Cloning of Chinese hamster DNA topoisomerase I cDNA and identification of a single point mutation responsible for camptothecin resistance.

A camptothecin-resistant (DC3F/C-10) Chinese hamster cell line that contains a catalytically altered and camptothecin (CPT)-resistant DNA topoisomerase I (top 1) (Tanizawa, A., and Pommier, Y. (1992) Cancer Res. 52, 1848-1854) and the parent cell line (DC3F) were used to compare top 1 mRNAs and cDNAs. Northern blot analysis showed a single 4.1-kilobase band without quantitative reduction between the two cell lines. We have cloned and sequenced top 1 cDNAs. DC3F and DC3F/C-10 top 1 c-DNA are 3591 and 3626 base pair long, respectively, and encode 767 amino acids. The homology of deduced amino acid sequences between Chinese hamster and mouse or human top 1 are 98.1 and 96.7, respectively. cDNAs from DC3F/C-10 and DC3F cells differ by a single base point mutation (G to A) which results in an amino acid change from Gly505 to Ser (Gly505-->Ser). G505 corresponds to Gly503 of human top 1 cDNA and is located 220 amino acids away from the presumed catalytic Tyr725. The point mutation in the Chinese hamster top 1 is located in a region that is highly conserved among all cloned top 1 cDNAs (plant ATH, vaccinia virus, Shope fibroma virus, Drosophila, Saccharomyces cerevisiae, Schizosaccharomyces pombe, mouse, and Human). A mutation of Asp533 to Gly in this same region has been shown to confer CPT resistance for human top 1. Chinese hamster top 1 protein with a Gly505-->Ser mutation that was expressed in bacteria was resistant to CPT, indicating that this single base mutation is involved in CPT resistance. Our results suggest that the highly conserved region around Gly505 plays an important role in the interactions among top 1, DNA, and CPT.

IT TO whom all correspondence should be addressed: Laboratory of inhibitor (1)(2)(3)(4), and its derivatives are among the most promising anti-cancer agents currently in clinical trials ( 5 ) . Despite the experimental evidence that CPT binds to top 1 .DNA complexes rather than to purified DNA or top 1 enzyme (6), the molecular interactions between the drug and its presumed receptor site remain hypothetical (4). DNA interaction is suggested by CPT's strong preference for guanine at the 5' terminus of the stabilized top 1-induced DNA breaks (7), camptothecin-induced DNA photodamage experiments (8), and protein interaction by photolabeling experiments (9).
Based on the findings obtained with several CPT-resistant cell lines, a decrease in the amount of top 1 protein seems to be the main mechanism for drug resistance (10). However, several other CPT-resistant cell lines exhibit either chromosomal rearrangement (11)(12)(13)(14) or top 1 point mutation (15)(16). Formation of top 1-mediated cleavable complex and replication fork arrest have been reported as the initial steps in the cascade leading to drug-induced cell death (17)(18)(19).
In an attempt to characterize the top 1 regions involved in CPT activity and to sequence for the first time the Chinese hamster top 1 enzyme, we have isolated the top 1 cDNAs from wild type (DC3F) and CPT-resistant (DC3F/C-10) Chinese hamster lung fibroblasts (20). We have reported previously that the amount of top 1 enzyme from DC3F/C-10 cells is increased rather than reduced and that its molecular size is similar (approximately 100 and 68 kDa) when compared with top 1 from parental DC3F cells. Furthermore, top 1 from DC3F/C-10 cells was CPT-resistant and showed reduced catalytic activity with approximately &fold less specific activity than that from parental cells (20).

MATERIALS AND METHODS
Cell Culture and Bacteria-The Chinese hamster lung fibroblast DC3F cell line and its CPT-resistant subline (DC3F/C-10) were grown in minimum essential medium with Earle's salt, supplemented with 10% of heat-inactivated fetal bovine serum, 2 mM glutamine, 100 units/ml penicillin, 100 Fg/ml streptomycin, 0.1 mM nonessential amino acids, and 1 mM sodium pyruvate (ABI, Columbia, MD) as described previously (20).
Construction and Screening of cDNA Libraries-Total cellular RNAs of both DC3F and DC3F/C-10 cells were extracted using the The abbreviations used are: CPT, camptothecin; top 1, DNA topoisomerase I; kb, kilobase(s); IPTG, isopropyl-1-thio-P-D-galactopyranoside; bp, base pair(s); PCR, polymerase chain reaction. Top 1 cDNA and CPT Resistance guanidinium thiocyanate-phenol-chloroform extraction method (21). Poly(A)+ RNA was purified with oligo(dT)-cellulose columns. The first cDNA libraries were constructed with the ZAP-cDNA synthesis kit using oligo(dT) primer (Stratagene, La Jolla, CA). Chinese hamster top 1 cDNA was screened with human top 1 cDNA (TIB) which was kindly provided by Dr. Earnshaw (Johns Hopkins University) (22). Since the longest clone was approximately 2.5 kilobase (kb) in the case of both DC3F (clone WTF8) and DC3F/C-10 (clone TAF4) cells (see Fig. 2), we constructed a second set of cDNA libraries using a specific primer for DC3F top 1 cDNA (5"TTTCGGATCTTGTCC AC-3', from position 1465 to 1449 (see Fig. 3) to obtain the 5'-end of top 1 cDNA. The second set of libraries from DC3F and DC3F/C-10 cells was screened with the EcoRIINdeI fragment of TIB (0.9 kb, from position 2 to 901). The longest clones from DC3F and DC3F/ C-10 cells were designated as WTSl2 and TAS17, respectively. Although TAS17 included the open reading frame, WTSl2 did not and was 319 bp shorter than TAS17. The 5'-end of DC3F top 1 cDNA (289 bp) was obtained and sequenced using PCR technique as described under "Results" (WTP2).
Blot Hybridization and Sequence Determination-Blot hybridization was performed according to standard procedures. Blots were hybridized with Chinese hamster top 1 cDNA (1.7-kb fragment of WTF8, nucleotide 1061-2801) or with a @-actin oligonucleotide probe (Oncogene Science, Inc., Manhasset, NY), which had been labeled with [32P]dCTP (Du Pont NEN) by random priming procedure, at 42 "C in the presence of 50% formamide and washed at the same temperature with a solution of 1 X SSC (0.15 M NaC1,0.015 M sodium citrate, pH 7.2), 0.1% sodium dodecyl sulfate (SDS), and 0.25 x SSC, 0.1% SDS. Signal intensities were compared using a Betascope 603 blot analyzer (Betagen Inc.). The cDNA sequences were determined by the dideoxynucleotide chain-termination method using Taq DNA polymerase (Stratagene, La Jolla, CA) and either 35Sor [33P]dATP (DuPont NEN).
Construction of Protein Expression Plnsmids for Top I-A recombinant plasmid containing the full-length top 1 cDNA from DC3F/ C-10 cells was first constructed using TAS17 and TAF4. Thereafter, a truncated top 1 cDNA containing the entire amino acids coding region (position 107 to the end) was inserted into pTrcHis A (PE/ C1) (Invitrogen, San Diego, CA). The recombinant plasmid was transfected into TOPlOF'. As for the expression of top 1 from DC3F cells, Bsu36 I fragment of PE/CI (position from 1256 to 2499), which contained the mutation site, was replaced with the counterpart of WTF8 (PE/WI).
Expression of Recombinant Topoisomerase I-Exponentially growing bacteria (ODm: 0.7) were treated with l mM isopropyl-l-thio-@-D-galactopyranoside (IPTG) (Stratagene, La Jolla, CA) for 1 h to express top 1. The bacteria in 50 ml of culture medium were spun down, resuspended in 5 ml of lysis solution (20 mM phosphate, 500 mM NaCl, pH 7.8), and treated with egg white lysozyme (100 pg/ml) (Boehringer Mannheim) for 15 min on ice. The cells were disrupted by three cycles of sonication and freeze-thawing. The cell lysate was treated with 5 pg/ml RNase (Sigma) for 15 min on ice. After centrifugation at 3000 X g for 15 min, the supernatant was used for catalytic assay.
DNA Oligonucleotide Cleavage Assays-The following doublestranded oligonucleotide was used.

SEQUENCE I
This oligonucleotide is derived from the Tetrahymena ribosomal DNA sequence containing a strong top 1 cleavage site (A) (24), in which the +1 base has been mutated to a G (italic letters) to enhance CPT cleavage (4, 7). 3'-End labeling of the upper strand was performed by using terminal deoxynucleotidyl transferase and 32Plabeled cordycepin (*A) (Du Pont NEN) as described previously (7). Reactions were performed in 10 pl of reaction buffer (0.01 M Tris-HC1, pH 7.5, 50 mM KCI, 0.1 mM EDTA, 15 pg/ml bovine serum albumin) for 30 min at 20 "C and were stopped by adding sodium dodecyl sulfate (0.5% final concentration) and 40 pl of loading buffer (80% formamide, 10 mM NaOH, 1 mM EDTA, 0.1% xylene cyanol, and 0.1% bromphenol blue). Four microliters were loaded into 16% acrylamide/urea sequencing gels. At the end of electrophoresis, gels were transferred to Whatman 3" paper sheets, dried, and analyzed using a PhosphorImager (Molecular Dynamics, Sunnyvale, CA). The full-length end-labeled oligonucleotide (33-mer, including cordycepin) was converted to a 19-mer after camptothecin-induced cleavage (caret between T and G on the upper strand shown above).

RESULTS
Northern Blot Analysis of Total RNA for Top 1 Message-Total cellular RNAs from DC3F and DC3F/C-10 cells which had been transferred onto nitrocellulose filters were examined as described under "Materials and Methods" (Fig. lA). The same blot was re-hybridized with a 40-base human @-actin oligonucleotide probe as an internal control. As shown in Fig. lA, the top 1 probe detected a 4.1-kb signal, and there was no size difference between top 1 messages from DC3F and DC3F/ C-10 cells. Comparison of signal intensities revealed that ratios of top 1 to actin message in DC3F/C-10 RNA were approximately 1.8-fold greater than those in DC3F cells. These results are consistent with our previous Western blot analysis showing that the amount of top 1 protein from DC3F/ C-10 cells was greater than that from DC3F cells (20).
Southern Blot Analysis for Genomic Top 1 DNA-Genomic DNA was extracted from DC3F and DC3F/C-10 cells using the salt extraction method (25). DNA (10 pg) was digested with the indicated enzymes ( Fig. lB), separated on 0.7% agarose gels, and transferred to nitrocellulose membranes. The membranes were hybridized with Chinese hamster partial top 1 cDNA. No difference was detected between restriction maps of the top 1 genes from DC3F and DC3F/C-10 cells. This suggests that there is no chromosomal rearrangement of Hamster Cells-As shown in Fig. 2, we first isolated partial A) which results in an amino acid change from Gly505 to Ser top 1 cDNA clones from a first set of libraries prepared from (Fig. 4). G l P 5 corresponds to Gly503 coded for in human top DC3F and DC3F/C-10 cells. The clones were approximately 1 cDNA (Figs. 2 and 7). Also, Top 1 cDNA from DC3F/C-10 2.5 kb in length starting from the poly(A)+ tail. We re-cells has 30 additional bases at the 5'-end (5"GTCGCCconstructed and screened the second set of libraries as de-GTCTGCTCGTGCCTCTCGGGAACT-3') and 5 additional scribed under "Materials and Methods." An additional 1477bases at the 3'-end (5'-AACGG-3') when compared with that bp clone, which contains an open reading frame (TAS171, and from DC3F cells. However, these changes did not affect the a 1126-bp clone (WTS12) were obtained from the second set amino acid sequence. Taken together, our results show that of libraries from DC3F/C-10 and DC3F cells, respectively. CPT resistance in DC3F/C-10 cells is associated with a single TAS17 and WTSl2 lacked 18 and 50 bases at their 3'-ends amino acid change, Gly505 to Ser.
which contain complementary sequences to the primer. TO Catalytic Activity of Expressed Top 1 in Bacteria-Once obtain base sequences of the missing amino terminus of top IPTG was added to culture medium, ODm values showed a 1 cDNA from DC3F cells, a 15-mer oligonucleotide (5'transient increase for 1 h, but thereafter decreased down to TGCCTCTCGGGAACT-3') which corresponds to base sethe initial value (data not shown). As negative control, quences of TAS17 (from position 16 to 30) was used as pTrcHisA with top 1 cDNA from DC3F/C-10 cells in the upstream primer for PCR reaction. After removal of external reverse orientation was transfected to TOPlOF' and analyzed primers and unincorporated dNTPs, PCR products were diin parallel (PE/Clrev). As reported previously (23), expresrectly sequenced with several internal primers ( WTP2 segsion of top 1 protein was toxic to host bacteria. ment in Fig. 2).
Lysates tire top 1 CDNAs from DC3F and DC3F/C-10 cells were catalytic activity was resistant to CPT up to 100 pM (Fig. 5, sequenced. Fig. 3 shows the nucleotide and deduced amino lower panel). In addition, the catalytic activity from the PE/ acid Sequences of the top 1 CDNA from DC3F cells. TOP 1 C1 plasmid was weaker than that from the wild type enzyme CDNA from DC3F cells is 3591 bp long and encodes 767 amino (3-5-fold). The lack ofmagnesium-independent DNA relaxing acids. The Chinese hamster cDNA contains 45 activity of the cell lysate from E. coli transformed by the PE/ bases from the EcoRI recognition sequence, which corre-Clrev plasmid indicated that the observed catalytic activity SPonds to the 3"end of human CDNA. As occurs in the mouse, in the cell lysates from PE/W1 and PE/C1 cultures was due Chinese hamster top has amino acids not to expressed Chinese hamster top 1 proteins (data not shown). present in the human* The and deduced amino Western blot analysis of the cell lysates was carried out using acid homologies between Chinese hamster and human are 86 human scleroderma SeNm which was kindly provided by D~. and 97%, respectively.
Earnshaw (Johns Hopkins University). When compared with cell lysate from bacteria carrying PE/Clrev, several additional

A C G T A C G T
Top 1 cDNA and CPT Resistance mutation is responsible for CPT resistance. As shown in Fig. 2 Arrows and S indicate negatively supercoiled SV40 DNA.
by continuous exposure to CPT-11 (26). The deduced amino acid sequence for CPT-K5 top 1 cDNA indicated two substitutions (Asp533 to Gly and Aspm to Gly) (26). However, sitedirected mutagenesis analysis revealed that only Asp533 + Gly is responsible for CPT resistance (27). The other CPT-resistant cell line is the non-small cell lung cancer cell line PC-7/ CPT which shows a single amino acid change of Thr7" to Ala (28). The three mutations found to date can be viewed in terms of their location. While Thr729 + Ala is very close to catalytic TYT~'~, Asp533 + Gly and GlYo5 + Ser (which corresponds to Gly503 in human top 1 (Fig. 7) are relatively far from the catalytic tyrosine but close to each other. As shown in Fig. 7, both of the mutations Asp533 -+ Gly and Gly"' * Ser are located in a highly conserved region among all the sequenced top 1 genes (human (22), mouse (29), Chinese hamster (this paper), vaccinia virus (30), Shope fibroma virus (31), S. cereuisiae (32), S. pombe (33), Drosophila (34), plant (Arabidopsis thulianu) (GenBank accession number X57544).
This similarity suggests that this highly conserved region probably plays an important role for the interaction of top 1 with its DNA substrate.
Morham and Shuman (35) reported extensive work on seven mutants of vaccinia top 1. Four mutants had changes located within the conserved region shown in Fig. 7 (GlyI3' + Ser, GlyI3' + Asp, ThrI4* + Ile, and Thr'47 + Ile). Mutant proteins with SerI3' or Ile147 exhibited substantial reduction of DNA relaxation activity. The Asp132 mutant was inert with respect to DNA relaxation, whereas the Ile14' mutation had less effect on activity than the other mutations. These four mutants are within the conserved region involved in CPT resistance for DC3F/C-10 and CPT-K5 top 1. The reduced specific activity of mutant top 1 from DC3F/C-10 cells is consistent with the mutant vaccinia top 1 data, indicating that the conserved noncatalytic region is critical for catalytic function. These results suggest that the top 1 catalytic site may consist of several protein domains, including the catalytic tyrosine region and the conserved region shown in Fig. 7. Hence, CPT resistance of mutant top 1 from DC3F/C-10 cells could be due to an altered secondary structure and protein conformation change which could affect the interaction of top 1 with the top 1. DNA complex.
In the case of the human top 1 mutant from CPT-K5 cells, the authors speculated that the specific activity of mutant top 1 is similar to that of wild type top 1, based on the results of Western blot analysis and total top 1 activity in the crude extracts (27). It is possible that amino acid substitution from Asp533 to Gly in human top 1 has a different effect than the other noncatalytic mutations listed in Fig. 7, even though they are relatively close each other. The predicted conformation change of mutant top l in CPT-K5 suggested that the mutated site becomes recessed from the surface in a protruding region of the molecule (28). We analyzed the possible conformational alteration of mutant top 1 due to the GlYo5 + Ser mutation, using the Gene Works software program (IntelliGenetics, Mountain View, CA) and found some differences between CPT-sensitive and -resistant top 1 in terms of Garnier protein structure predictions LY and 0 (36). However, it is not clear whether these differences are relevant to CPT resistance and reduced specific activity. It is possible that the glycine to serine substitution reduces the flexibility around the wild type Gly506 residue.
Amino acids surrounding the catalytic tyrosine compose another region where mutations can alter specific activity and CPT sensitivity, for example amino acid substitution from Thr7" to Ala in PC-7/CPT cells (15). Replacement of the vaccinia virus top 1 active tyrosine site region (SKRAY) with the counterpart of yeast top 1 (SKINY) does not make vaccinia top 1 sensitive to CPT, whereas the mutated yeast top 1 which has SKRAY instead of SKINY becomes resistant to CPT (35). The authors speculate that in addition to the region mentioned above, there is another critical region for CPT sensitivity where there is no homology between yeast and vaccinia virus top 1. One possible site could be Aspzz1 in vaccinia virus top 1, because the Aspzz1 + Val mutant becomes sensitive to CPT (37). Aspzz1 corresponds to Val5= in human top 1, which is conserved among all the known top 1.
Another possible approach to analyzing top 1 mutant is to compare the effects of CPT derivatives on mutant top 1. Structure-activity studies of CPT derivatives are consistent with the existence of a stereospecific drug binding pocket (3,4,38,39).
CPT-resistant top 1 might still be sensitive to other CPT analogues. Such an analysis may provide useful information on the molecular mechanism of drug-induced inhibition of top 1 activity and on ways to overcome such resistance in cancer chemotherapy.