Proc. Nail. Acad. Sci. USA Vol. 89, pp. 2307-2311, March 1992 Biochemistry

Isolation and characterization of human cDNA clones encoding a high mobility group box protein that recognizes structural distortions to DNA caused by binding of the anticancer agent cisplatin (structure-specific recognition protein/DNA sequence/drug resistance/chromosome mapping)

SUZANNE L. BRUHN*, PIETER M. PIL*, JOHN M. ESSIGMANN*, DAVID E. HOUSMANt, STEPHEN J. LIPPARD*

AND

Departments of *Chemistry and tBiology, Massachusetts Institute of Technology, Cambridge, MA 02139

Contributed by Stephen J. Lippard, November 27, 1991

physical characterization of its DNA adducts (3-7). Proteins that specifically recognize such altered structures could act alone or in concert with other-factors to bind DNA. Gel mobility-shift analyses of mammalian cellular extracts have revealed the existence of proteins that bind specifically to biologically relevant DNA adducts of active antitumor platinum complexes (8-11). Moreover, these proteins can differentiate among DNA structural changes produced by individual cisplatin adducts (9). To characterize such structure-specific recognition proteins (SSRPs) more fully, two related cDNAs were cloned from mammalian cells (12) by using the methodology developed for the isolation of sequence-specific DNA binding proteins such as transcription factors (13). The proteins expressed from these cDNAs bind specifically to adducts ofactive antitumor platinum complexes when assayed by modified Western blotting. Similar binding specificity occurs for native proteins in cell extracts as detected by gel mobility-shift assays (9). In the present article we describe the isolation and sequencet of full-length cDNA clones that encode the previously reported cisplatin-DNA SSRP (12), hereafter SSRP1. Analysis of the predicted amino acid sequence of these clones reveals important homologies to other proteins that recognize altered DNA structures. Included are the high mobility group (HMG) proteins 1 and 2 (14-16). Homology is also observed with the HMG-box domain in human upstream binding factor (hUBF), which activates transcription of RNA polymerase 1 (17). Other recently identified HMG-box proteins include the sexdetermining region Y (SRY) (18, 19); mitochondrial transcription factor 11 (20); lymphoid enhancer binding factor 1 (LEF-1) (21); a T cell-specific transcription factor, TCF-la (22); and the yeast autonomously replicating sequence factor ABF2 (23). While this manuscript was in preparation, the murine homolog of SSRP1 was discovered by workers who screened a mouse cDNA library with variable-(diversity)-joining region [V-(D)J] recombination signal sequence probes (24). This latter finding and the homology to HMG-1 suggest novel mechanisms by which cisplatin and related compounds may function as antitumor drugs.

ABSTRACT Human cDNA clones encoding a structurespecific recognition protein, SSRP1, that binds specifically to DNA modified with cisplatin have been isolated and characterized. The SSRP1 gene maps to human chromosome 11q12. The cDNA clones, obtained by using partial-length cDNAs described previously, predict an 81-kDa protein containing several highly charged domains and a stretch of 75 amino acids 47% identical to a portion of the high mobility group (H1MG) protein HMG1. This HMG box most likely constitutes the structure recognition element for cisplatin-modified DNA, with the probable recognition motif being the local duplex unwinding and bending toward the major groove that occurs upon formation of intrastrand cis-{Pt(NH3)2}2+ d(GpG) and d(ApG) cross-links. Although the DNA recognition properties of members of the HMG-box family of proteins have been characterized with respect to their sequence specificity, the present work demonstrates that proteins with this domain can recognize particular DNA structures as well. The Pt-DNA SSRP described here is the human homolog of a recently identified mouse protein that binds to recombination signal sequences [Shirakata, M., Huppi, K., Usuda, S., Okazaki, K., Yoshida, K. & Sakano, H. (1991)Mol. Cell. Biol. 11, 4528-4536]. These sequences have been postulated to form stem-loop structures, further implicating local bends and unwinding in DNA as a recognition target for HMG-box proteins. Expression analysis in a variety of tissues and cisplatin-resistant cell lines and the inability of cisplatin to induce the message in HeLa cells argue against a direct link between SSRP1 mRNA levels and the response of cells to the drug. The highly effective antitumor drug cisplatin forms covalent adducts with DNA that alter its structure and block replication and transcription (1). Intracellular interactions with these adducts are likely to be of central importance in explaining the toxicity of the drug towards rapidly dividing tumor cells. Accordingly, attention has turned to the study of proteins that could interact in vivo with cisplatin-modified DNA. Understanding such interactions has the potential not only to reveal aspects of the molecular mechanism of the drug but also to delineate underlying principles of the structure-specific recognition of DNA. Structural alterations that occur normally throughout the cell cycle, in DNA packaging and processing, or as the result of environmental damage serve as cues for a variety of biological processes (2). Studies of this type of DNA-protein interaction are facilitated by a well-defined DNA structure at the recognition site. The anticancer drug cisplatin provides an excellent model system for the study of such structure-specific recognition because of the extensive

METHODS Labeling of Probes for Hybridization. The APt2 probe (12) used for hybridization and library screening was radiolabeled Abbreviations: HMG, high mobility group; hUBF, human upstream binding factor; SRY, sex-determining region Y; LEF-1, lymphoid enhancer binding factor 1; TCF-la, T cell-specific transcription factor; V-(D)-J, variable-(diversity)-joining regions; SSRP, structure-specific recognition protein; RSS, recombination signal sequences. fThe sequence reported in this paper has been deposited in the GenBank data base (accession no. M86737).

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.

2307

2308

by random oligonucleotide priming as described (25). Typically, 50-100 ng of DNA in low-melting-point agarose was boiled, primed with pd(N)6 oligonucleotides (Pharmacia), and labeled with [a-32P]dCTP by Escherichia coli DNA polymerase I (Klenow fragment). Labeled fragments were purified by spin dialysis over Sephadex G-50 columns, and the extent of incorporation of radioactivity was monitored by scintillation counting. Library Screening. For the primary screen of each cDNA library, 5 x 106 recombinant phage were plated on E. coli host strain Y1088. Duplicate replica nitrocellulose filters were prepared and then denatured (0.5 M NaOH/1.5 M NaCl), neutralized (1 M Tris, pH 7.4/1.5 M NaCl), and rinsed with 2x SSC (20x SSC = 3 M NaCl, 0.3 M Na3C6H5O7). After baking for 2 hr at 80'C in a vacuum oven, the filters were preincubated at 420C for 4 hr with hybridization fluid [50%o formamide/1 M NaCI/50 mM Tris, pH 7.5/0.5% SDS/10o dextran sulfate/1 mg of denatured salmon sperm DNA per ml/1x Denhardt's solution (0.02% polyvinylpyrrolidone/ 0.02% Ficoll/0.02% bovine serum albumin)]. Probe was then added at a concentration of 1 x 106 cpm of labeled DNA per ml of hybridization fluid, and the incubation was continued for an additional 16 hr. The filters were washed once at room temperature in 2x SSC/0.1% SDS, twice at 650C in 2x SSC/0.1% SDS, and twice at 65°C in 0.1 x SSC/0.1% SDS for 15 min each. The filters were air-dried briefly and analyzed by autoradiography. Multiple rounds of screening were used to isolate plaque-pure bacteriophage clones. Single plaques were amplified in liquid culture for DNA preparation and further analysis. Subcloning. Purified phage DNA was digested with EcoRI to release the cDNA inserts. The EcoRI fragments were isolated from low-melting-point agarose gels by using Geneclean (Bio 101, La Jolla, CA) and ligated into the EcoRI site of plasmid pBluescript SKII(+). After transformation of competent E. coli XL-1 cells, single colonies were isolated and amplified in liquid culture. DNA was purified by using Qiagen affinity chromatography. Sequence Determination and Analysis. Sequence determination was performed on double-stranded plasmid DNA by using the chain-termination method (26) and Sequenase T7 DNA polymerase (United States Biochemical). Sequence analysis used software from the Genetics Computer Group (GCG) at the University of Wisconsin (27). Homology searches were made by using the BLAST (basic local alignment search tool) Network Service at the National Center for Biotechnology Information (28). Northern Analysis. RNA was isolated by using standard procedures (29). Northern analysis was performed as described (12). Inducibility. HeLa cells were grown in a suspension culture of Dulbecco's modified Eagle's medium and 5% (vol/vol) horse serum. The drug concentrations used were 0, 0.05, 0.5, and 1.0 ,ug of cisplatin per ml of medium. Cisplatin was added at time 0, and cells were removed at 0, 6, 12, 24, and 48 hr. Cytoplasmic RNA from each time point studied was isolated and used for Northern analysis. The amount of signal in each lane of the filter was quantitated by using a Molecular Dynamics (Sunnyvale, CA) phosphorimager.

RESULTS The initial cDNA clones APtl and APt2 obtained from expression screening of a human B-cell cDNA library represented just over half of the complete sequence (12). One of these clones, APt2, was used to screen additional cDNA libraries to obtain the complete sequence. A schematic representation of the alignment of relevant clones is presented in Fig. 1. All cDNA clones were completely sequenced in both directions and were identical in overlapping regions.

Proc. Natl. Acad. Sci. USA 89 (1992)

Biochemistry: Bruhn et al.

By using the sequence information from these clones, a composite human sequence representing 2839 bases of DNA was generated (Fig. 2). There is a continuous open reading frame of 2127 bases beginning at position 275. The sequence surrounding the methionine start codon conforms well with the initiation sites of other vertebrate cDNAs (30) and is conserved in homologs isolated from mouse (24) and Drosophila melanogaster (S.L.B., D.E.H., and S.J.L., unpublished work). A consensus polyadenylylation signal AATAAA is present within the 435 bases of the 3' untranslated sequence beginning at position 2800. The sequence predicts a 709-amino acid protein of M, 81,068. The amino acid composition reveals a strikingly high percentage of charged residues (36%). Further analysis of the protein sequence indicated the presence of several highly charged domains, illustrated in Fig. 3. There is an acidic domain, positions 440-496, which contains 26 negatively charged and 4 positively charged amino acids. Two basic domains, denoted Basic I and Basic II, are located at amino acids 512-534 and 623-640, respectively. At the carboxyl terminus of the protein, positions 661-709, there is another highly charged series of amino acids containing 14 negative and 9 positive residues. Analysis of the hydropathy profile shows the entire region from position 400 to the carboxyl terminus of the protein to be highly hydrophilic (not shown). A search of protein data bases with the predicted amino acid sequence revealed some interesting homologies. SSRP1 showed the greatest homology to HMG1 and HMG2 proteins from several species (14, 15) and to a transcription factor containing HMG-box domains, hUBF (17). The location of the HMG box is indicated in Fig. 3, together with an alignment with the sequence of human HMG1 and the corresponding domains of other HMG-box protein family members. The acidic region of SSRP1 has limited homology to nucleolin (31), which is involved in transcriptional control of rRNA genes. Northern analysis with the cDNA clones isolated previously revealed a 2.8-kilobase (kb) message, which was conserved in humans and rodents (12). Further studies have more completely characterized the expression pattern of this protein. To determine its tissue specificity, total RNA was isolated from baboon brain, heart, ileum, jejunum, kidney, liver, muscle, and spleen tissue and was used for Northern analysis with the clone APt2 (Fig. 4). The 2.8-kb message is expressed in all tissues examined. Rehybridization of the filter with a fragment of human ,3actin allowed normalization for RNA loading levels (data not Eco RI

Size (bp)

A't2 YPtl HEK 402 HEK 1001 M 801 BG 801

1444

-1 -

1898 2692

12665 1

2176

1827 2839

275

ORF

2404

FIG. 1. Schematic representation of the relationship among huSSRP1. Clones APtl and APt2 were isolated by expression screening and are described in ref. 12. The cDNA APt2 was used here as a probe to screen several libraries for additional sequence information. One clone from a human embryonic kidney library contained the complete cDNA sequence and polyadenylylation signal and is denoted HEK 402. A second clone, HEK 1001, had all of the coding sequence and lacked only the consensus polyadenylylation signal. A third clone, clone M 801, isolated from a fetal muscle library, contained 147 bases of additional 5' untranslated sequence. A fourth clone, BG 801, from a basal ganglia cDNA library, was also sequenced. man cDNA clones encoding

Proc. Natl. Acad. Sci. USA 89 (1992)

Biochemistry: Bruhn et al.

GAATT CCGTA CGGCT TCCGG TGGCG GGACG CGGGG CCGCG CACGC GGGAA AAGCT TCCCC GGTGT CCCCC CATCC CCCTC CCCGC GCCCC CCCCG CGTCC CCCCA GCGCG CCCAC CTCTC GCGCC GGGGC CCTCG CGAGG CCGCA GCCTG AGGAG ATTCC CAACC TGCTG AGCAT CCGCA CACCC ACTCA GGAGT TGGGG CCCAG CTCCC AGTTT ACTTG GTTTC CCTTG TGCAG CCTGG GGCTC TGCCC AGGCC ACCAC AGGCA GGGGT

90 180 270

CGAC ATG GCA GAG ACA CTG GAG TTC AAC GAC GTC TAT CAG GAG GTG AAA GGT TCC ATG AAT GAT GGT CGA CTG AGG TTG AGC L S> N D G R L R V V K G S A E T L E F N D M M Y Q E

352

CGT CAG GGC ATC ATC TTC AAG AAT AGC AAG ACA GGC AAA GTG GAC AAC ATC CAG GCT GGG GAG TTA ACA GAA GGT ATC TGG E G I T G K V N I E Q G I I F K N S K Q A G L T N> R D

433 53

CGC CGT GTT GCT CTG GGC CAT GGA CTT AAA CTG CTT ACA AAG AAT GGC CAT GTC TAC AAG TAT GAT GGC TTC CGA GAA TCG K Y G F R E S> L L T K N G H V R V A L H L K D R G G Y

514 80

GAG TTT GAG AAA CTC TCT GAT TTC TTC AAA ACT CAC TAT CGC CTT GAG CTA ATG GAG AAG GAC CTT TGT GTG AAG GGC TGG H C V K W> K D G F K T H L L L E E S D F F K E E L Y R

595 107

AAC TGG GGG ACA GTG AAA TTT GGT GGG CAG CTG CTT TCC TTT GAC ATT GGT GAC CAG CCA GTC TTT GAG ATA CCC CTC AGC I P L P V F S> N W T V K L L S F I Q E G F Q D G D G G

676 134

AAT GTG TCC CAG TGC ACC ACA GGC AAG AAT GAG GTG ACA CTG GAA TTC CAC CAA AAC GAT GAC GCA GAG GTG TCT CTC ATG H> Q D A V S L V T L F H N E N V S Q C T T K N D G E E

757 161

GAG GTG CGC TTC TAC GTC CCA CCC ACC CAG GAG GAT GGT GTG GAC CCT GTT GAG GCC TTT GCC CAG AAT GTG TTG TCA AAG K> S F A Q N V L D V P V A V R F Y V P P T E Q E G D E

838 188

GCG GAT GTA ATC CAG GCC ACG GGA GAT GCC ATC TGC ATC TTC CGG GAG CTG CAG TGT CTG ACT CCT CGT GGT CGT TAT GAC Y D> 0 R L T P R G A V I A T I C I F L C D D A R E Q G

215

ATT CGG ATC TAC CCC ACC TTT CTG CAC CTG CAT GGC AAG ACC TTT GAC TAC AAG ATC CCC TAC ACC ACA GTA CTG CGT CTG R P T T V L> I R H K T F K I Y L I P T F L H L Y G D Y

1000 242

TTT TTG TTA CCC CAC AAG GAC CAG CGC CAG ATG TTC TTT GTG ATC AGC CTG GAT CCC CCA ATC AAG CAA GGC CAA ACT CGC P N P I K G T F F F V I S L P Q R> L H K D R Q L D Q Q

1081 269

TAC CAC TTC CTG ATC CTC CTC TTC TCC AAG GAC GAG GAC ATT TCG TTG ACT CTG AAC ATG AAC GAG GAA GAA GTG GAG AAG N V K> N E Y H D I S L T L N E F I L L F S K E E L D E

1162 296

CGC TTT GAG GGT CGG CTC ACC AAG AAC ATG TCA GGA TCC CTC TAT GAG ATG GTC AGC CGG GTC ATG AAA GCA CTG GTA AAC H N V N> N S F V K A L E S L Y V S R R G R L T K N G E

1243 323

CGC AAG ATC ACA GTG CCA GGC AAC TTC CAA GGG CAC TCA GGG GCC CAG TGC ATT ACC TGT TCC TAC AAG GCA AGC TCA GGA R C S Y K A S G> H S A C I T S K I T V P N F Q G G Q G

1324

CTG CTC TAC CCG CTG GAG CGG GGC TTC ATC TAC GTC CAC AAG CCA CCT GTG CAC ATC CGC TTC GAT GAG ATC TCC TTT GTC S F V> L F D E II Y V H K P P V H I R L Y P L E F I R G

1405 377

AAC TTT GCT CGT GGT ACC ACT ACT ACT CGT TCC TTT GAC TTT GAA ATT GAG ACC AAG CAG GGC ACT CAG TAT ACC TTC AGC F T T S> N F S F D F I T K Y A T T T T Q G Q R G E E R

1486 404

AGC ATT GAG AGG GAG GAG TAC GGG AAA CTG TTT GAT TTT GTC AAC GCG AAA AAG CTC AAC ATC AAA AAC CGA GGA TTG AAA K> S N I K N R G L F D F V N A K K L I E R E K L E Y G

1567 431

GAG GGC ATG AAC CCA AGC TAC GAT GAA TAT GCT GAC TCT GAT GAG GAC CAG CAT GAT GCC TAC TTG GAG AGG ATG AAG GAG 94 K R E> A Y L A D S D H E E M N P S G Y D E Y E D Q D

1648 458

GAA GGC AAG ATC CGG GAG GAG AAT GCC AAT GAC AGC AGC GAT GAC TCA GGA GAA GAA ACC GAT GAG TCA TTC AAC CCA GGT P N G> K I T D S F R E D S S D S E E G E N A N D E E G

1729 485

GAA GAG GAG GAA GAT GTG GCA GAG GAG TTT GAC AGC AAC GCC TCT GCC AGC TCC TCC AGT AAT GAG GGT GAC AGT GAC CGG S D S N D R> E E V A D S N A S A S S S E G E E D E F E

1810

GAT GAG AAG AAG CGG AAA CAG CTC AAA AAG GCC AAG ATG GCC AAG GAC CGC AAG AGC CGC AAG AAG CCT GTG GAG GTG AAG P V V K> 0 E K K K K R A K A K D K S R D E K L K K M R

1891

AAG GGC AAA GAC CCC AAT GCC CCC AAG AGG CCC ATG TCT GCA TAC ATG CTG TGG CTC AAT GCC AGC CGA GAG AAG ATC AAG K I K> K K P S R N A E G D N A P K P M S A Y M L W L R

1972 566

TCA GAC CAT CCT GGC ATC AGC ATC ACG GAT CTT TCC AAG AAG GCA GGC GAG ATC TGG AAG GGA ATG TCC AAA GAG AAG AAA N K K K> S D S K E H P S L S K K A I G M G I I T D G E

2053 593

GAG GAG TGG GAT CGC AAG GCT GAG GAT GCC AGG AGG GAC TAT GAA AAA GCC ATG AAA GAA TAT GAA GGG GGC CGA GGC GAG G E> R Y G G E E W D K A A R R D Y E K A M K E E R E D

2134 620

TCT TCT AAG AGG GAC AAG TCA AAG AAG AAG AAG AAA GTA AAG GTA AAG ATG GAA AAG AAA TCC ACG CCC TCT AGG GGC TCA H P S S> R G S S K R K K V K S T D K S K K K K V K E K

2215 647

TCA TCC AAG TCG TCC TCA AGG CAG CTA AGC GAG AGC TTC AAG AGC AAA GAG TTT GTG TCT AGT GAT GAG AGC TCT TCG GGA S S S G> 0 S S V S S D E K S S S E S F K S K E F R L S

2296

GAG AAC AAG AGC AAA AAG AAG AGG AGG AGG AGC GAG GAC TCT GAA GAA GAA GAA CTA GCC AGT ACT CCC CCC AGC TCA GAG S S E> A S T P P E N K S K S D S L K K R R E E E E E R

2377 701

GAC TCA GCG TCA GGA TCC GAT GAG TAG A AACGG AGGAA GGTTC TCTTT GCGCT TGCCT TCTCA CACCC CCCGA CTCCC CACCC *> D S A S G S D E

2460 709

ATATT TTGGT GGCAG TGGGG GATCC AAATC AGGCC TGTAG AAACC AAAAA

ACCAG TTTCT CCTCA TGAAA AGACG TCTTA ACTCT GCTGC CTCAT CTTAC TTTCC CGACC CTCCT ACCTG GGGCC TATTT AAGGA ATTC

TGCAG TTCCC TTAAG CTACT

TCCCT AAGGA GATGT TTCAT

GGATT TGGCT AGCTG TTTGT

CTGTG GTTTA CTGCT ATTTC

CCATC TAATT TGTCC TGGTC

TGAAC TGGGG TGTTC TGTGA

ATGCT CTCCT GTTGG TGTGT ATGTC AGAGA TAGGG TGGGA GGCAG GGCAA AAGTT GCTGG AGCAG GGGTC ATGTG AAATG ATTTA.AIMA GGGAA CTGAC

2309

26

919

350

512 539

674

ACTAG

2550

TGCAG AGGCC TTTGG

2640 2730 2820

2839

FIG. 2. Sequence and translation of human cDNA encoding SSRP1. The polyadenylylation signal is underlined.

shown) and revealed the relative levels of expression to be similar, except for brain tissue, in which it is higher.

Because of the exceptional success of cisplatin in treating testicular cancer, a more detailed analysis of expression was

Acidic

Basic I

HMG

Basic II

Mixed Charge

440-496

512 - 534

539 - 614

623 - 640

661- 709

4 E D K? NPNPK SSR1 539 hW tbl 104Y KI 5 DO D i5T (LG T D MR~o s I hUb2 188 QN T. M ( hUb3 399 GGKGGSE GGE I-I IFS 0 H E W SAI I SRY 576 E EQDI S LP[ L KTT atTli 1 .SSVLAS A2DAYM I KISRF EQjA 289 LKESAAI AMW2b1 43 RPELIK Q K F Q r OK!JSR K rMJ&)W AB12W 1 1 6 Wg3FDEXP KKY QS

KPK

LE-1

EQEPKTPH

I

HEj

FIG. 3. (Upper) Schematic representation ofvari ous domains of human cisplatin-DNA SSRP1.

(Lower) Optimal alignment of human cisplatin-DNA SSRP1 with human HMG1, revealing 47% identity in the regions compared. Also depicted are compari-

sons among HMG-box family members (17-23). Identical amino acids are boxed. No obvious consensus sequence emerges. mt, mitochondrial.

2310

Proc. Natl. Acad. Sci. USA 89 (1992)

Biochemistry: Bruhn et al. :3

P N

11. Further refinement with a series of hybrid cell lines containing only small defined segments of human chromosome 11 on a rodent genomic background (33) localized the position of the clone to 11q12. Further details may be found in ref. 34. Placement of the sequence on the long arm of human chromosome 11 is particularly interesting because the murine homolog has been mapped to mouse chromosome 2 (24). Previously, a syntonic relationship had been demonstrated only for mouse chromosome 2 and human chromosome lip (35). DISCUSSION

a

c

-

FIG. 4. Northern analysis of total RNA from baboon tissues with clone APt2 as probe.

carried out in a series of testicular carcinoma cell lines. Several bladder cancer cell lines (32) were studied concurrently because cisplatin is less active against this type of cancer. SSRP1 was expressed in all of the bladder and testicular cell lines examined; no general trends were apparent (data not shown). These data indicate that the mRNA level does not correlate with the antitumor activity of cisplatin for a particular tissue type. Since the protein described here specifically recognizes DNA adducts of active antitumor platinum complexes, its possible role in acquired resistance of cells to cisplatin was also investigated. Fig. 5 shows the results of Northern analysis in which the APt2 clone was used to probe cytoplasmic RNA levels in a series of cisplatin-resistant human, mouse, and hamster cell lines (S.L.B., J.M.E., D.E.H., and S.J.L., unpublished results). These data indicate that the level of expression does not correlate with resistance in these cell lines. To study whether expression of the SSRP1 could be induced in cells treated with the drug, cytoplasmic RNA was isolated from HeLa cells exposed to a range of concentrations of cisplatin. The 2.8-kb species was not inducible by a range of cisplatin concentrations over the course of 48 hr (data not shown). The human map position of SSRP1 was determined by using a panel of human chromosome-specific human-rodent hybrids. Initial experiments placed the gene on chromosome To

N

CD

~-

Z)

--

GX =

N

T

0o

0

N1 ,~

,-

CD

FIG. 5. Northern analysis of cytoplasmic RNA isolated from human cisplatin-resistant cell lines HeLa (parental), HeLa/1.2 (35), HeLa/3.0 (60), HeLa/6.0 (95), mouse L1210 (parental), L1210/7.5 (15), L1210/15 (25), hamster V79 (parental), V79/7.5 (30), and V79/15 (70) with clone APt2 as probe. Cell lines are named by placing the concentration (in ,Ag per ml of cisplatin) in which the cells are routinely subcultured after the name of the parental cell line. The resistance level follows in parentheses (e.g., HeLa/3.0 is 60-fold resistant). The lighter bands for the mouse and hamster lanes arise from weaker cross-species hybridization between the human probe and the rodent homologs.

The present results reveal the sequence of cDNA clones that encode a protein recognizing the specific structures formed when DNA is modified by cisplatin. Previous work has shown that the major cis-{Pt(NH3)2}2+ adducts, intrastrand d(GpG) and d(ApG) cross-links, unwind the duplex by 130 and cause a 340 bend in the direction of the major groove (3, 36, 37). Important clues for identifying the type of protein that might interact with such an altered structure are provided by the striking homology of SSRP1 to HMG1, which is known to bind cruciform DNA (38), and the near identity to a mouse protein that binds to signal sequences for V-(D)-J recombination (24). The common DNA structural element recognized by SSRP1 and HMG1, while not yet defined, most likely mimics the unwinding and bending known to occur in cisplatin-modified DNA. Moreover, our results raise the possibility that HMG1, the family of HMG-box proteins, and recombination functions may be involved in the molecular mechanism of the drug. The initial clones APtl and APt2 were obtained by using a functional screen-namely, that of protein binding to cisplatin-modified DNA. Therefore, the shorter clone, APt2, serves to define the region containing the DNA binding activity. This region extends from residues 149-627 of the full length protein and includes the acidic domain, Basic I, and the HMG box (Fig. 3). Among these, the HMG box is the one most likely to contain the Pt-DNA structure-recognition site, since parallel work from our laboratory has revealed the strong and specific binding of cisplatin-modified oligonucleotide probes to HMG1 (P.M.P. and S.J.L., unpublished results). HMG-box domains are emerging as an important recognition element of proteins for DNA. Deletion analysis of HMG-box family members hUBF (17) and TCF-1a (22) has demonstrated that a single HMG-box domain is sufficient for the specific interactions of these proteins with DNA. Lack of a clearly defined consensus sequence among the HMG-box domains in a variety of proteins (Fig. 3) may indicate either that such proteins recognize different DNA structures or that their different shapes are capable of recognizing similar DNA structures. Whereas mutations in the sequences of target recognition sites in DNA alter binding of the HMG-box proteins, such changes could also modify the shape of the recognition site, reducing its protein affinity. The suggestion (23) that HMG-box proteins recognize DNA structure rather than sequence is strongly supported by the fact that SSRP1 binds to cisplatin-modified DNA probes but not to unmodified probes (12). The known properties of HMG1 are fully consistent with its role in binding to altered DNA structures. HMG1 suppresses nucleosome core particle formation (39), and it can selectively unwind negatively supercoiled DNA, protecting it from relaxation by E. coli topoisomerase I and preventing the formation of higher order secondary structure (40). It binds preferentially to A+T-rich regions (41), single-stranded DNA (42), BZ junctions (43), and cruciform structures (38). Studies of plasmid DNA containing a number of structural domains suggest that HMG1 can differentiate among various DNA conformations (43).

Biochemistry: Bruhn et al.

Proc. Natl. Acad. Sci. USA 89 (1992)

Several studies suggest that HMG1 and HMG2 act by

binding to specific structural elements in DNA upstream from actively transcribed genes to preserve conformations necessary for the binding of sequence-specific transcription factors (44-48). In particular, HMG1 removes the transcriptional block caused by cruciforms in supercoiled DNA (49). Eukaryotic DNA contains palindromic sequences that form such structures, which may have elements in common with the intrastrand d(GpG) and d(ApG) cross-linked complexes formed by cisplatin-modified DNA. Additional insights into the possible role of SSRP1 are provided by the recent characterization of a mouse cDNA clone isolated by screening an expression library with oligonucleotides containing recombination signal sequences (RSS) (24). RSS sequences are signals for somatic DNA recombination to generate antibody diversity through V-(D)-J joining. The mouse protein is 95.5% identical with the human DNA SSRP1 described here and clearly is the mouse homolog of the human and Drosophila (S.L.B., D.E.H., and S.J.L., unpublished work) clones. Interestingly, V-(D)-J recombination is postulated to proceed via stemloop structures formed by RSS sequences (50, 51). The similarity among stem-loop DNA, cruciforms recognized by HMG1, and the bent, unwound cisplatin intrastrand crosslinked conformations is intriguing and indicates that binding of this HMG-box protein to RSS involves shape and sequence recognition elements. Whether SSRP1 plays a role in the cisplatin antitumor mechanism is as yet unclear. Its expression pattern suggests a function that is critical to a variety of tissues. Its presence does not correlate with the tissue-specific antitumor activity of cisplatin, however, nor with drug sensitivity in a series of resistant cell lines. Expression of the encoded message was not inducible in HeLa cells treated with a range of drug concentrations. It is possible that binding of SSRP1 or other SSRPs in vivo protect Pt-DNA adducts from repair, allowing them to block DNA replication and effect cell death (9). This action could also titrate such proteins away from their normal cellular functions, contributing to cytotoxicity (9, 52). The present work suggests that, for either of these possibilities, two specific proteins that might be diverted from their normal regulatory intracellular roles are HMG1 and the SSRP1, and that somatic DNA recombination might be a specific function affected by the platinum anticancer drug family. Understanding the shape recognition elements of these proteins may provide a basis for the design of future generations of chemotherapeutic agents. We thank C. A. Pabo, P. A. Sharp, J. Stubbe, and J. C. Wang for comments on the manuscript; B. A. Donahue and J. H. Toney for valuable advice and experimental assistance; L. A. Doucette-Stamm and A. Buckler for helpful discussions; B. Z. Stanger for the physical mapping studies; J. R. W. Masters for providing the testicular and

bladder carcinoma cell lines used for Northern analysis; and B. Nelson, L. Kunkel, and Integrated Genetics for providing cDNA libraries. This research was supported by grants from the National Cancer Institute (CA 32134 to S.J.L. and CA 52127 to J.M.E.) and Bristol-Myers (to D.E.H. and S.J.L.) and by a Howard Hughes Medical Institute Predoctoral Fellowship to P.M.P. 1. Bruhn, S. L., Toney, J. H. & Lippard, S. J. (1990) Prog. Inorg. Chem. 38, 477-516. 2. Churchill, M. E. A. & Travers, A. A. (1991) Trends Biochem. Sci.

16, 92-97. 3. Bellon, S. F., Coleman, J. H. & Lippard, S. J. (1991) Biochemistry 30, 8026-8035. 4. Lepre, C. A. & Lippard, S. J. (1990) Nucleic Acids Mol. Biol. 4,

10. 11.

12. 13. 14. 15.

16. 17. 18. 19. 20. 21. 22.

23. 24. 25. 26. 27. 28. 29.

30. 31.

32. 33. 34.

35. 36. 37. 38. 39. 40. 41. 42.

43. 44. 45. 46. 47.

48. 49.

9-38.

5. 6. 7. 8. 9.

Sherman, S. E. & Lippard, S. J. (1987) Chem. Rev. 87,1153-1181. Reedijk, J. (1987) Pure Appl. Chem. 59, 181-192. Eastman, A. (1987) Pharmacol. Ther. 34, 155-166. Chu, G. & Chang, E. (1988) Science 242, 564-567. Donahue, B. A., Augot, M., Bellon, S. F., Treiber, D. K., Toney,

50. 51. 52.

2311

J. H., Lippard, S. J. & Essigmann, J. M. (1990) Biochemistry 29, 5872-5880. Andrews, P. A. & Jones, J. A. (1991) Cancer Commun. 3, 1-10. Chao, C.-K., Huang, S.-L., Lee, L.-Y. & Lin-Chao, S. (1991) Biochem. J. 277, 875-878. Toney, J. H., Donahue, B. A., Kellett, P. J., Bruhn, S. L., Essigmann, J. M. & Lippard, S. J. (1989) Proc. Natl. Acad. Sci. USA 86, 8328-8332. Singh, H., LeBowitz, J. H., Baldwin, A. S., Jr., & Sharp, P. A. (1988) Cell 52, 415-423. Einck, L. & Bustin, M. (1985) Exp. Cell Res. 156, 295-310. Bustin, M., Lehn, D. A. & Landsman, D. (1990) Biochim. Biophys. Acta 1049, 231-243. van Holde, K. E. (1988) Chromatin (Springer, New York). Jantzen, H. M., Admon, A., Bell, S. P. & Tjian, R. (1990) Nature (London) 344, 830-836. Sinclair, A. H., Berta, P., Palmer, M. S., Hawkins, J. R., Griffiths, B. L., Smith, M. J., Foster, J. W., Frischauf, A.-M., Lovell-Badge, R. & Goodfellow, P. N. (1990) Nature (London) 346, 240-244. Gubbay, J., Collignon, J., Koopman, P., Capel, B., Economou, A., Munsterberg, A., Vivian, N., Goodfellow, P. & Lovell-Badge, R. (1990) Nature (London) 346, 245-250. Parisi, M. A. & Clayton, D. A. (1991) Science 252, 965-969. Travis, A., Amsterdam, A., Belanger, C. & Grosschedl, R. (1991) Genes Dev. 5, 880-894. Waterman, M. L., Fischer, W. H. & Jones, K. A. (1991) Genes Dev. 5, 656-669. Diffley, J. F. X. & Stillman, B. (1991) Proc. Natl. Acad. Sci. USA 88, 7864-7868. Shirakata, M., Huppi, K., Usuda, S., Okazaki, K., Yoshida, K. & Sakano, H. (1991) Mol. Cell. Biol. 11, 4528-4536. Feinberg, A. P. & Vogelstein, B. (1983) Anal. Biochem. 132, 6-13. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Natl. Acad. Sci. USA 74, 5463-5467. Devereux, J., Haeberli, P. & Smithies, 0. (1984) Nucleic Acids Res. 12, 387-395. Altschul, S. F., Gish, W., Miller, W., Myers, E. W. & Lipman, D. J. (1990) J. Mol. Biol. 215, 403-410. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual (Cold Spring Harbor Lab., Cold Spring Harbor, NY). Kozak, M. (1987) Nucleic Acids Res. 15, 8125-8132. Srivastava, M., Fleming, P. J., Pollard, H. B. & Burns, A. L. (1989) FEBS LetI. 250, 99-105. Masters, J. R. W., Hepburn, P. J., Walker, L., Highman, W. J., Trejdosiewicz, L. K., Povey, S., Parkar, M., Hill, B. T., Riddle, P. R. & Franks, L. M. (1986) Cancer Res. 46, 3630-3636. Glaser, T. (1989) Ph.D. dissertation (Massachusetts Institute of Technology, Cambridge, MA). Bruhn, S. L. (1992) Ph.D. dissertation (Massachusetts Institute of Technology, Cambridge, MA). Nadeau, J. H., Davisson, M. T., Doolittle, D. P., Grant, P., Hillyard, A. L., Kosowsky, M. & Roderick, T. H. (1991) Mamm. Genome 1, S461-S515. Bellon, S. F. & Lippard, S. J. (1990) Biophys. Chem. 35, 179-188. Rice, J. A., Crothers, D. M., Pinto, A. L. & Lippard, S. J. (1988) Proc. Natl. Acad. Sci. USA 85, 4158-4161. Bianchi, M. E., Beltrame, M. & Paonessa, G. (1989) Science 243, 1056-1059. Waga, S., Mizuno, S. & Yoshida, M. (1989) Biochim. Biophys. Acta 1007, 209-214. Sheflin, L. G. & Spaulding, S. W. (1989) Biochemistry 28, 56585664. Reeves, R. & Nissen, M. S. (1990) J. Biol. Chem. 265, 8573-8582. Isackson, P. J., Fishback, J. L., Bidney, D. L. & Reeck, G. R. (1979) J. Biol. Chem. 254, 5569-5572. Hamada, H. & Bustin, M. (1985) Biochemistry 24, 1428-1433. Tremethick, D. J. & Molloy, P. L. (1986) J. Biol. Chem. 261, 6986-6992. Tremethick, D. J. & Molloy, P. L. (1988) Nucleic Acids Res. 16, 11107-11123. Watt, F. & Molloy, P. L. (1988) Nucleic Acids Res. 16, 1471-1486. Waga, S., Mizuno, S. & Yoshida, M. (1988) Biochem. Biophys. Res. Commun. 153, 334-339. Singh, J. & Dixon, G. H. (1990) Biochemistry 29, 6295-6302. Waga, S., Mizuno, S. & Yoshida, M. (1990) J. Biol. Chem. 265, 19424-19428. Max, E. E., Seidman, J. G. & Leder, P. (1979) Proc. Natl. Acad. Sci. USA 76, 3450-3454. Sakano, H., Huppi, K., Heinrich, G. & Tonegawa, S. (1979) Nature (London) 280, 288-294. Scovell, W. M. (1989) J. Macromol. Sci. Chem. A26, 455-480.