p19 skpl and p45 skp2 are essential elements of the cyclin a-cdk2

Cell, Vol. 82, 915-925, September 22, 1995, Copyright © 1995 by Cell Press

p19 skpl and p45 skp2 Are Essential Elements of the Cyclin A-CDK2 S Phase Kinase

Hui Zhang, Ryuji Kobayashi, Konstantin Galaktionov, and David Beach Howard Hughes Medical Institute Cold Spring Harbor Laboratory Cold Spring Harbor, New York 11724

Summary

In normal human fibroblasts, cyclin A-CDK2 exists in a quaternary complex that contains p21 and PCNA. In many transformed cells, p21 disappears, and a substantial fraction of cyclin A-CDK2 complexes with p9 cKsl/cKs2, p19, and p45. To investigate the signifi- cance of these rearrangements, we have isolated cDNAs encoding p19 and p45. In vitro reconstitution demonstrated that binding of p19 to cyclin A-CDK2 requires p45. Addition of these proteins to the kinase had no substantial effect on the kinase activity in vitro. Interference with p45 function in vivo by microinjection of antibodies or antisense oligonucleotides prevented entry into S phase in both normal and transformed cells. Cyclin A-CDK2 has previously been identified as a kinase whose activity is essential for S phase. Our results identify p45 as an essential element of this activity. The abundance of p45 is greatly increased in many transformed cells. This could result in changes in cell cycle control that contribute to the process of cellular transformation.

Introduction

In eukaryotes, entry into each phase of the cell division cycle is controlled by a class of enzymes known as the cyclin-dependent kinases (CDKs) (reviewed by Sherr, 1993, 1994; Hunter and Pines, 1994). In their simplest active form, these consist of a catalytic subunit (the CDK) and a positive regulatory subunit known as a cyclin. In mammalian cells, the CDK family consists of at least seven members. Combination with an equally diverse family of cyclins yields numerous cell cycle regulatory enzymes, each with a potentially unique function.

Considering the role of CDKs in controlling cell proliferation, it is not surprising that these enzymes are subject to multiple levels of regulation. First, the activity of CDKs depends upon the availability of their cognate cyclin subunits, many of which either oscillate in abundance during the cell cycle or require the presence of growth factors for expression (reviewed by Sherr, 1993). Once formed, cyclin-CDK complexes still require phosphorylation of a threonine residue (usually near amino acid 160) for activity (reviewed by Solomon, 1993; Dunphy, 1994). Active enzymes can be further regulated by inhibitory phosphorylation (e.g., of Thr-14 and Tyr-15 in cdc2) and by binding of inhibitory proteins (e.g., p21 or p16) (reviewed by Hunter and Pines, 1994; Xiong et al., 1993b; Harper et al., 1993; Serrano et al., 1993).

Differences in the abilities of normal and transformed cells to proliferate had long predicted that alterations in the pathways that control cell cycle progression must ac- company cellular transformation. This prompted a comparison of cyclin-CDK complexes present in normal fibroblasts to those present in their transformed derivatives (Xiong et al., 1993a). In normal cells, each cyclin-CDK enzyme exists in a quaternary complex that also contains proliferating cell nuclear antigen (PCNA) and p21 (Zhang et al., 1993). Quaternary complexes are lost from many transformed cells owing to the absence of the tumor suppressor p53, which directly controls the expression of p21 (Xiong et al., 1993a; EI-Deiry et al., 1993). In the presence of DNA damage, elevation of p53 levels leads to the accumulation of p21 (Dulie et al., 1994; EI-Deiry et al., 1994), which acts as an inhibitor of cyclin-CDK enzymes when bound at greater than unit stoichiometry (Zhang et al., 1994).

In normal cells, CDK4 and CDK6 partner with D-type cyclins in quaternary complexes (Matsushime et al., 1992; Xiong et al., 1992; Zhang et al,, 1993; Meyerson and Har- low, 1994). However, in virally transformed cells, a 16 kDa protein, p16, is the exclusive partner of these kinase subunits (Xiong et al., 1993a). The critical role for CDK4 and CDK6 in cell cycle progression is to phosphorylate the tumor suppressor pRb and consequently permit the transition from G 1 into S phase (Matsushime et al., 1994; Meyer- son and Harlow, 1994). p16 is a specific inhibitor of CDK4 and CDK6 whose effect would normally be to halt cell cycle progression (Serrano et al., 1993, 1994). However, in the virally transformed cells, pRb function is compromised, and p16 is therefore without effect (Xiong et al., 1993a; Serrano et al., 1994).

Our original comparisons of cyclin kinase complexes in normal and transformed fibroblasts also identified a third class of subunit rearrangements that specifically affected cyclin A-CDK2 complexes (Xiong et al., 1993a). In many transformed cells, a cyclin A-CDK2 complex(es) containing proteins of 9, 19, and 45 kDa replaces the quaternary cyclin A-CDK2-p21-PCNA complexes found in normal cells (see below).

Cyclin A-associated enzymes have been established as key promoters of progression through the S phase of the cell cycle (reviewed by Heichman and Roberts, 1994). The expression of cyclin A and the activity of cyclin A-CDK2 peaks during late G1 and S phase in both normal and transformed cells (Pines and Hunter, 1990). Furthermore, microinjection of antibodies to cyclin A or introduction of antisense cyclin A expression constructs can prevent DNA replication (Girard et al., 1991; Pagano et al., 1992). In some cell types, cyclin A has even been shown to localize to sites of ongoing DNA synthesis (Cardoso et al., 1993). Finally, ectopic expression of cyclin A was sufficient to allow adhesion independent DNA synthesis in normal rat kidney cells (Guadagno et al., 1993).

Thus far, our studies of the rearrangements of cyctin kinase complexes that occur upon cellular transformation

Cell 916

have identified at least two direct links between tumor suppression and the enzymes that regulate cell cycle progression. p21 is a direct transcriptional target of the tumor suppressor p53 and is a likely effector of the cell cycle arrest that is induced by p53 in response to DNA damage (EI-Deiry et al., 1993, 1994; Duli~ et al., 1994). The p16 protein is a major tumor suppressor in its own right and acts in a regulatory pathway that also contains another tumor suppressor, pRb, and an oncogene, cyclin D (reviewed by Hunter and Pines, 1994). In this paper, we de- scribe the isolation of cDNAs encoding the 19 and 45 kDa subunits of the cyclin A-CDK2 complex that is abundant in tumor cells. Our data suggest that the p45-bound form of cyclin A-CDK2 is required for execution of DNA replication in both normal and transformed cells.

Results

Cyclin A-CDK2 Complexes Differ in Normal and Transformed Cells In normal human fibroblasts, cyclin A-CDK2 exists in a quaternary complex with p21 and PCNA (Zhang et al., 1993). We have previously reported the dissolution of this complex in fibroblasts transformed by viral oncoproteins (Xiong et al., 1993a). In these cells, cyclin A instead complexes with CDK2 and proteins of 9, 19, and 45 kDa (p9, p19, and p45) (Figure 1). These proteins are the predomi- nant partners for cyclin A not only in virally transformed cells (e.g., HeLa and 293; Figure 1) but also in many transformed cells that express no viral oncoprotein (e.g., ML1 ; Figure 1A). Under no circumstances have we observed p19 or p45 in association with other cyclin-CDK complexes (data not shown). However, p9 can also be found in cyclin B-CDC2 immunoprecipitates (data not shown).

Since p9, p19, and p45 invariably appear together in cyclin A immunoprecipitates, we considered the possibility that these proteins might exist, with CDK2, in a single complex. To test this possibility, proteins present in a lysate prepared from ~S-labeled 293 cells were separated by sedimentation in a glycerol gradient. Cyclin A and its associated proteins were recovered from each fraction of the gradient by immunoprecipitation with a cyclin A antiserum. Peak levels of p9, pl 9, and p45 were found in fraction 17, which also contained substantial amounts of cyclin A and CDK2 (Figure 1B). This strongly suggested that these five proteins coexist in a single complex. Western blotting demonstrated that the p19-p45-containing complex was well resolved from the cyclin A-CDK2 complex that contains p107 and E2F (peaked in fraction 25; data not shown).

Cyclin A-CDK2 complexes are active from the onset of S phase through G2 phase and into early M phase, at which point the complex is inactivated by degradation of cyclin A. To gain some clue to the function of the p19- p45-containing complex, we asked whether its abundance varied during the cell cycle. Treatment of HeLa cells with nocodazole, a drug that causes arrest in late G2/early M phase, reduced the amount of p19 and p45 in cyclin A immunoprecipitates (data not shown). However, synchro- nization of HeLa cells in S phase by treatment with hy-

A

200-

97-

68 -

43 -

30 -

2 9 3

o

WI38 ML1 IMR90

HeLa HS68

+ +

--' z~ -' - _ ~..-, ._, ._,

~ m

~,m,~ Cyc A mtlf

' ~ " ~ ~ PCNA

18-

14-

3 p 9

B

200-

96-

68-

43-

30-

1 3 5

293/anti-cyc A

7 9 11 13 15 17 19 21 23 25 27 29 31

- cyc A

- p45

3 CDKs

18-

14-

- p19

- p9

Figure 1. The Presence of Two Types of Cyclin A-CDK2 Complexes in Transformed and Nontransformed Cells (A) Comparison of cyclin A-CDK2 complexes between transformed and nontransformed cells. Cell lysates from either [35S]methionine- labeled, nontransformed human fibroblasts (Wl38, IMR90, and Hs68) or transformed cells (293, ML1, and HeLa) were immunoprecipitated with either normal rabbit serum (NRS), anti-cyclin A (anti-Cyc A), or anti-CDK2 antibodies or with the anti-CDK2 antibody plus CDK2 anti- genic peptide as indicated. The electrophoretic positions of cyclin A and CDK2-associated proteins are indicated on the right. The protein molecular mass markers (in kilodaltons) are indicated on the left. (B) Glycerol gradient analysis of quinary cyclin A-CDK2-p9-p19-p45 protein complex. Cell lysate from [3SS]methionine-labeled 293 cells was loaded directly onto a 15%-35% glycerol gradient and sedimented. The fractions (250 ~.1 each) were collected from the top (left) to bottom (right) of the gradient. Cyclin A (cyc A) complexes were recovered from the fractions by cyclin A antibody and analyzed by SDS-polyacylamide gel electrophoresis (SDS-PAGE). The peak of bovine serum albumin, sedimented in parallel as molecular mass marker, migrates at fraction 15.

p19 and p45: Essential Elements of Cyclin A-CDK2 917

droxyurea resulted in a dramatic increase in cyclin A-bound p9, p19, and p45 (data not shown). These results suggested that the relative abundance of the cyclin A-CDK2- p9-p19-p45 complex was increased during S phase.

Isolation of cDNA Clones That Encode p19 and p45 As a prerequisite to understanding the consequences of the association of p9, p19, and p45 with cyclin A in transformed cells, it was necessary to identify the protein constituents of this complex. The mobility of p9 was reminis- cent of previously characterized CDK-associated proteins, p9 cKsl and p9 cKs2 (Richardson et al., 1990). These homologs of fission yeast p13 s"°~ have been shown to bind to CDKs in mammalian cells. A partial V8 protease digest of the cyclin A-associated p9 was virtually identical to similar digests of in vitro translated CKS1 or CKS2 (data not shown), indicating that the cyclin A-associated p9 was a member of the human CKS family. However the high de- gree of amino acid identity between CKS1 and CKS2 made it impossible to determine whether the cyclin A-associated p9 was identical to either CKS1 or CKS2 or was instead a mixture of the two.

Neither p19 nor p45 had a mobility that was identical to a known CDK-associated protein. We therefore sought to isolate cDNAs encoding these potentially uncharacter- ized components of cyclin A-CDK2. Cyclin A and its associated proteins were purified from lysates of human 293 cells using an anti-cyclin A immunoaffinity column. Pro- teins were eluted from the affinity matrix by boiling in Laemmli's sample buffer. Individual proteins present in the eluent were separated on a preparative SDS-polyacrylamide gel. Bands corresponding to p19 and p45 were ex- cised from the gel, and the proteins were digested in situ with achromobacter endoprotease I. Following elution from the gel, individual peptide fragments from each protein were purified using high pressure liquid chromatogra- phy and sequenced by an automated peptide sequencer. Sequences of several peptides were obtained from each protein (Figures 2A and 2B).

The sequences of the p19 and p45 peptides were used to design degenerate oligonucleotide primers. These were used to amplify fragments of the p19 and p45 cDNAs from a HeLa cell library. In each case, the fragments contained sequences encoding other p19 or p45 peptides obtained from protein sequencing. Therefore, the fragments were used as probes to identify full-length cDNAs encoding candidate p19 and p45 proteins.

The cDNA encoding the candidate for pl 9 contained an open reading frame of 163 amino acid residues corresponding to a protein with a calculated molecular mass of 18,645 Da (Figure 2B). The deduced amino acid sequence of p19 indicated that it was not a member of the p21 family. A search of available databases revealed no significant homology to any protein with a clearly defined function (see below).

In vitro translation of the candidate p19 cDNA produced a protein with an electrophoretic mobility identical to that of the p19 present in an anti-cyclin A immunoprecipitate from 293 cells (Figure 3A, left). A comparison of a partial

V8 protease digest of the in vitro translated p19 to a similar digest of p19 from a cyclin A immunoprecipitate confirmed that our cDNA encoded a complete p19 protein (data not shown). A similar approach was used to verify our candidate p45 cDNA (Figure 3A).

The p45 cDNA encoded a polypetide of 435 amino acids with a calculated molecular mass of 48,928 Da (see Figure 2A). As with p19, a search of available databases yielded no informative homologies. However, this search did re- veal that the carboxy-terminal half of the p45 protein is composed of seven imperfect repeats of 26 amino acid residues (see Figure 2C). This motif shares homology with the leucine-rich repeats that are found in a number of func- tionally diverse proteins, including adenylate cyclase, SDS22, and the RAD7 proteins of Saccharomyces cerevisiae and Schizosaccharomyces pombe (Kataoka and Wigler, 1985; Perozzi and Prakash, 1986; Ohkura and Ya- nagida, 1991). The leucine-rich repeated motifs are thought to be involved in protein-protein interactions (Kobe and Deisenhofer, 1994).

Since p19 and p45 are previously unreported proteins that associate with the cyclin A-CDK2 preferentially during S phase in transformed cells, we have named them pl 9 skpl and p45 skp2 (for S phase kinase-associated protein).

p19, Cyclin A, and CDK2 Are the Major p45-Associated Proteins To examine p19 and p45 protein complexes in vivo, we used specific antibodies raised against p45 and p19 proteins. Immunoprecipitation of p45 from [35S]methionine- labeled 293 cell lysates revealed that cyclin A, CDK2, p9, and p19 are the major p45-associated proteins in human cells (Figure 3C, lane labeled anti-p45). This cyclin A-CDK2- pg-p19-p45 complex was an active kinase capable of phosphorylating histone H1 (Figure 3C).

The cyclin A-CDK2-p19-p45 complex was initially identified as a major cyclin A-CDK2 complex in many transformed cells (see Figure 1). However, p45 was also present at low level in normal human fibroblasts (Figure 3B). As in transformed cells, p45 in normal cells associates with cyclin A and CDK2 (Figure 3B).

Using a p19 antiserum, we did not recover cyclin A, CDK2, or p45 in the immunoprecipitates (Figures 3B and 3C). Furthermore, we could not detect any histone H1 kinase activity in the anti-p19 immunoprecipitates. This indicated either that the p19 antiserum did not recognize the complexed form of p19 or that the vast majority of p19 in human cells is not bound to the cyclin A-CDK2 enzyme.

Reconstitution of Cyclin A-CDK2-p19-p45 Complexes To examine the biochemical consequences of complex formation between cyclin A-CDK2 and p9, p19, and p45, we performed in vitro reconstitution experiments using proteins expressed in baculovirus-infected insect cells. Addition of p45 alone to cyclin A-CDK2 resulted in the formation of a ternary complex (Figure 4A). p45 could also bind cyclin A alone, but to a much lower extent than in the presence of CDK2.

Cell 918

A i0 30 50

G~TTCCGGGCTGTAGAGCTTGCCCGCGCAGTGGGGATGG~CGTTGCTAGGCTTAGCGG 70 90 ii0

GTCTGGCTGCTGGAGGCCCGAGCAGCACGCTCGAGCCGACGCGCGCCAAAGCGGG~TCT 130 150 170

GG~GGCG~GCAGCTCTGC~GTTT~TGCACGTATTTAAAACTCCCGGGCCTGCGGAC M H V F K T P G P A D

190 210 230 GCTATGCACAGG~GCACCTCCAGGAGATTCCAGACCTGAGTAGC~CGTTGCCACCAGC A M H R K H L Q E I P D L S S N V A T S

250 270 290 TTCACGTGGGGATGGGATTCCAGC~GACTTCTG~%CTGCTGTCAGGCATGGGGGTCTCC [ T W G W D S S K T S E L L S G M G V S

310 330 350 GCCCTGGAGAAAGAGGAGCCCGACAGTGAG~CATCCCCCAGG~CTGCTCTCAAACCTG A L E K E E P D S E N I P Q E L L S N L

370 390 410 GGCCACCCGGAGAGCCCCCCACGGAAACGGCTG~GAGCAAAGGGAGTGACAAAGACTTT

G H P E S P P R K R L K S K G S D K D [

430 450 470 GT~TTGTCCGCAGGCCT~GCTAAATCGGGAG~CTTTCCAGGTGTTTCATGGGACTCT V I V R R P K L N R E N F P G V S W D S

490 510 530 CTTCCGGATGAGCTGCTCTTGGG~TCTTTTCCTGTCTGTGCCTCCCTGAGCTGCTAAAG

L P D E L L L G I F S C L C L P E L L K 550 570 590

GTCTCTGGTGTTTGT~GAGGTGGTATCGCCTAGCGTCTGATGAGTCTCTATGGCAGACC V S G V C K R W Y R L A S D E S L W Q T

610 630 650 TTAGACCTTACAGGTA~TCTGCACCCGGATGTGACTGGTCGGTTGCTGTCTC~GGG L D L T G K N L H P D V T G R L L S Q G

670 690 710 GTGATTGCCTTCCGCTGCCCACGATCATTTATGGACC~CCATTGGCTG~CATTTCAGC V I A F R C P R S F M D Q P L A E H F S

730 750 770 CCTTTTCGTGTACAGGACATGGACCTATCG~CTCAGTTATAG~GTGTCCACCCTCCAC P F R V Q D M D L S N S V I E V S T L H

790 810 830 GGCATACTGTCTCAGTGTTCC~GTTGCAG~TCT~GCCTGG~CTGCGGCTTTCGGAT G I L S Q C S K L Q N L S L E L R L S D

850 870 890 CCCATTGTC~TACTCTCGCAAAAAACTCAAATTTAGTGCGACTT~CCTTCCTGGGTGT P I V N T L A K N S N L V R L N L P G C

910 930 950 CCTGGATTCCCTAAATTTCCCCTGCAGACTTTCCT~GCAGCTGTCCCAGACTGGATGAG P G F P K F P L Q T F L S S C P R L D E

970 990 i010 CTG~CCTCTCCTGGTGTTTT~TTTCACTGAAAAGCATGTACAGGTGGCTGTTGCGCAT L N L S W C F N F T E K H V Q V A V A H

1030 1050 1070 GTCTCAGAGACCATGACCCAGCTG~TCT~GCGGCTACAGAAAG~TCTCCAGAAATCA V S E T M T Q L N L S G Y R K N L Q K S

1090 iii0 i130 GATCTCTCTACTTTAGTTAG~GATGCCCC~TCTTGTCCATCTAGACTT~GT~TAGT D L S T L V R R C P N L V H L D L S N S

1150 1170 1190 GTCATGCTAAAG~TGACTGCTTTCAGG~TTTTCCCAGCTC~CTACCTCC~CACCTA

V M L K N D C F Q E F S Q L N Y L Q H L 1210 1230 1250

TCACTCAGTCGGTGCTATGATAT~TACCTGAAACTTTACTTG~CTTGGAGAAATTCCC S L S R C Y D I I P E T L L E L G E I P

1270 1290 1310 ACACTAAA~kCACTAC~GTTTTTGG~TCGTGCCAGATGGTACCCTTC~CTGTTAAAG

T L K T L Q V F Q ~ V P D G T L Q L L K

1330 1350 1370 G~GCCCTTCCTCATCTACAGATT~TTGCTCCCATTTCACCACCATTGCCAGGCC~CT E A L P H L Q I N C S H F T T I A R P T

1390 1410 1430 ATTGGC~CAAAAAG~CCAGGAGATATGGGGCATCAAATGCCGACTGACACTGCAAAAG I G N K K N Q E I W G I K C R L T L Q K

1450 1470 1490 CCCAGTTGTCTATG~GTATTTATTGCAGGATGGTGTCTCTTCTTTAG~CAGGGAAAAT

P S C L * 1510 1530 1550

AGGCAGG~GCCC~TTGCTGGAGTACTTAGCTAGTTTTATTCTTGGTTTTCCCTTTTGC

1570 1590 CTGTCATTCTGC~GTATACTAGGGAGCCCATTTTGAGAG

The p19 protein could not independently associate with cyclin A-CDK2 in vitro. However, binding of p19 was observed in the presence of p45 (Figure 4A, right), suggesting that p45 mightbridge the interaction between p19 and cyclin-CDK complexes. In agreement with this model, p19 and p45 can form an independent binary complex (data not shown), p9 cKsl or p9 cKs2 was capable of joining cyclin A-CDK2 irrespective of the presence of p45 or pl 9. However, p9 itself had no measurable effect on the ability of p45 or p19 to bind cyclin A complexes (data not shown).

Using complexes reconstituted in vitro, we have asked

a i0 30 50 AGGATCCTCGAGGCCACGAAGGCCCTC TCCCTTCGCAAACGCCTCCCGGCTCTCGTAAGC

70 90 ii0 CTCC CGCCGGCCGTCTCCTTAACACCGAACACCATGCCTTCAATTAAGTTGCAGAGTTCT

M P S I K L Q S S 130 150 170

GATGGAG~GATATTTGAAGTTGATGTGGAAATTGCCAAACAATCTGTAACTATTAAGACC D G E I F E V D V E • A K Q S V T I K T

190 210 230 ATGTTGGAAGATTTGGGAATGGATGATGAAGGAGATGATGACC CAGTTC CTC TACCAAAT M L E D L G M D D E G D D D P V P L P N

250 270 290 GTGAATGCAGCAATATTAAAAAAGGTCATTCAGTGGTGCACCCACCACAAGGATGAC CC T V N A A I L K K V [ Q W C T H H K D D P

310 330 350 CCTCCTCCTGAAGATGATGAGAACAAAGAAAAGCGGACAGATGATATCCCTGTTTGGGAC P P P E D D E N K E K R T D D I P V W D

370 390 410 CAAGAATTC CTGAAAGTTGACCAAGGAACACTTTTTGAACTCATTCTGGCTGCAAACTAC Q E F L K V D Q G T L F E L I L A A N Y

430 450 470 TTAGACATCAAAGGTTTGCTTGATGTTACATGCAAGACTGTTGCCAATATGATCAAGGGG L D I K G L L D V T C K T V A N M I K G

490 510 530 AAAACTC CTGAGGAGATTCGCAAGACCTTCAATATCAAAAATGACTTTACTGAAGAGGAG K T P E E I R K T F N I K N D F T E E E

550 570 590 GAAGCCCAGGTACGCAAAGAGAACCAGTGGTGTGAAGAGAAGTGAAATGTTGTGCCTGAC E A O V R K E N Q W C E E K *

610 AC TGTAACACTGTAAGGAT

C p45 leucine-rich repeats (a.a.220-400):

220 LQNLSLELRLSDPIVNTLAKNSN ~ LVR

246 LNLPGCPGFPKFPLQTFLSSCPR-LDE

272 LNLSWCFNFTEKHVQVAVAHVSETMTQ

299 LNLSGYRKNLQKSDLSTLVRRCPNLVH

326 LDLSNSVMLKNDCFQEFSQLNY--LQH

351 LSLSRCYDIIPETLLELGEIPT -LKT

376 LQVFGIVPDGTLQLLKEALPHL--QIN

consensus: LNLS-C--L --LL-L ...... L--

F QF

245

271

298

325

350

375

400

Figure 2. The DNA and the Deduced Amino Acid Sequences of p19 and p45 (A) DNA and the deduced amino acid sequences of p45. The peptides that were obtained from microsequencing of the purified p45 protein are underlined. (B) DNA and the deduced amino acid sequences of pl 9. The peptides that were obtained from microsequencing of the purified p19 protein are underlined. (C) p45 contains a leucine-rich repeated domain at its carboxyl region.

whether binding of p45 alone or of p45 and p19 had any effect on the enzymatic activity of cyclin A-CDK2. Even at saturating concentrations of p45 or both p45 and p19, we observed no discernible change in the ability of cyclin A-CDK2 to phosphorylate several candidate substrates such as histone H1, the tumor suppressor pRb, or the 34 kDa subunit of the eukaryotic single-stranded binding protein, replication protein A (RPA) (Figure 4A; data not shown). This is consistent with the observation that the cyclin A - C D K 2 - p 9 - p 1 9 - p 4 5 complexes isolated from h u- man cells contained an active kinase (see Figure 3C).

Proteins that are associated with cyclin-CDK complexes could be either regulators or substrates of these kinases. We therefore tested whether p45 or p19 could serve as a substrate for cyclin A-CDK2. Glutathione S-trans- ferase (GST)-p45 fusion protein (Figure 4B, left), but not purified GST (data not shown), was phosphorylated by purified cyclin A-CDK2 (Figure 4B, right). Phosphorylation of p45 was also detected upon addition of [7-32P]ATP to cyclin A immunoprecipitates from human cells (data not


A m -¢-J

>~ Lt3

43 - p45

30 - ~ ~ - CDK2

18 - " "~"~ -p19

1 4 -

- p9

~-p45

B 293 Wl38

v o~ u~ v oh u~ U U Q. t~ U

IP: ~ ~ ~ ~ ~ ~ ~

~-p45 " ' ' ,,,,,, . . . . . p45

c~-CDK2 ",ql , Wil~ ~ - CDK2

C 293

<

9 7 -

6 8 -

- c y c A

43- ' ~ p45

30- - CDK2

18- ~'- '- p19

1 4 -

- p9

~ I I I D ~ - histone H1

Figure 3. Comparison of Cloned p19 and p45 with the Cyclin A-CDK2 Complex-Associated Proteins

(A) (Left) The in vitro translated (IVT) candidate p19 and p45 proteins from the cloned cDNAs were compared with those of p19 and p45 in anti-cyclin A (a-cyc A) immunoprecipitation from 293 cells. (Right) Specificity of affinity- purified p45 antibodies. Total cell lysates were prepared from HeLa cells arrested in S phase by 10 mM hydroxyurea. Approximately 75 p.g of total proteins were loaded onto an SDS-polyacrylamide gel and electrophoresed as indicated. The recombinant p45 protein (approximately 25 ng) was also electrophoresed in parallel. The proteins were immunoblotted with the affinity-purified anti-p45 antibody. (B) Cyclin A-CDK2-p45 complexes were isolated from log phase-growing transformed cells (293) or normal human fibroblasts (Wl38) by immunoprecipitation (IP) with specific cyclin A (cyc A), CDK2, p19, and p45 antibodies, as indicated at the top of the lanes. The proteins were detected by immunoblot using either anti- p45 (upper set of gels) or anti-CDK2 (lower set of gels) antibodies. Double amounts of Wl38 cells were used in the experiment. (C) In vivo p19-p45 complexes. Cell lysates from [~S]methionine-labeled 293 cells were immunoprecipitated with specific cyclin A (cyc A), p45, or p19 antibodies. The immunoprecipitates were analyzed for either protein composition (top) or kinase activity using histone H1 as the substrate (bottom).

A anti-cyclin A B

+ ~.~ 88 + + < <

68-

~ - CyC A

43- ~ - p45

cyc A + CDK2 + pl 9 +

0 N c~

6 8 -

~ ~ - cyc A 9 6 - ~ 43- "~ ~-~- ~ ~':" -p45

~ w , = ~ 68-

30- ~ , ,4 ,= ,=.~,~ - CDK2 43-

30- ~ - "~ " CDK2 18- - ~ "~ - pl 9 30- ~,,

~ - hfstone H1

18-

I R I I , ~ - histone H1

18- "

14. ~-.

I -GST-p45

I - cyc A

Figure 4. p45 and p19 Complex with Cyclin A-CDK2 In Vitro

(A) In vitro formation of cyclin A-CDK2-p19- p45 complex. (Left) Baculovirus-infected insect Sf9 cells were [3SS]methionine labeled. The cell lysates that contained cyclin A (cyc A) and CDK2 were incubated in the absence or presence of p45 as indicated. The protein complexes were immunoprecipitated with anti- cyclin A antibody and assayed for both protein composition (top) and kinase activity (bottom) using histone H1 as the substrate. (Right) ~S- labeled insect Sf9 cell lysates that contain either cyclin A, CDK2, or p19 were mixed. To this mixture, increasing amounts of p45-containing Sf9 cell lysates (0, 5, 10, 20, and 40 p.I) were added as indicated. The protein complexes were immunoprecipitated with anti-cyclin A antibody. Both the protein composition and the histone H1 kinase activity were analyzed. (B) Phosphorylation of p45 by cyclin A-CDK2. (Left) Coomassie staining of the purified GST- p45 protein. The GST-p45 fusion protein was isolated from Sf9 cells infected by a baculovirus encoding the recombinant protein. (Right)

Phosphorylation assay. Active cyclin A (cyc A)-CDK2 was assembled from lysates prepared from baculovirus infected insect Sf9 cells as described in (A). The kinase was isolated by anti-cyclin A immunoprecipitation. Purified GST-p45 protein was added to the purified cyclin A-CDK2 in the presence of [7-32P]ATP. The incorporation of 32p into p45 and cyclin A is shown.

Cell 920

shown). These studies suggest that the biological activity of p45 could be regulated by cyclin A-CDK2. In contrast, there was no detectable phosphorylation of p19 by cyclin A-CDK2 in these assays (data not shown).

p45 Is Essential for Entry into S Phase Since neither p19 nor p45 had a negative effect on cyctin A-CDK2 activity in vitro, it seemed unlikely that either of these proteins would act as an inhibitor of cell proliferation in vivo, We therefore sought to determine whether the presence of p19 or p45 might be required for any aspect of cell cycle progression, particularly for entry into or passage through S phase. We initially chose to interfere with p19 and p45 function by antibody microinjection. For these experiments, HeLa cells were synchronized by mitotic shake-off (Pagano et al., 1992). These cells enter S phase approximately 12-14 hr after replating on gridded cov- erslips. At approximately 6 hr after replating, cells were microinjected with affinity purified p19 or p45 antibodies or with purified GST antibodies as a control. Incorporation of bromodeoxyuridine (BrdU) was used to monitor DNA synthesis.

Injection of cells with GST antibodies had no effect on the percentage of BrdU-positive cells (Figures 5G-51). However, injection of p45 antibodies severely reduced the fraction of cells that underwent DNA synthesis (Figures 5A-5C). This antibody recognized only a single p45 protein in the total HeLa cell lysates (see Figure 3A, right). To test further the specificity of the S phase inhibitory effect, we coinjected cells with p45 antibodies and with an equal molar amount of purified recombinant GST-p45 protein (Figures 5D-5F). This resulted in a complete reversal of the inhibition of DNA synthesis (Figures 5D-5F). Cyclin A-CDK2 function is required both for entry into S phase and for ongoing DNA synthesis. To ask whether the requirement for p45 was similar to the requirement for cyclin A-CDK2, we injected p45 antibodies into cells that were synchronized in S phase. In this case, microinjection had no effect on the incorporation of BrdU (Figures 5J- 5L), suggesting that p45 is required for entry into S phase but not for progression through S phase,

As with the GST antibodies, microinjection of p19 antibodies was without effect (data not shown). Although this is somewhat at odds with the results of the p45 antibody injections, it may not be surprising in light of the fact that the p19 antibodies do not recognize the p19 protein in cyclin A-containing complexes (see Figure 3C).

Although p45 is most abundant in tumor cells, it is also expressed to a low extent in normal cells (see Figure 3B). To determine whether these cells also require p45 function for entry into S phase, we again microinjected purified p45 antibody and measured the effect on DNA synthesis. Normal human fibroblasts were synchronized by serum starvation and were subsequently stimulated by serum addition to reenter the cell cycle. Microinjection of the p45 antibody during G1 essentially abolished DNA synthesis (Figures 5M-50), whereas injection of the control GST antibody had no inhibitory effect (Figures 5P-5R). Again, coinjection of p45 protein with the p45 antibody completely rescued DNA synthesis (data not shown). These results

suggest that p45 is required for execution of S phase in both normal and transformed cells. Furthermore, the ability of p45 antibodies to inhibit DNA synthesis in normal cells strongly argues against the possibility that the p45 antibody exerts its effect simply by interfering with the function of all cyclin A-CDK2 complexes since cyclin A-CDK2 is in vast excess over p45 in normal cells.

As a complement to antibody microinjection, we have also injected synthetic antisense oligonucleotides that were directed against the p45 mRNA (Figure 6; data not shown). Microinjection of p45 antisense oligonucleotides into synchronized G1 HeLa cells prevented DNA synthesis (Figure 6), whereas a control oligonucleotide had no effect. The inclusion of purified p45 protein in the injection mixture completely reversed the inhibition of DNA synthesis (Figure 6).

p45 Expression Is High in Transformed Cells and during S Phase Since the proportion p19-p45-containing cyclin A complexes is increased in transformed cells that have been blocked in S phase by treatment with hydroxyurea (data not shown), we asked whether p45 or p19 expression might vary during the cell cycle. HeLa cells were synchronized at the G1/S boundary by a double thymidine block (Heintz et al., 1983; Pines and Hunter, 1990). Following release from the second block, cells immediately entered S phase. They proceeded synchronously through G2/M after approximately t 0-12 hr and again entered S phase at 24-28 hr following release (Figure 7A). Northern blotting of RNA prepared from synchronous populations revealed that expression of the p45 mRNA was clearly cell cycle dependent (Figure 7A). Levels were high in S phase cells and lower in G1 and G2/M fractions. This pattern of expression roughly paralleled that of cyclin A (Figure 7A). Expres- sion of p19 was also cell cycle dependent, with peak levels in S phase (Figure 7A). However, since pl 9 protein is present in excess of p45, it is unclear whether changes in p19 levels contribute to the cell cycle dependence of p19- p45-containing complexes.

Cyclin A complexes containing p19 and p45 are much more abundant in many transformed cells than in their normal counterparts. To begin to address the mechanisms that underlie these changes in the subunit composition of cyclin A-CDK2, we have examined patterns of p19 and p45 expression (Figure 7B). A comparison of several cell lines indicated that both p19 mRNA and p19 protein were equally abundant in normal and transformed cells. In contrast, both p45 mRNA and p45 protein levels were greatly increased in transformed cell lines. Since the presence of p45 is necessary for binding of p19 to cyclin A-CDK2 complexes (see Figure 4A), it is likely that changes in the abundance of the p45 protein contribute to the rearrangement of cyclin A-CDK2 complexes that accom- pany cellular transformation.

Discussion

The principal characteristic of all transformed cells is their ability to proliferate under conditions that would arrest the

p19 and )45: Essential Elements of Cyclin A-CDK2 921

#,;~. '?

"4 ~,,

Figure 5. p45 Is Essential for S Phase Entry

(A-I) Synchronized G1 phase HeLa cells were microinjected with either the affinity-purified anti-p45 antibody (A-C) or coinjected with anti- p45 antibody plus equal moles of purified p45 protein (D-F) or with equal amount of a control antibody, the anti-GST antibody (G-I). The cells were then labeled with BrdU continuously for another 12-14 hr. (J-L) Synchronized S phase HeLa cells were microinjected with anti-p45 antibody. After 2 h r, the cells were labeled with BrdU for 6 hr. (M-R) Synchronized G1 human normal fibre- blast cells (IMR90) were microinjected with either anti-p45 (M-O) or anti-GST (P-R) antibodies. The cells were subsequently labeled with BrdU for 24 hr. The cells were fixed and processed for the staining of either the injected immunoglobulin G proteins (B, E, H, K, N, and Q)or the incorporation of BrdU in the nuclei with an anti-BrdU antibody (C, F, I, L, O, and R). The phase- contrast photography of injected cells was also presented (A, D, G, J, M, and P).

growth of their normal counterparts. Accumulating evi- dence suggests that loss of proliferation control in tumor cells is accompanied by changes in the regulation of the cyclin-CDK enzymes that drive cell cycle progression (Xiong et al., 1993a). In previous work, we have demonstrated that two major routes of CDK inhibition, those in- volving p16 and p21, are commonly compromised during cellular transformation. In this paper, we have examined changes in the cyclin A-CDK2 complex that also accom- pany cellular transformation.

In normal cells, cyclin A-CDK2 exists predominantly in a quaternary complex with p21 and PCNA (Zhang et al., 1993). In many transformed cells, these are replaced by quinary complexes containing cyclin A, CDK2, p9 cKsl or p9 cKs2, p19, and p45 (Figure 1). This switch may be ex- plained, in part, by the fact that p45 is more abundant in most transformed cells than in their normal precursors (Figures 3 and 7). Although p19 is equally abundant in

~ an~i-GST [ ] ant~-sense

[ ] antJ-p45 [ ] conttO ] Oll90$

[ ] ~nti-p4$ [] ant~-sens~ +~5 +p45

10o-

80-

80£ zo~

sel 4o £ $0-

ZO-

10

0

HeLa Gl-phase HeLa S-phase

i {MR-90 Gl-phase HeLa Gl-phase

Figure 6. Summary of the Microinjection Experiments

Cell 922

A 2 6 10 14 18 22 Z6 30 (hour)

p19 ~ ' ~ ' ~

p 4 5

cyc A m

c0~ oWi~mm,m

1oo 9o go

30

10

0 ~ 10 14 18 ~ Z6

time after release (hours)

B

m

Figure 7. Analysis of p19 and p45 Expression (A) (Top) Cell cycle analysis. HeLa cells were synchronized at the GI/S border by double thymidine block. Total RNAs were prepared at various timepoints after releasing the cells into fresh medium. The RNA blots were probed with either ~P-labeled p19, p45, cyclin A, or CDK2 cDNA as indicated. (Bottom) FACS analysis of synchronized cells. Open squares, S phase; open triangles, G2/M phase; closed circles, G1 phase. (B) Expression of p45 and p19 in transformed and nontransformed human cells. Total RNA was prepared from active growing nontransformed human fibroblasts (Wl38 and IMR90) or transformed cells (A431, ML1, HeLa, VA13, and 293). Equal amounts of RNA were used in each lane. The blots were probed with either s2p-labeled p45 or p19 cDNA.

normal and transformed cells (Figure 7B), the presence of p45 is required for binding of p19 to cyclin A-CDK2 (Figure 4A), thus explaining the concerted appearance of these proteins in cyclin A-CDK2 complexes. We have also found that p45 and p21 compete for binding to the cyclin A-CDK2 enzyme (data not shown). Therefore, the loss of p21 may also contribute to the redistribution of cyclin A-CDK2 into the p45-containing complex that is enriched in transformed cells.

Binding of p19 and p45 to cyclin A-CDK2 has no discernible effect on the ability of the kinase to phosphorylate histone H 1, Rb, or p34 RPA in vitro. However, in vivo ablation of p45 by microinjection of either p45 antibodies or antisense oligonucleotides clearly indicates that the p19- p45-containing complex plays a role in promoting cell cycle progression that is not only essential but is also distinct from that of the cyclin A-CDK2 binary enzyme (Figures 5 and 6). Ablation of p45 in transformed cells prevented DNA synthesis. This was only effective if cells were injected prior to the onset of S phase. In normal cells, p45 and p45-associated cyclin A complexes are quite low in abundance (Figures 1,.3, and 7). Nevertheless, injection

of p45 antibody still prevented DNA synthesis (Figures 5 and 6). In these cells, the binding of p45 to its antibody could affect only a small fraction of total cyclin A-CDK2 enzyme. Therefore, that fraction must have a specific role in the promotion of S phase.

Our studies do not rule out the possibility that p45 plays an essential role in the execution of DNA synthesis that is independent of its interaction with cyclin A-CDK2. How- ever, antibody depletion experiments demonstrate that >90% of the p45 present in 293 cell lysates is bound to cyclin A-CDK2 (data not shown). Considered together with the well-established role of cyclin A-CDK2 in S phase progression, our results argue that the requirement for p45 at the onset of DNA synthesis reflects a requirement for the p19-p45-bound cyclin A-CDK2 enzyme. Our results also indicate that the biological activity of p45 may potentially be regulated by cyclin A-CDK2 through phosphorylation (Figu re 4B). The binding of p45 could facilitate the phosphorylation of p45 by the kinase.

Since p45 contains a leucine-rich repeat motif that is implicated in protein-protein interactions, p45 could function as an adaptor for the cyclin A-CDK2 enzyme. Such a function has been suggested for Grb2, which binds to both the receptor tyrosine kinases and Sos or Shc proteins (Skolnik et al., 1993), and for SDS22, another leucine-rich repeat-containing protein that associates with phosphatase PP1 in S. pombe (Ohkura and Yanagida, 1991). In this scenario, p45 could target the cyclin A-CDK2 to its specific substrates.

To date, the sequence of p19 has provided no clues to the precise mechanism by which p19 functions. A search of the sequence database has revealed extensive homology between p19 and two other published sequences. The first, A39L, shares approximately 50% amino acid identity with human p19. A39L is derived from the Chlorella virus that infects algae (Lu et al., 1995). This virus shares simi- larity to members of poxvirus family and encodes its own DNA replication machinery, including homologs of DNA polymerase ~z and PCNA (which shares only 29% identity with its human counterpart) (Grabherr et al., 1992; J. L. Van Etten, personal communication). The pl 9 protein also shares 65% amino acid identity with a protein, FP21, from Dictyostelium (Kozarov et al., 1995). Since neither the function of A39L nor the function of FP21 is known, these homologies have been somewhat uninformative. A search of expressed sequence tag databases also revealed the existence of highly conserved p19 homologs in variety of other evolutionarily diverse organisms such as Caenorhab- ditis elegans and Arabidopsis thaliana (data not shown). The striking conservation of p19 homologs through evolu- tion and the association of p19 with cyclin-CDK enzymes suggest that these proteins may play a fundamental role in regulating the growth of diverse organisms.

Here we have identified the constituents of the cyclin A-CDK2 complex that predominates not only in transformed cells (VA13, 293, and HeLa) that express viral oncoproteins (SV40 T antigen, adenovirus E1A, and papillo- mavirus E6/E7 proteins), but also in transformed cells that express no viral proteins (e.g., ML1 and A431; Figures 1 and 7). In an effort to identify the mechanism that results


in e leva ted leve ls o f the qu ina ry c o m p l e x in t r a n s f o r m e d

cel ls, w e have e x a m i n e d cycl in A - C D K 2 c o m p l e x e s in hu-

man f i b rob las ts that cons t i tu t i ve ly e x p r e s s pap i l l omav i rus

E6, E7, o r both. The e x p r e s s i o n o f t h e s e viral p ro te ins

has been s h o w n to c a u s e g e n o m i c instabi l i ty , but is not

suf f ic ient to i nduce ce l lu la r t r ans fo rma t i on (Whi te et al.,

1994, and re fe rences there in) . Ou r da ta ind ica ted that

the re was no subs tan t ia l e leva t ion o f the cyc l in A - C D K 2 -

p 9 - p 1 9 - p 4 5 c o m p l e x (da ta not shown) in t hese cells.

There fore , addi t ional t rans fo rmat ion-assoc ia ted events are

requ i red for the i nc reased level o f this c o m p l e x in t rans-

f o rmed cells.

The p 1 9 - p 4 5 - a s s o c i a t e d e n z y m e is requ i red for en t ry

into S p h a s e bo th in no rma l ce l ls and in t r a n s f o r m e d cells.

T h e i nc reased a b u n d a n c e of the p45 -con ta in ing com-

p lexes in t r a n s f o r m e d cel ls cou ld imply a par t ia l de regu la -

t ion of en t ry into S phase . In this way, the e n h a n c e d ex-

p ress ion o f p45 in t u m o r cel ls may con t r i bu te to the

p rocess o f ce l lu la r t rans fo rma t ion . Also, the abi l i ty to exe -

cu te S p h a s e u n d e r i napp rop r i a te cond i t i ons cou ld pro-

mo te the ka ryo typ ic instabi l i ty that is a c o m m o n fea tu re

of the t r a n s f o r m e d pheno type .

Experimental Procedures

Cells Human diploid cells of lung fibroblasts WI38 and IMR90, foreskin fibroblast Hs68, and an SV40-transformed human fibroblast VA13 were obtained from American Type Cultu re Collection, The nontransformed cells were used between passages 13-23. HeLa, 293, ML1, A431, and other human cells used in the experiments were obtained from cell culture facilities at Cold Spring Harbor Laboratory. The cells were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum. The culture of insect Sf9 cells and the production of baculoviruses that encode various proteins used in the experiments were as described previously (7hang et al., 1994).

Antibodies Antibodies raised against various cyclins and CDKs were described previously (Xiong et al., 1993a; Zhang et al., 1993). For the production of anti-p45 and anti-p19 antibodies, GST-p45 or GST-p19 (the fusion proteins between the GST and the full-length human p45 or p19) was purified from bacteria and used to immunize rabbits. For the affinity purifications of anti-p45 antibody, MBP-p45 (a fusion protein between the maltose-binding protein and p45) was purified on amylose-aga- rose beads (New England Biolabs). The MBP-p45 was covalently cross-linked to CNBr-activated Sepharose 4B beads (Bethesda Re- search Laboratories) and used as the affinity medium. The p45 antibody was affinity purified as described previously (Xiong et al., 1992). The affinity purification of anti-p19, anti-GST, or anti-cyclin A antibodies was performed similarly using the bacterially produced recombinant proteins as the antigens. The affinity-purified antibodies were concentrated by Centricon concentrators (Amicon).

Purification of Cyclin A-CDK2 Complex For the purification of cyclin A-CDK2 complex, 2 mg of affinity-purified anti-cyclin A antibody was loaded onto 1 ml of protein A beads (Pierce). After washing the column extensively with 0.1 M triethanolamine (pH 8.3) (Sigma), the antibody was cross-linked to protein A with 50 mM dimethyl pimelimidate (Pierce) in 0.1 M triethanolamine (pH 8.3) at room temperature for 4 hr. The reaction was terminated by 0.1 M ethanolamine (pH 9) (Sigma) at room temperature for 2 hr.

Actively growing suspension 293 cells, a human embryo kidney cell that was transformed by adenovirus DNA, were cultured in DMEM supplemented with 5% calf serum to 1 x 106 cells per milliliter. Typi- cally, 40 liters of 293 cells were pelleted and washed three times with phosphate-buffered saline. The cells were then lysed with 200 ml of lysis buffer (0.5% NP-40, 50 mM Tris [pH 7.4], 150 mM NaCI, 100 mM

NaF, 1 mM NaVO4, and 0.5 mM PMSF) and 5 p.g/ml each of aprotinin, leupeptin, soybean trypsin inhibitor, and 100 mM benzamidine (all from Sigma). The lysate was clarified by centrifugation at 18,000 x g and was then loaded onto the anti-cyclin A immunoaffinity column. The bound proteins were released in the Laemmli sample buffer and separated in an SDS-polyacrylamide gel. The protein bands were visualized by staining with 0.05% Coomassie brilliant blue G (Sigma). Each individual protein band that corresponded to p19 or p45 protein was isolated and microsequenced as described previously (Xiong et al., 1993b).

Isolation of cDNA Clones Encoding p19 and p45 Microsequencing of p19-derived peptides produced several peptide sequences: K7, ENQw?EEK; K9, TFNIK; K10, NDFTEEEEAQVRK; K19, ?MLEDL?Mdd (small letter and question mark indicate the uncer- tainty of the sequence). A degenerating oligonucleotide, 5'-CGAATTCA- A(CT)GC(CT)TT(CT)AC(ACGr)GA(AG)GA(AG)GA(AG)GA(AG)GC-3', deduced from the sequence NDFTEEEEA of K10, was synthesized, Th is oligon ucleotide and a 20-mer of oligo(dT) specific for the 3' poly(A) tail were used as the primers for polymerase chain reaction (PCR) using a HeLa cDNA as the templates (Strategene), A 400 bp PCR product that contained both K10 and K7 amino acid sequences and the 3' untranslated region of the p19 cDNA was obtained. This DNA fragment was used to screen a HeLa ~.ZAP cDNA library (Strategene) for the full-length p19 cDNA.

For the cloning of p45, the following peptide sequences were obtained from the p45-derived peptides: K12, ?FVIVRRPK; K13, ?LQEIP- DLSSNVATSF; K15, ??ELLSGMG?SALEK; K19, EEPDSENIPQEL- LSN; K24, ?LQVFglVP. Two oligonucleotides, 5'-CGAATrC(ACGT) CA(AG)AT(ACT)CC(ACGT)GA(CT)(CF)T-3', which was deduced from the sequence LQEIPDL of K13, and 5'-CGAATTC(CT)TC(CT)TG (ACGT)GG(TGA)AT(GA)TT(CF)TC-3', which was derived from the antisense direction of peptide sequence ENIPQE of K19, were synthesized. A DNA fragment that had a 147 bp insertion was obtained in PCR using the HeLa cDNA as the templates. This PCR product contained the known amino acid sequences of K13, K15, and K19 and was used subsequently as the probe to isolate the full-length p45 cD NA from the HeLa cDNA library.

Immunological Procedures and Kinase Assays The immunoprecipitation, Western blot, kinase assays using histone H1 or other potential substrates, and glycerol gradient analysis were conducted as described previously (Zhang et al., 1994).

Reconstitution Assay Insect Sf9 cells were infected with individual baculoviruses that encodes either cyclin A, CDK2, p19, p45, or p9 cKs7 or p9 cKs2. The cell lysates were prepared from the [~S]methionine-iabeled infected cells as described previously (Zhang et al., 1994). For a typical reconstitution reaction, 5 I~1 of cyclin A and 5 p.I of CDK2-containing lysates were mixed with 0-100 p.I of a lysate that contained either p19, p45, or both and incubated at 30°C in the presence of ATP as described previously (7hang et al., 1994). After the reaction, the protein complexes were isolated by immunoprecipitation and analyzed for both protein composition or kinase activity using histone H1 as the substrate.

RNA Analysis and Cell Cycle Analysis HeLa cells were synchronized at the GIlS border by the double thymidine arrest procedure (Heintz et al., 1983; Pines and Hunter, 1990). Cell cycle analysis using fluorescence-activated cell sorting (FACS) was conducted as described previously (Xiong et al., 1993b). Northern blot analysis using total RNA was performed as described by Pines and Hunter (1990).

Microinjection of Cells Microinjection experiments were conducted using the Zeiss microinjection apparatus that contains the 5242 microinjector and the 5170 micromanipulator (Eppendort) mounted on a Zeiss Aviovert-10 micro- scope. The femtotips were from Eppendorf (catalog number 952.008). All the affinity-purified antibodies were adjusted to 4 mg/ml in 100 mM glycine (pH 7~2). Synchronized HeLa cells were obtained by mitotic shake-off (Pagano et al., 1992). HeLa cells were treated with 40 ng/ ml nocodazole for 12 hr. The mitotic shake-off cells were removed of

Cell 924

the drug and were seeded onto a gridded glass coverslip (Bellco) in fresh medium. Approximately 70%-80% of cells enter S phase synchronously at 13-15 hr. For G1 phase HeLa cells, antibodies were microinjected into the cytoplasm of the cells at 6 hr postreplating, and the cells were labeled continuously with 10 p.M BrdU (Amersham). Typically, approximately 150-200 cells were injected for each sample, and each experiment was repeated at least three times independently. At 18-22 hr postreplating, the cells were fixed and immunostained for the injected IgG using Texas red-conjugated goat anti-rabbit IgG antibody (Cappel) and the FITC-conjugated mouse anti-BrdU antibody (Becton Dickinson) according to a method described by Leonhardt et al. (1992). The percentage of injected BrdU-positive cells were normal- ized with that of surrounding cells that were not injected.

For S phase injection, the mitotic shake-off HeLa cells were replated in the presence of 10 mM hydroxyurea. At 22 hr postreplating, antibodies were microinjected into the S phase cells. Hydroxyurea was then removed from the medium. Cells were labeled 2 hr later with 10 mM BrdU for another 6 hr and then fixed for immunostaining.

For the competition experiment, a baculovirus that is programmed to produce a GST-p45 fusion protein was constructed. The fusion protein was harvested from baculovirus-infected insect Sf9 cells and was purified by glutathione-Sepharose beads (Zhang et al., 1994). The purified protein was >95O/o pure (Figure 4B, left) and was concentrated by Centricon concentrators (Amicon). The purified GST-p45 was mixed, prior to injection, at 1:1 molar ratio with anti-p45 antibody. As a control, the antibody was mixed with the same glutathione solution that was used to elute GST-p45 protein. For the injection of nontransformed cells, low passage (10-14 passages) IMR90 human fibroblast cells were cultured to 70%-80% confluency. The cells were then serum-starved in DMEM without serum for 60 hr. The cells were stimulated to reenter the cell cycle by addition of 20% fetal bovine serum. Approximately 65°/0-70% cells reentered the cell cycle. The cells were microinjected at 6-8 hr poststimulation and then continuously labeled with BrdU for 18-20 hr. The cells were fixed and proceeded for immunostaining analysis as described above.

For microinjection of antisense oligonucleotides, an antisense (25- mer) phosphorothioate oligonucleotide (S-oligonucleotide) directed against p45 RNA was designed (Oligos Etc.). The anti-p45 antisense oligonucleotide has sequence 5'-TCCTGTGCATAGCGTCCGCAG- GCCC-3', which corresponds to p45-encoding nucleotide sequence 22-46. As a control, an S-oligonucleotide that had the same nucleotide composition as the p45 antisense oligonucleotide but with reverse orientation was used. The S-oligonucleotides were dissolved in 0.1 M glycine (pH 7.2) for microinjection. Synchronized G1 HeLa cells were microinjected at 6-8 hr postmitotic shake-off with 25 p.M S-oligonucleotides. To mark the injected cells, 1 mg/ml purified rabbit IgG protein was coinjected with the oligonucleotides. After microinjection, the cells were processed as described above.

Acknowledgments

Correspondence should be addressed to D. B. We thank Drs. Gregory J. Hannon and Maureen Caligiuri for critical reading of the manuscript and Hong Sun for help in setting up the microinjection experiments. We thank Mike Ockler, Jim Duffy, and Phil Renna for preparation of the figures. H. Z. is a postdoctoral fellow of the Howard Hughes Medical Institute, and D. B. is an Investigator of the Howard Hughes Medical Institute. This work was supported in part by grants from the National Institutes of Health (GM39620-07 to B. D. and NIH5P30-CA45508-08 to R. K.).

Received April 19, 1995; revised July 25, 1995.

References

Cardoso, M. C., Leonhardt, H., and NadaI-Girard, B. (1993). Reversal of terminal differentiation and control of DNA replication: cyclin A and Cdk2 specifically localize at subnuclear sites of DNA replication. Cell 74, 979-992. Dulib, V., Kaufman, W. K., Wilson, S. J., Tlsty, T. D., Lees, E., Harper, J. W., Elledge, S. J., and Reed, S. I. (1994). p53-dependent inhibition of cyclin-dependent kinase activities in human flbroblasts during radia- tion-induced G1 arrest. Cell 76, 1013-1023.

Dunphy, W. G., (1994). The decision to enter mitosis. Trends Cell Biol. 4, 202-207.

El-Deiry, W. S., Tokino, T., Velculescu, V. E., Levy, D. B., Parsons, R., Trent, J. M., Lin, D., Mercer, W. E., Kinzler, K. W., and Vogelstein, B. (1993). WAF1, a potential mediator of p53 tumor suppression. Cell 75, 817-825.

EI-Deiry, W. S., Harper, J. W., O'Connor, P. M., Velculescu, V. E., Canman, C. E., Jackman, J., Pietenpol, J. A., Burrell, M., Hill, D. E., Wang, Y., Wiman, K. G., Mercer, W. E., Kastan, M. B., Kohn, K. W., EUedge, S. J., Kinzler, K., W., and Vogelstein, B. (1994). WAF1/CIP1 is induced in p53-mediated G1 arrest and apoptosis. Cancer Res. 54, 1169-1174. Girard, F., Strausfeld, U., Fernandez, A., and Lamb, N. (1991). Cyclin A is required for the onset of DNA replication in mammalian fibroblasts. Cell 67, 1169-1179. Grabherr, R., Strasser, P., and Van Etten, J. L. (1992). The DNA polymerase gene from Chlore//a virus PBCV-1 and NY-2A contains an intron with nuclear splicing sequences. Virology 188, 721-731.

Guadagno, T. M., Ohtsubo, M., Roberts, J. M., and Assoian, R. K. (1993). A link between cyclin A expression and adhesion-dependent cell cycle progression. Science 262, 1572-1575.

Harper, J. W., Adami, G. R., Wei, N., Keyomarsi, J., and Elledge, S. J. (1993). The p21 CDK-interacting protein Cipl is a potent inhibitor of G1 cyclin-dependent kinases. Cell 75, 805-816.

Heichman, K., and Roberts, J. M. (1994). Rules to replicate by. Cell 79, 557-562.

Heintz, H., Sive, H. L., and Roeder, R. G. (1983). Regulation of human histone gene expression: kinetics of accumulation and the changes in the rate of synthesis and in the half-lives of individual histone mRNAs during the HeLa cell cycle. Mol. Cell. Biol. 3, 539-550.

Hunter, T., and Pines, J. (1994). Cyclins and cancer I1: cyclin D and CDK inhibitors come of age. Cell 79, 573-582.

Kataoka, T., and Wigler, M. (1985). DNA sequence and characterization of the S. cerevisiae gene encoding adenylate cyclase. Cell 43, 493-505.

Kobe, B., and De isenhofer, J. (1994). The leucine-rich repeat: a versa- tile binding motif. Trends Biol. Sci. 19, 415-421.

Kozarov, E., Van der Wel, H., Field, M., Gritzali, M., Brown, Jr. R. D., and West, C. M. (1995). Characterization of FP21, a cytosolic glycopro- tein from Dictyostelium. J. Biol. Chem. 270, 3022-3030.

Leonhardt, H., Page, A. W., Weier, H.-U., and Bestor, T. H. (1992). A targeting sequence directs DNA methyltransferase to sites of DNA replication in mammalian nuclei. Cell 71,865-873.

Lu, Z., Li, Y., Zhang, Y., Kutish, G. F., Rock, D. L., and Van Etten, J. L. (1995). Analysis of 45 kb of DNA located at the left end of the Chlorella virus PBCV-1 genome. Virology 206, in press.

Matsushime, H., Ewen, M., Strom, D. K., Kato, J., Hanks, S. K., Rous- sel, M. F, and Sherr, C. J. (1992). Identification and properties of an atypical catalytic subunit (p34 psK~'~4) for mammalian D type G1 cyclins. Cell 71,323-334.

Matsushime, H., Quelle, D. E., Shurtleff, S. A., Shibuya, M., Sherr, C. J., and Kato, J.-Y. (1994). D-type cyclin-dependent kinase activity in mammalian cells. Mol. Cell. Biol. 14, 2066-2076.

Meyerson, M., and Harlow, E. (1994). Identification of G1 kinase activity for cdk6, a novel cyclin D partner. MoI. Cell. Biol. 14, 2077-2086.

Ohkura, H., and Yanagida, M. (1991). S. pombe gene sds22 + essential for a midmitotic transition encodes a leucine-rich repeat protein that positively modulates protein phosphatase-l. Cell 64, 149-157.

Pagano, M., Pepperkok, R., Verde, F., Ansorge, W., and Draetta, G. (1992). Cyclin A is required at two points in the human cell cycle. EMBO J. 11,961-971.

Perozzi, G., and Prakash, S. (1986). RAD7 gene of Saccharomyces cerevisiae: transcripts, nucleotide sequence analysis, and functional relationship between the RAD7 and RAD23 gene products. Mol. Cell. Biol. 6, 1497-1507.

Pines, J., and Hunter, T. (1990). Human cyclin A is an adenovirus E1A-associated protein p60 and behaves differently from cyclin B. Nature 346, 760-763.


Richardson, H. E., Stueland, C. S., Thomas, J., Russell, P., and Reed, S. I. (1990). Human cDNA encoding homologs of the small p34 cdc2~cdc2- associated protein of Saccheromyces cerevisiae and Schizosaccharo- myces pombe. Genes Dev. 4, 1332-1344.

Serrano, M., Hannon, G. J., and Beach, D. (1993). A new regulatory motif in cell-cycle control causing specific inhibition of cyclin D/CDK4. Nature 366, 704-707. Serrano, M., Gomez-Lahoz, E., Depinho, R. A., Beach, D., and Bar- Sagi, D. (1994). Inhibition of ras-induced proliferation and cellular transformation by p16 'NK4. Science 267, 249-262. Sherr, C. J. (1993). Mammalian G1 cyclins. Cell 73, 1059-1065. Sherr, C. J. (1994). G1 phase progression: cycling on cue. Cell 79, 551-555. Skolnik, E. Y., Batzer, A., Li, N., Lee, C. H., Lowenstein, E., Moham- madi, M., Margolis, B., and Schlessinger, J. (1993). The function of GRB2 in linking the insulin receptor to Ras signaling pathways. Sci- ence 260, 1953-1955.

Solomon, M. (1993). Activation of the various cyclin/cdc2 protein kinases. Curr. Opin. Cell Biol. 5, 180-186.

White, A. E., Livanos, E. M., and Tlsty, T. D. (1994). Differential disrup- tion of genomic integraty and cell cycle regulation in normal human fibroblasts by the HPV oncoproteins. Genes Dev. 8, 666-677. Xiong, Y., Zhang, H., and Beach, D. (1992). D type cyclins associate with multiple protein kinases and the DNA replication and repair factor PCNA. Cell 71,505-514. Xiong, Y., Zhang, H., and Beach, D. (1993a). Subunit rearrangement of the cyclin-dependent kinases is associated with cellular transformation. Genes Dev. 7, 1572-1583 Xiong, Y., Harmon, G. J., Zhang, H., Casso, D., Kobayashi, R., and Beach, D. (1993b). p21 is a universal inhibitor of cyclin kinases. Nature 366, 701-704. Zhang, H., Xiong, Y., and Beach, D. (1993). Proliferating cell nuclear antigen and p21 are components of multiple cell cycle kinase complexes. Mol. Biol. Cell 4, 897-906. Zhang, H., Hannon, G. J., and Beach, D. (1994). p21-containing cyclin kinases exist in both active and inactive states. Genes Dev. 8, 1750- 1758.

GenBank Accession Numbers

The accession number for p19 is U33760 and for p45 is U33761.

p19 skpl and p45 skp2 are essential elements of the cyclin a-cdk2

Documents