette transporters. These data underscore the role of the pleiotropic drug resistance network in regulating camptothecin toxicity and are consistent with a model of decreased intracellular concentrations of camptothecin resulting from the increased expression of the SNQ2 transporter.
Eukaryotic DNA topoisomerase I is a highly conserved enzyme that
catalyzes changes in DNA topology during such processes as transcription and DNA replication (reviewed in Refs. 1-4). The enzyme
transiently nicks and religates a single strand of duplex DNA,
resulting in the relaxation of positively or negatively supercoiled DNA. This reaction is characterized by the formation of a covalent linkage between the active site tyrosine and the 3-phosphate of the
cleaved DNA strand, which conserves the energy of the broken phosphodiester bond in the DNA backbone. Thus, an exogenous energy source is not required for completion of the catalytic cycle.
Camptothecin is a potent anti-tumor drug that specifically targets DNA
topoisomerase I by reversibly stabilizing a covalent enzyme-DNA
intermediate (reviewed in Refs. 2, 4-7). In the yeast
Saccharomyces cerevisiae, DNA topoisomerase I has been shown to constitute the cellular target of the drug (8-14). Yeast strains deleted for TOP1
(top1)1 are
camptothecin-resistant; however, the cytotoxic response of these
top1
strains to camptothecin is restored by the
expression of yeast or human TOP1 from plasmid-borne
sequences (8, 12, 13). Formation of the ternary drug-enzyme-DNA
complexes alone is insufficient to cause cell death. Studies with DNA
synthesis inhibitors suggest that the collision of DNA replication
forks with the drug-stabilized covalent intermediates produces
double-stranded DNA breaks, which in turn result in cell cycle arrest
in G2 phase and cell death (reviewed in Refs. 2 and 5).
Other mammalian cell responses to camptothecin treatment include an
inhibition of DNA synthesis (15, 16) and the increased expression of DNA damage responsive genes, such as p53 and p21 (17, 18). The
increased expression of the yeast DNA damage-inducible genes DIN3 and RNR3 in response to camptothecin
treatment and the camptothecin hypersensitivity of yeast cells
defective in DNA double strand break repair, due to deletion of
RAD52, suggest a conserved mechanism of drug-induced DNA
damage and repair (reviewed in Refs. 9 and 11).
Camptothecin resistance in mammalian cells typically results from a decrease in DNA topoisomerase I activity or the expression of TOP1 mutants that produce a camptothecin-resistant enzyme (10, 19-21); although, two recent reports of camptothecin-resistant cell lines are intriguing because they suggest alternative mechanisms of drug resistance (22, 23). Mammalian cells overexpressing the ATP-binding cassette (ABC) transporter MDR1 acquire resistance to several important anticancer drugs including taxol, doxorubicin, amsacrine, and etoposide (24) but are not resistant to camptothecin or the analogues 10-hydroxycamptothecin and 10,11 methylenedioxycamptothecin (25-27). MDR1 overexpression confers an approximate 10-fold resistance to topotecan, SN-38, and 9-aminocamptothecin in cultured cells; however, no difference in tumor clearing is evident in animal models (25).
We recently developed a yeast genetic screen to identify mutations in genes other than TOP1 that suppress the cytotoxic action of camptothecin (14). We reported the isolation of several dominant SCT1 mutants that abrogated the cytotoxic action of camptothecin and the terminal G2-arrested phenotype, although the cells expressed high levels of wild-type DNA topoisomerase I. However, these SCT1 mutants were unable to suppress the camptothecin sensitivity accompanying human DNA topoisomerase I expression and remained sensitive to UV and x-ray-induced DNA damage.
In this study, we report that the camptothecin resistance of the SCT1 mutants results from a dominant mutation in PDR1, which encodes a transcription factor that regulates the expression of a large family of gene products responsible for pleiotropic drug resistance in yeast (reviewed in Ref. 28). Downstream targets of PDR1 include the ABC transporters PDR5, YOR1, and SNQ2 (28-30) that belong to the same transporter superfamily as mammalian MDR1. Our results show that camptothecin sensitivity in yeast is also modulated by the PDR1 regulatory network, primarily through SNQ2. Although the related PDR5 transporter was not required for camptothecin resistance, overexpression of PDR5 from the GAL1 promoter was sufficient to confer camptothecin resistance in wild-type PDR1 strains. The promiscuous nature of pleiotropic drug resistance in yeast suggests that the increased expression of other, as yet, undefined ABC transporters might also affect the activity of camptothecin analogues in human tumor cells.
Camptothecin, cycloheximide, and oligomycin were
obtained from Sigma. Camptothecin was dissolved in dimethyl sulfoxide
(Me2SO) at 4 mg/ml. Cycloheximide and oligomycin were
dissolved at a final concentration of 1 mg/ml in water and ethanol,
respectively. Aliquots of each were stored at 20 °C.
The plasmids and yeast strains used in these studies are described in Tables I and II, respectively. Except as noted, all strains were derived from parent strains FY250 and FY251. Due to differences in DNA sequence, the wild-type PDR1 allele in plasmid pSK-PDR1I, isolated from strain IL125-2B,2 is designated PDR1I and that isolated from the parent strain of EKY3 (FY250) is designated PDR1E. The PDR1E allele was isolated by the gap repair method of Rothstein (31). First, the ApaI site in the polylinker of plasmid YCpPDR1-101-T was excised by digesting with KpnI and SalI, repairing the ends with T4 DNA polymerase and ligating the blunt ends. The plasmid was then digested with ApaI and SacII, and the 6350-kb plasmid backbone with flanking PDR1 sequence was transformed into FY250 by the lithium acetate procedure (32). Following selection for tryptophan prototrophs, the YCpPDR1E plasmid was recovered by vortexing the cells in the presence of phenol and glass beads (33). The identity of the gap-repaired PDR1E and PDR5 sequences (in plasmid YCpPDR5, Table I) was confirmed by detailed restriction enzyme mapping and sequence analysis of the purified plasmid DNA.
|
|
Plasmid pKO-pdr1::URA3, which contains
pdr11::URA3, was made by digesting
plasmid YCpPDR1-101-T with BclI, repairing the ends with T4
DNA polymerase, and purifying the 6.8-kb DNA backbone. This was ligated
to a 1.6-kb BamHI fragment with repaired ends, which was
excised from plasmid YDpU (34) and contains the URA3 gene.
The resulting plasmid, with URA3 inserted in the opposite orientation of the PDR1 promoter, was digested with
SacI and KpnI, and a 3-kb fragment containing
pdr1[/delta]1::URA3 was used for single step gene replacement (31). Deletion of PDR1 in the
resulting uracil prototrophs was confirmed by screening for increased
cycloheximide sensitivity and by PCR using primers
5
-GAGACAATGGTTGACCTTTTGTCC and 5
-ACGCCCTTATACTGTCCAGTTGAG that flank
the integration site.
Plasmid pKO-pdr5::URA3 was constructed in several steps.
First, the PDR5 gene was PCR-amplified from FY250 genomic
DNA using primers 5-CGTGTATTTCGCTGCTTAGGAAAG and
5
-TGAGAAGACGGTTCGCCATTC and the GeneAmp XL kit (Perkin-Elmer). The
resulting 6-kb fragment was treated with T4 DNA polymerase and T4
polynucleotide kinase, ligated to ApaI DNA linkers as per
the manufacturer's instructions (New England Biolabs), and ligated
into the ApaI site of pBluescriptII-KS(+) to give plasmid
pBlue-PDR5. The coding and 5
-noncoding sequences of PDR5,
between internal BglII and MscI sites, were
excised. Blunt ends, prepared by treatment with T4 DNA polymerase, were ligated to the 1.6-kb URA3 fragment described above. The
pdr5
1::URA3 sequences were then
excised from pKO-pdr5::URA3 in a 3-kb SacI and
KpnI fragment and used to generate the
pdr5
1 strains indicated in Table II. The URA+
transformants were examined for increased cycloheximide sensitivity,
and integrations were confirmed by PCR using primers 5
TGCAAAGGAATGGCAACTGAG and 5
GCCGTATATGAGAAGACGGTTCG that flank the
integration site.
To identify the dominant mutation conferring camptothecin resistance in SCT1 mutant strains, a genomic DNA library was constructed from SCT1-1 strain EKY7 (Table II) using a modified version of that described by Rose and Broach (35). Briefly, EKY7 cells were grown in YPD, converted to spheroplasts, and lysed as described. The extracts were then sequentially treated with RNase and proteinase K. Following extraction with chloroform/isoamyl alcohol (24:1), the DNA was fractionated by CsCl gradient centrifugation and collected in 1-ml fractions. 10-µl aliquots were analyzed in 0.6% agarose gels, and the most viscous fractions containing high molecular weight DNA were pooled, dialyzed against TE buffer (0.1 M Tris, pH 8.0, 10 mM Na2EDTA), ethanol-precipitated, and resuspended in TE buffer. 25 µg of genomic DNA was partially digested with BamHI, under conditions that maximized the yield of DNA fragments in the 8-12-kb size range. Digestion was terminated by the addition of a final 25 mM Na2EDTA and 0.2 M NaCl. The DNA fragments were then pooled, phenol-extracted, ethanol-precipitated, and resuspended in TE at a final concentration of 0.3 mg/ml.
Approximately 10 µg of YEp24-PL DNA (Table I) was digested to completion with BamHI and dephosphorylated by incubation with calf intestinal alkaline phosphatase (1 unit of calf intestinal alkaline phosphatase/pmol of DNA ends). After phenol extraction and ethanol precipitation the DNA was resuspended in TE, pH 7.5. The DNA library was constructed by ligating 5 µg of the partially digested genomic DNA with equimolar amounts of the dephosphorylated YEp24-PL DNA. The multicopy genomic DNA library was amplified using SURE cells (Stratagene) and CsCl-purified (36). Based on restriction enzyme digestion of representative plasmid DNAs obtained from single bacterial transformants, the average insert size was approximately 6 kb, and 50% of the vectors contained insert DNA.
The genomic library DNA was electroporated into EKY5 cells, in which pGPD-TOP1 was integrated into the chromosome at the TOP1 locus (Table II), using a Bio-Rad Gene Pulser (37). The constitutive high levels of TOP1 expression from the GPD promoter were necessary to score cell sensitivity to camptothecin (14). Following electroporation, the cells were plated on SC-uracil, histidine media containing 1 M sorbitol, 25 mM HEPES, pH 7.2, 10 µg/ml camptothecin, and a final 0.25% Me2SO. The plasmids were recovered from the drug-resistant colonies (33), and the DNA inserts were excised and subcloned into the BamHI site of plasmid pRS416. Plasmid YCp-clone1 (YCpSCT1-1 in Table I) consistently conferred camptothecin resistance upon transformation into TOP1 expressing yeast cells.
Plasmid YIp-clone1 was constructed (Table I), linearized at a unique
Bsu36I site within the insert, and transformed into EKY7 to
target integration to the chromosomal position of the clone 1 sequences. The resulting strain, RRY9-1 (Table II), was mated with EKY1
and sporulated. The meiotic products of 26 tetrads were grown,
transformed with YCpGPD-TOP1, and examined for co-segregation of
camptothecin resistance and uracil prototrophy. SCT1-1 was then localized to chromosome VII by hybridizing blots containing the
entire yeast genome in ordered, overlapping arrays of lambda clones and
cosmids (obtained from ATCC) (38) with a 32P-labeled probe
prepared from the 6.4-kb clone 1 insert using the Megaprime kit
(Amersham Corp.). This was confirmed by aligning restriction maps of
clone 1 DNA with those derived from the published sequence of the
centromere-proximal left arm of chromosome VII (39). The C-terminal 800 bp of the PDR1I, PDR1E, and
PDR1-101 alleles were sequenced using primers F1
(5-CCAACCTCGATGTTATCTCC), F2 (5
-CGCAACCAACAAATGGG), R1
(5
-GAGAAGGAGATCGCCCTAG), and R2 (5
-GGACTTGCGAATTATTTTGCCC) on an
Applied Biosystems Sequencer by cycle sequencing.
RRY4-3 and RRY5-3 cells (Table
II), transformed with plasmid YCpGPD-TOP1-L, were grown in SC-leucine
media supplemented with 25 mM HEPES, pH 7.2. At an
A595 = 0.3, the cultures were treated with 100 µM camptothecin or Me2SO alone. At the times
indicated, aliquots were taken, and -galactosidase activity was
assayed in crude cell extracts as described (40). Briefly, the cells were concentrated and lysed by vortexing with glass beads, and extracts
were clarified by centrifugation. Cleavage of o-nitrophenyl
-D-galactopyranoside was measured by absorbance at
420 nm, and
-galactosidase activity was calculated as
A420 × assay volume/0.0045 × total
protein concentration × extract volume × time. Protein concentrations were determined with the Bio-Rad dye reagent as per the
manufacturer's recommendations.
To assay yeast cell sensitivity to cycloheximide, oligomycin, or camptothecin, overnight cultures were serially 10-fold diluted, and 5-µl aliquots were spotted onto the appropriate selective agar plates containing the indicated drug concentrations. Oligomycin sensitivity was assessed on YP glycerol plates, and plates containing camptothecin were buffered with 25 mM HEPES, pH 7.2, and contained a final 0.125% Me2SO as described (41, 42).
Hybridization AnalysisFor Northern blots, total yeast RNA was prepared by mechanical shearing of the cells as described previously (43). Briefly, 40 ml of exponentially growing cells in SC-uracil media, or cells induced for 8 h with a final 2% galactose in SC-uracil plus raffinose media, were washed and resuspended in 300 µl of RNA buffer (200 mM Tris, pH 7.0, 500 mM NaCl, and 10 mM EDTA). 300 µl of unbuffered, H2O-saturated phenol/chloroform and 250 µl of acid-washed glass beads were added and the cell suspensions vortexed 10 min at 4 °C. Following the addition of 300 µl of RNA buffer, the cell debris and glass beads were removed by centrifugation. The aqueous phase was extracted two more times with H2O-saturated phenol/chloroform. The RNA was then ethanol-precipitated and resuspended in 50 µl of RNA buffer.
Northern hybridization analysis was performed as described (44) using 10 µg of total RNA electrophoresed in a 1% agarose, 6% formaldehyde gel. Probes were labeled by random priming using the Megaprime kit (Amersham Corp.) and hybridized at 65 °C in buffer containing 5% (w/v) SDS, 0.5 mM sodium phosphate, pH 7.0, and 1% bovine serum albumin (45). The blots were washed repeatedly in 5% SDS, 40 mM sodium phosphate and once in 1% SDS, 40 mM sodium phosphate (45). The bands were visualized by autoradiography and quantitated using a Molecular Dynamics PhosphorImager.
We previously developed a yeast genetic screen to identify the cellular factors that participate in the process of camptothecin-induced cell killing (14). Dominant camptothecin-resistant SCT1 mutants were isolated that express wild-type DNA topoisomerase I, yet suppress the terminal G2-arrested phenotype and cell lethality induced by drug treatment of wild-type strains. In contrast, the camptothecin sensitivity of SCT1 mutants expressing human DNA topoisomerase I is not suppressed at high drug concentrations (14).
To resolve this discrepancy, we first examined the cytotoxic response
of yeast cells expressing the yeast or human enzyme, over a range of
drug concentrations. In experiments with decreasing concentrations of
camptothecin, top1 strains expressing human DNA
topoisomerase I from the GAL1 promoter on a single copy
vector were approximately 50-fold more sensitive than the same strains expressing yeast TOP1 (data not shown). Similar levels of
cell killing were apparent at drug concentrations of 0.1 µg/ml with human TOP1 expressing cells versus 5 µg/ml for
cells expressing the yeast enzyme. However, less than a 10-fold
increment in the camptothecin sensitivity of the human enzyme was
detected in DNA cleavage assays in comparisons of yeast and human DNA
topoisomerase I purified from the same strains (41, 42, and data not
shown). These data suggest that differences in the in vivo
activity of the endogenous and heterologous enzymes in yeast might
render human DNA topoisomerase I a more potent poison in the presence of camptothecin. This is consistent with the limited resistance of
SCT1-1 mutants expressing human TOP1 observed at
lower doses of camptothecin, in comparisons with isogenic wild-type
strains (data not shown).
A reduction in the levels of camptothecin-induced DNA damage in
SCT1-1 cells was evident in assays of DNA damage-inducible gene expression. Yeast cells carrying an integrated bacterial lacZ gene fused to the damage-inducible DIN3 and
RNR3 promoters express -galactosidase in response to DNA
damage induced by UV, methyl methanesulfonate, and camptothecin (12,
46). In the present studies, DIN3-lacZ reporter constructs
were integrated into the genome of SCT1-1 and wild-type
SCT1+ strains to yield strains RRY4-3 and RRY5-3
(Table II). These cells were transformed with an ARS/CEN
vector expressing TOP1 from the constitutive GPD
promoter, YCpGPD-TOP1 (Table I) (14), and treated with
100 µM camptothecin, and the levels of
-galactosidase activity were assayed over time. As shown in Fig. 1,
SCT1-1 strains expressed approximately 50% of the
-galactosidase activity of isogenic SCT1+
strains over a 9-h period following camptothecin treatment.
-Galactosidase activity in control cultures treated with
Me2SO was negligible. Similar results were obtained using
the damage-inducible RNR3-lacZ fusions (data not shown).
Consistent with these data, the camptothecin sensitivity of
SCT1-1 cells expressing yeast TOP1 was restored when the cells' ability to repair double-stranded DNA breaks was impaired by disrupting the RAD52 gene (data not shown).
rad52
strains are deficient in the recombinational repair
of double-stranded DNA breaks and, as such, are hypersensitive to low
doses of camptothecin (8, 12). Thus, drug-induced DNA damage does occur
in the SCT1-1 mutant strain, albeit at levels tolerated in
the presence of functional RAD52.
Cloning and Characterization of PDR1-101 from SCT1-1 Cells
To
identify the dominant mutation responsible for the camptothecin
resistance of the SCT1 mutants, a 2-µm vector based DNA library was constructed with a partial BamHI digest of
genomic DNA purified from strain EKY7 (Table II). To
simplify screening for camptothecin-resistant transformants, the
library DNA was electroporated into the wild-type
SCT1+ strain, EKY5, which also contains a
GPD-promoted TOP1 construct integrated into the
chromosome. The constitutive expression of TOP1 in these
cells results in a 104-fold drop in cell viability in the
presence of camptothecin, which obviates the need for a second plasmid
expressing DNA topoisomerase I. Over 50,000 transformants were screened
for camptothecin resistance, by replica plating onto media containing
drug or Me2SO alone. One vector, YEp-clone1 (Table I),
contained a 6.4-kb DNA fragment (Fig. 2A)
that also conferred camptothecin resistance when the genomic DNA
insert was subcloned into an ARS/CEN vector (YCp-clone1) (Fig.
3).
The genetic identity of clone 1 as SCT1 was confirmed by integrating the DNA into the chromosome of SCT1-1 strain EKY7 with the URA3 marked plasmid YIpSCT1-1 (YIp-clone1, Table I). This strain was mated to wild-type SCT1+ strain EKY1, sporulated, and the spores tested for camptothecin sensitivity and for growth in the absence of uracil. The integrated URA3 marker co-segregated with the camptothecin-resistant phenotype in 23 of 26 tetrads examined, indicating linkage between clone 1 and SCT1-1. Tetratypes were identified as 3:1 segregation of camptothecin resistance in which one camptothecin-resistant haploid was auxotrophic for uracil. A map distance of 5.8 centimorgans between clone 1 and SCT1-1 was due to the presence of the intervening URA3 sequences.
The 6.4-kb genomic DNA was mapped to the centromere-proximal left arm of chromosome VII by probing a panel of yeast lambda clones and cosmids containing the complete yeast genome (ATCC). A comparison of the clone 1 restriction map with the restriction maps derived from published sequences (39) revealed the presence of complete ERG4 and PDR1 genes, as well as portions of SCL1 and open reading frame YGL023 (Fig. 2A). Yeast PDR1 mutants display a semidominant pleiotropic drug-resistant phenotype to such structurally diverse drugs as cycloheximide and oligomycin (28, 47). To determine if PDR1 was responsible for the drug-resistant phenotype, a 1.6-kb SacI fragment was removed from clone 1 to delete over half of the ERG4 gene. The plasmid YCpPDR1-101-T was then cotransformed with YCpGPD-TOP1 into a wild-type yeast strain, and the transformants were tested for drug sensitivity (Fig. 3). The pattern of camptothecin resistance in cells containing clone 1 or PDR1-101 sequences was identical. In comparison with the vector control, the increased copy number of the wild-type PDR1I allele produced only a slight increment in cycloheximide resistance. Thus, the camptothecin-resistant phenotype of SCT1-1 cells results from a dominant mutation in PDR1-101.
Regulation of Camptothecin Sensitivity by PDR1To identify the specific lesion, DNA fragments spanning the PDR1-101 allele were used to replace the corresponding sequences in wild-type PDR1I (obtained from Dr. John Golin, Catholic University) in plasmid YCpPDR1I. As shown in Fig. 2A, plasmids containing the C-terminal end of PDR1-101 conferred camptothecin resistance. In Fig. 2B, a comparison of PDR1-101 and PDR1I sequences spanning this region indicated four differences: a C to T transition at nucleotide 2636 resulting in a Thr-879 to Met substitution, a T to C transition at nucleotide 3218 resulting in a Ile-921 to Thr substitution, a T to A transversion at 3397 resulting in a Ser-981 to Thr substitution and an additional 15 nucleotides inserting five Asn residues after residue Ile-1010. Since the PDR1 locus exhibits strain-specific differences,2 the wild-type PDR1E allele was recovered from the EKY3 progenitor strain, FY250, by gap repair of the plasmid-borne PDR1-101 allele. As with PDR1I, transformation of EKY1 cells with the plasmid-encoded wild-type PDR1E allele produced a slight decrease in cycloheximide sensitivity yet did not confer camptothecin resistance (Fig. 3). While higher levels of PDR1E expression from the GAL1 promoter did enhance the cycloheximide resistance of the cells, there was no effect on camptothecin-induced cytotoxicity (data not shown). Thus, the camptothecin-resistant phenotype was the direct result of mutation(s) in PDR1E. The C-terminal PDR1E and PDR1-101 sequences were identical except for the Thr-879 to Met change (Fig. 2B), which occurs near an asparagine-rich domain that may function in transcriptional activation (48). Replacing the C-terminal PDR1E sequence with sequence from PDR1-101 containing the Thr-879 to Met substitution resulted in a dominant camptothecin-resistant PDR1 allele, confirming that this single point mutation was responsible for the drug-resistant phenotype (data not shown).
The camptothecin-resistant strains EKY7 and EKY8 were also resistant to
cycloheximide and oligomycin (Fig. 4), consistent with a
model of pleiotropic drug resistance mediated by the increased expression of several target genes. Moreover, this effect is not restricted to dominant mutations near the putative transactivation domain in PDR1-101. A previously reported mutation of
residue Pro-298 in PDR1-7 produces a similar pattern of
multi-drug resistance (47). This mutation, in strain JG204, was also
sufficient to render the cells resistant to camptothecin, in comparison
to the isogenic PDR1I strain, JG200 (Fig. 4).
Etoposide is a potent anti-tumor drug that targets eukaryotic DNA
topoisomerase II by stabilizing the covalent enzyme-DNA intermediate
(reviewed in Refs. 7 and 49). Mammalian cells overexpressing the ABC
transporter MDR1 are resistant to a wide variety of
chemotherapeutic agents, including etoposide and taxol (24), but are
not resistant to camptothecin and only weakly resistant to the
water-soluble camptothecin analogue, topotecan (25-27). Yet, despite
the similarities between the downstream targets of PDR1 and the
MDR1 encoded P-glycoprotein, the PDR1-101 allele had no obvious effects on yeast cell sensitivity to etoposide (Fig.
5). Strain JN394t2-4 contains a dominant ISE2
mutation that renders cells permeable to etoposide and a single copy
plasmid carrying the TOP2 gene fused to a constitutive
DED1 promoter to complement the top2-4 ts mutant
(50). Following transformation of strain JN394t2-4 with plasmids
expressing PDR1E or PDR1-101, the cells
were assayed for growth in media containing etoposide. In all cases,
9 h of etoposide treatment produced a 75% reduction in the number
of viable cells, in contrast to the 4-fold increase in the number of
untreated cells. Nevertheless, the PDR1-101 transformants
were resistant to cycloheximide, which is indicative of
PDR1-101 function in this strain background (data not
shown).
To establish that PDR1 function is absolutely required for the
camptothecin resistance of the SCT1 strains, the
PDR1 gene sequences were deleted in wild-type
PDR1E, and mutant PDR1-101,
PDR1-102, and PDR1-104 strains by one-step gene
replacement (31). The latter correspond to the SCT1-1, SCT1-2, and SCT1-4 strains, EKY7, EKY8, and EKY9,
respectively (14). The resultant strains were transformed with plasmid
YCpGPD-TOP1 and assayed for camptothecin sensitivity. As seen in Fig.
6, deletion of the dominant PDR1 alleles not
only abolished the camptothecin-resistant phenotype at 5 µg/ml drug
but actually increased cell sensitivity to camptothecin (compare the
growth of the PDR1E cells with that of the
pdr11 strains at 1 µg/ml camptothecin). This
hypersensitive phenotype of pdr1
1 cells shows
that, even in wild-type yeast strains, the PDR1 network is a
major determinant of camptothecin metabolism.
The Function of PDR5 in PDR1-101-induced Camptothecin Resistance
The PDR5 gene product (PDR5) is a
transmembrane ATPase in the ABC family of transporters (28, 51, 52).
PDR5 is overexpressed in PDR1 mutant strains and is the
principle effector of cycloheximide resistance (28, 51, 53). The
PDR5 gene was disrupted in strains EKY3
(PDR1E) and EKY7 (PDR1-101) to determine
the role of PDR5 in mediating the drug-resistant phenotypes. As shown
in Fig. 7A, the PDR1E,
pdr51 strain and the PDR1-101,
pdr5
1 strain were both hypersensitive to
cycloheximide, in comparison with isogenic PDR5 control
strains. However, deletion of PDR5 had no effect on the
camptothecin sensitivity of PDR1E or
PDR1-101 strains, which suggests that camptothecin is not a
good substrate for the PDR5 transporter.
Surprisingly, elevated levels of PDR5 expression from the
GAL1 promoter did suppress the camptothecin sensitivity of
wild-type PDR1E cells. Plasmid YCpGAL1-PDR5-L or
YCpGAL1-L (as a vector control) was transformed into
PDR1E and PDR1-101 strains and induced by
growth on galactose. In Fig. 7B, overexpression of
PDR5 caused a slight defect in cell growth in the absence of
drug (compare YCpGAL1-PDR5-L transformants with YCpGAL1-L transformants
grown on dextrose and galactose Me2SO plates).
Nevertheless, GAL1-induced PDR5 expression in the
PDR1E strain conferred resistance to both
cycloheximide and camptothecin, at levels comparable to the
drug-resistant PDR1-101 mutant containing the YCpGAL1-L
control. Consistent with these findings, expression of
PDR1-101 produced only a 4-fold increment in PDR5
mRNA levels in comparison with wild-type PDR1I
and PDR1E controls (Fig. 8). On the
other hand, induction of PDR5 expression from the
GAL1 promoter construct resulted in greater than a 16-fold increase in PDR5 mRNA levels, relative to the uninduced
control (Fig. 8). Taken together, these data suggest that while
camptothecin is a poor substrate for PDR5, elevated levels of
PDR5 expression can result in promiscuous transport.
PDR1-101-induced Camptothecin Resistance Requires SNQ2
Two
other recently characterized downstream targets of PDR1,
SNQ2, and YOR1 are required for resistance to
4-nitroquinoline oxide and oligomycin, respectively (28-30). To assess
the function of SNQ2 and YOR1 in suppressing camptothecin-mediated cell
death, SNQ2 and YOR1 were individually disrupted
in PDR1-101 or PDR1E strains (using
constructs kindly provided by W. Scott Moye-Rowley). As with
PDR5, deletion of YOR1 had little effect on the
camptothecin resistance of PDR1-101 cells expressing
wild-type DNA topoisomerase I (Fig. 9). In contrast, the
camptothecin sensitivity of PDR1-101, snq21 cells was close to that of the wild-type
PDR1E control at 5 µg/ml.
PDR1E, snq2
1 cells also
exhibited a hypersensitive phenotype at lower camptothecin
concentrations (Fig. 9). Thus, SNQ2 appears to be the primary effector
of camptothecin resistance in the PDR1-1E, mutant
strains, with minor contributions from other members of the pleiotropic
drug resistance network regulated by PDR1.
Camptothecin is a potent antineoplastic agent that reversibly
stabilizes the covalent intermediate formed between eukaryotic DNA
topoisomerase I and DNA. In drug-treated cells, double strand DNA
breaks accumulate resulting in cell cycle arrest in G2 and cell death (reviewed in Refs. 2, 4, and 5). Experiments in yeast
support a similar mechanism of drug-induced DNA damage during S phase
that leads to a G2-arrested terminal phenotype (9, 11, 14).
As TOP1 is nonessential in yeast (54), this genetically
tractable system has been exploited to define the cellular processes
and structural aspects of DNA topoisomerase I essential for the
cytotoxic action of camptothecin. For example, DNA topoisomerase I
constitutes the sole cellular target of the drug, as cells deleted for
TOP1 are camptothecin-resistant (12, 13). Furthermore, the
mechanism of drug-induced cell killing appears to be conserved in that
drug sensitivity is restored when plasmid-borne yeast or human
TOP1 sequences are expressed in these top1
cells (8, 12, 13). Specific amino acid substitutions have also been
described in the yeast and human DNA topoisomerase I that render the
catalytically active enzymes resistant to the drug (9, 10, 41, 42). In
addition, we recently described the isolation of dominant
SCT1 mutants that suppressed the cytotoxic activity of
camptothecin in yeast, despite the fact that these cells expressed
catalytically active, camptothecin-sensitive DNA topoisomerase I (14).
In this report, we have identified SCT1-1 as a dominant
mutation in PDR1.
PDR1 is the major transcriptional regulator of a network of gene products that respond to cellular toxins (reviewed in Ref. 28). Dominant PDR1 mutations have been reported that confer pleiotropic drug resistance by increasing the expression of target genes that affect drug metabolism (28). These include ABC transporters for which drug or peptide substrates have been defined, including PDR5, YOR1, STE6, and SNQ2 (28-30, 53, 55). Other targets of PDR1 include ABC transporter homologues, whose functions have yet to be determined, membrane-associated proteins, and the glycolytic enzyme glyceraldehyde-3-phosphate dehydrogenase (28). The regulated expression of this diverse set of genes suggests a general response to toxins and stress that includes drug efflux.
The T879M mutation in PDR1-101 occurs near an
asparagine-rich domain that may function similarly to other
asparagine-rich transcriptional activators (48). The activation
domain of the homologous PDR3 protein is also localized to the C
terminus (56). In addition to camptothecin, PDR1-101 cells
exhibited resistance to other structurally unrelated drugs, including
cycloheximide and oligomycin, suggesting the increased expression of
several downstream ABC transporters. Consistent with previous studies (57), the observed cycloheximide resistance of PDR1-101
strains correlated with elevated levels of PDR5 expression
and was suppressed in PDR1-101, pdr5 strains.
However, the camptothecin resistance of these cells was
indistinguishable from isogenic PDR-101, PDR5 strains. Instead, the camptothecin resistance phenotype of
PDR1-101 cells appeared to be mediated primarily by
SNQ2. Deletion of PDR1-101 or SNQ2
sequences not only restored drug sensitivity but actually produced an
increase in camptothecin sensitivity at lower drug doses. Although the
intracellular concentrations of camptothecin were not assayed in these
mutants, these data indicate that PDR1 and SNQ2 also play a role in
mediating camptothecin-induced toxicity in wild-type yeast strains. In
this regard, deletion of PDR1 may prove advantageous when
using yeast to screen for novel therapeutics or to establish the
mechanism of drug action.
Although the PDR1-101 mutant suppressed the cytotoxic
response of TOP1 expressing yeast cells, significant levels
of drug-induced DNA damage were indicated by elevated levels of
damage-inducible DIN3 and RNR3 expression.
Moreover, experiments with rad52 strains suggest that
this level of DNA damage was sufficient to kill cells impaired in their
ability to repair double strand DNA breaks. These results are also
consistent with the inability of PDR1-101 to suppress the
camptothecin sensitivity of yeast cells expressing human DNA
topoisomerase I. In yeast cells, the human enzyme appears to constitute
a more effective target for camptothecin than the yeast enzyme. Thus,
only at significantly reduced concentrations of drug was
PDR1-101 able to reduce the amount of camptothecin-mediated DNA damage to non-lethal levels.
ABC transporter function is conserved among eukaryotes. One mode of multi-drug resistance in mammalian cells involves the overexpression of the MDR1 ABC transporter (gp170). gp170 overexpressing cells are resistant to a wide variety of clinically important drugs including taxol, doxorubicin, etoposide, and amsacrine (24) but show no resistance to camptothecin or the camptothecin analogues 10-hydroxycamptothecin and 10,11-methylenedioxycamptothecin (25-27). gp170 overexpression confers an approximate 10-fold resistance to camptothecin analogues topotecan, SN-38, and 9-aminocamptothecin in culture; however, no difference in tumor clearing is evident in animal models (25). Examples of functional conservation across species include complementation of yeast cells deleted for the a factor transporter, STE6, by the mouse MDR3 gene (58) and the restoration of cadmium resistance to yeast cells lacking the YCF1 transporter by the mammalian CFTR gene (59). The recent report of a mitoxantrone-resistant human breast carcinoma cell line exhibiting cross-resistance to camptothecin, without any discernible alterations in DNA topoisomerase I (22), raises the intriguing possibility that a drug transporter might be involved.
It was surprising to us that the PDR1-101 mutant which results in resistance to multiple drugs including camptothecin did not confer resistance to etoposide, a substrate of MDR1 protein in human cells (24). In our studies, two distinct mechanisms of camptothecin resistance were observed. One resulted from a dominant mutation in PDR1-101 that requires SNQ2 function. The other was induced by the overexpression of PDR5 from the GAL1 promoter in wild-type PDR1E cells. This apparent promiscuity in ABC transporter function further supports the notion that similar mechanisms of camptothecin resistance might be achieved by the increased expression of other as yet unidentified drug transporters in human cells.
We thank John Golin, Thomas McQuire, and W. Scott Moye-Rowley for plasmids, strains, and many helpful discussions; Elisabetta Balzi and Andre Goffeau for communicating results prior to publication, and Joyce Agati and Jolanta Fertala for their excellent assistance.