Oncogene (1999) 18, 5423 ± 5434 ã 1999 Stockton Press All rights reserved 0950 ± 9232/99 $15.00 http://www.stockton-press.co.uk/onc Functional evaluation of tumour-speci®c variants of p16INK4a/CDKN2A: correlation with protein structure information

Margarida Ruas1, Sharon Brookes1, Neil Q McDonald1 and Gordon Peters*,1

1Imperial Cancer Research Fund, PO Box 123, 44 Lincoln's Inn Fields, London WC2A 3PX

Inherited mutations in the CDKN2A/INK4a/MTS1 may harbour CDKN2A mutations, a proper under- tumour suppressor gene on chromosome 9p21 are standing of the impact of these mutations is desirable associated with familial predisposition to melanoma and (Dracopoli and Fountain, 1996; Hayward, 1996). other tumour types. Nonsense and missense mutations The major challenge is to distinguish between are also found in a variety of sporadic cancers, and over innocuous polymorphisms and real germline muta- 140 sequence variants have already been recorded in the tions that impair the function of CDKN2A to di€erent literature. In assessing the relevance of these variants degrees. Moreover, since the sporadic mutations and for counselling members of a€ected families, it is described thus far have invariably been found by important to distinguish inactivating mutations from analysing tumour cells, those regarded as polymorph- harmless polymorphisms. Existing functional assays have isms, because of their presence in adjacent una€ected frequently reached con¯icting conclusions and no single tissue, could represent rare germline variants that test appears adequate. Here we evaluate a number of predispose to cancer. Conversely, it is possible that alternatives including a novel assay based on retroviral the sequence variants of CDKN2A recorded in sporadic delivery of p16INK4a cDNAs into human diploid ®bro- tumours may re¯ect a mutator phenotype caused by a blasts. Among the 17 sequence variants analysed, three lesion in DNA repair or `caretaker' genes (Kinzler and distinct categories can be distinguished: those that Vogelstein, 1997), and may not necessarily a€ect the abrogate the binding of p16INK4a to CDK4 and CDK6, function of the protein. The prevalent use of PCR in those that alter the properties of the protein without the analysis of CDKN2A gene sequences also raises the preventing it from interacting with CDKs, and those that possibility that some recorded mutations are PCR have no discernible e€ect on protein function. These artefacts. distinctions can be rationalized by considering the impact The product of the CDKN2A gene is a 16 kDa of the amino acid changes on the three-dimensional protein, p16INK4a, comprising four ankyrin-type repeat structure of the protein. motifs, that binds speci®cally to and blocks the function of the cyclin dependent kinases CDK4 and Keywords: cyclin-dependent kinase inhibitor; tumour CDK6 (Serrano et al., 1993). When associated with D- suppressor; familial melanoma; ankyrin repeats; type cyclins, CDK4 and CDK6 promote passage replicative senescence through the G1 phase of the cell cycle by contributing to the phosphorylation and functional inactivation of the retinoblastoma gene product, pRb (reviewed in Sherr, 1994; Weinberg, 1995). Thus, p16INK4a and other Introduction members of the INK4 family (p15INK4b, p18INK4c and p19INK4d) are all capable of imposing a G1 cell cycle The CDKN2A/INK4a/MTS1 tumour suppressor gene arrest that is dependent on the presence of functional on human chromosome 9p21 is a frequent target of pRb (Sherr and Roberts, 1995; Ruas and Peters, 1998). inactivating mutations in both sporadic and familial All four proteins bind directly to CDK4 and CDK6 in cancers (reviewed in Ruas and Peters, 1998). For vitro but, despite their biochemical and structural example, the gene sustains homozygous deletions in similarities, only p16INK4a has the credentials of a 40 ± 60% of bladder cancers, gliomas and T-cell acute tumour suppressor. A likely explanation is that lymphoblastic leukaemias (T-ALL), is silenced by p16INK4a accumulates as cells reach the end of their hypermethylation of the promoter region in colon, replicative lifespan (senescence) so that inactivation of breast and non-small cell lung cancers (NSCLC), and p16INK4a favours the emergence of immortal cell clones sustains missense and nonsense mutations in a (Alcorta et al., 1996; Hara et al., 1996; Loughran et al., signi®cant proportion of biliary tract, pancreatic and 1996; Rezniko€ et al., 1996; Yeager et al., 1998). oesophageal cancers (Ruas and Peters, 1998). Over 140 The known properties of p16INK4a have been single amino acid changes have already been noted in exploited in the development of functional assays, human tumour cells, a€ecting 95 of the 156 residues in which fall broadly into three types: (i) the ability of the protein. Germline mutations are associated with p16INK4a to bind directly to CDK4 and CDK6; (ii) its inherited prediposition to melanoma and other cancers ability to inhibit the phosphorylation of pRb by cyclin (Ruas and Peters, 1998) and since it is estimated that D-CDK complexes; and (iii) its ability to cause cell approximately 10% of all melanomas have a familial cycle arrest when ectopically expressed. Only a component and as many as half of the familial cases minority of the known missense variants have been functionally analysed and although the severe loss-of- function mutants are easily detected, a substantial *Correspondence: G Peters Received 25 January 1999; revised 12 March 1999; accepted 15 April proportion of the variants tested thus far behave 1999 ambiguously in di€erent assay systems or have given Function and structure of p16 mutants MRuaset al 5424 inconsistent results when analysed by di€erent labora- in a simple in vitro association assay (Parry and Peters, tories (Koh et al., 1995; Lukas et al., 1995; Ranade et 1996). The di€erent single amino acid changes were al., 1995; Reymond and Brent, 1995; Shapiro et al., introduced in the wild-type p16INK4a cDNA sequence by 1995; Wick et al., 1995; Yang et al., 1995; Enders et al., site-directed mutagenesis and the proteins were 1996; Lilischkis et al., 1996; Parry and Peters, 1996; synthesized by coupled transcription and translation Tevelev et al., 1996; Yarbrough et al., 1996; Zhang and in rabbit reticulocyte lysates. Binding assays were Peng, 1996; Arap et al., 1997; Harland et al., 1997; Sun performed by mixing the p16INK4a variants with et al., 1997). There are a number of possible reasons. CDK4 or CDK6 produced by in vitro translation in The ®rst is that any assay that requires the production the presence of [35S]methionine. The mixtures were of recombinant p16INK4a in bacteria is vulnerable to immunoprecipitated with antiserum against full length aggregation of the protein (Boice and Fairman, 1996; p16INK4a (Parry and Peters, 1996) and the amount of Tevelev et al., 1996; Zhang and Peng, 1996) and to labelled CDK that co-precipitated with p16INK4a was variability caused by having to recover the protein assessed by SDS ± PAGE (Figure 1). In preliminary from insoluble inclusion bodies. A second point is that tests it was con®rmed that the variant p16INK4a proteins any assay conducted in vitro, such as CDK binding or were synthesized and immunoprecipitated with similar kinase inhibition, may not be a true re¯ection of the in eciency (data not shown). vivo situation. For example, it is now clear that cyclin D-dependent kinase activity is associated with high molecular weight complexes in vivo and that D cyclins are rarely if ever found in simple binary associations with CDK4 or CDK6 (Della Ragione et al., 1996; Mahony et al., 1998; Musgrove et al., 1998; McConnell et al., 1999). Similarly, although p16INK4a does participate in binary complexes with CDKs, it does not associate detectably with D cyclins in vivo (Hall et al., 1995; Parry et al., 1995; Guan et al., 1996; McConnell et al., 1999). This could in part explain why the ability of p16INK4a to inhibit cyclin-CDK complexes assembled using baculovirus driven expres- sion in insect cells does not give a linear dose response (for example, see Serrano et al., 1993; Koh et al., 1995; Lukas et al., 1995; Ranade et al., 1995; Parry and Peters, 1996). Finally, virtually all assays based on the ectopic expression of p16INK4a in eukaryotic cells are vulnerable to variations in the levels of expression achieved. One of the most frequently used is based on the co-transfection of p16INK4a cDNA with a cell surface marker so that the proportion of transiently transfected cells arrested in G1 can be measured by ¯ow cytometry (Koh et al., 1995; Lukas et al., 1995; Parry and Peters, 1996). Assays of this type do not lend themselves to quantitative appraisal of variants that may be only partially or conditionally impaired. Because of these problems, we have sought to devise functional tests that are relatively simple to perform and interpret, and that can detect partially as well as fully impaired variants of p16INK4a. Here we compare the ecacy of established and novel assays by testing a series of sporadic and familial variants of p16INK4a, several of which have given ambiguous results in previous studies. Two assays emerge as the most reliable indicators of p16INK4a function: binding to CDK4 and CDK6 in vitro and measurements of cell proliferation following infection with retroviruses encoding p16INK4a. Our con®dence in these assays is supported by relating the observed properties of each variant to the predicted e€ects of the amino acid substitution on the three- dimensional structure of p16INK4a. Figure 1 Binding of p16INK4a variants to CDKs in vitro. The ability of p16INK4a variants to bind to CDK4 and CDK6 was tested by mixing unlabelled p16INK4a with 35S-labelled CDKs, Results produced by in vitro translation in reticulocyte lysates. The mixtures were immunoprecipitated with antiserum against p16INK4a and the co-precipitated CDKs were analysed by SDS ± Interaction of p16INK4a variants with CDKs in vitro PAGE in a 12% polyacrylamide gel and visualized by autoradiography. Load shows the direct analysis of the input The most direct functional test for INK4 proteins is CDKs translated in vitro and Vector corresponds to a mock whether they can bind eciently to CDK4 and CDK6 binding assay using the empty pRSETA plasmid Function and structure of p16 mutants MRuaset al 5425 Under the conditions used, the A20P, D84H, D84N and a non-binding control (A20P) were produced as and D84Y variants showed no detectable binding to recombinant proteins in bacteria and puri®ed via the

CDK4 or CDK6, while there was very little residual His6 tag. Preliminary titration experiments were carried binding with L16P, P48L, E88K and InsR113 (see out to determine the linear range for the quantitative Table 1 for a description of these mutants). The co-precipitation of CDK4. Since p16INK4a is prone to remaining variants, N71S, V95A, H98Y, G122S, aggregation, the appropriate amount can vary con- A127S, A132P, were indistinguishable from wild-type siderably depending on the quality and age of the p16INK4a in their capacity to interact with CDK4 and protein preparations. As illustrated in Figure 2, the CDK6. These data are summarized in Table 1 along G101W mutant showed a progressive reduction in with corresponding results for three additional variants CDK binding when the reaction was performed at that were analysed in previous studies. Thus, M53I increasing temperatures (Parry and Peters, 1996) failed to bind to CDK4 and CDK6 whereas G101W whereas wild-type p16INK4a was largely una€ected. In and A118T retained binding (Parry and Peters, 1996; our experience, it is essential to consider the trend Harland et al., 1997). This meant that around half (8/ across a range of temperatures since even the wild-type 17) of the variants under consideration here appeared protein can occasionally show reduced binding at 408C, functionally normal in the in vitro assay but only as evident here, which can be o€set by an increase in A127S was suspected of being a polymorphism (see non-speci®c background (Parry and Peters, 1996). Ruas and Peters, 1998). Signi®cantly, three of the Thus, although several of the other variants showed variants in this category, N71S, G101W and A118T, some decrease in binding at higher temperatures (see segregate with familial melanoma (Hussussian et al., Table 1), only N71S and A118T showed a progressive 1994; Kamb et al., 1994; Harland et al., 1997). decline in binding as the temperature was increased (Figure 2). It should be emphasized that neither of these variants is as convincingly ts as the G101W Conditional binding of p16INK4a variants to CDK4 mutant. As we have previously shown that the G101W mutant is temperature sensitive (ts) in a variety of functional Ectopic expression of p16INK4a variants in TIG-ER cells tests (Parry and Peters, 1996), it was of interest to establish whether any of the current panel of variants There were still variants that remained indistinguish- showed a ts phenotype. The variants that retained able from wild-type in the in vitro binding assay. An CDK binding activity, together with wild-type p16INK4a obvious source of concern was that the interaction

Table 1 List of mutants tested and functional conclusions Residual Cdk binding Growth AA change Base change Tumour type binding at 378C arrest Reference L16P CTG4CCG familial melanoma/bile duct 7 ND 7 (Yoshida et al., 1995; Sou®r et al., 1998) A20P GCG4CCG lung cell line 7 ND 7 (Liu et al., 1995) P48L CCG4CTG familial melanoma/ 7 ND 7 (Zhang et al., 1994; Liu et al., 1995; Pollock et al., CCG4TTG melanoma/HNSCC 1995;Platz et al., 1997) M53I ATG4ATC familial melanoma 7 ND + (Walker et al., 1995; FitzGerald et al., 1996; Flores et al., 1997; Harland et al., 1997; Sun et al., 1997; Sou®r et al., 1998) N71S AAC4AGC familial melanoma + 87+12 + (Hussussian et al., 1994) D84Y GAC4TAC NSCLC/prostate line 7 ND ND (Hayashi et al., 1994; Komiya et al., 1995; Liu et al., 1995; Tamimiet al., 1996) D84N GAC4AAC NSCLC/esophagus/HNSCC 7 ND ND (Hayashi et al., 1994; Mori et al., 1994; Zhang et al., 1994) D84H GAC4CAC NSCLC 7 ND ND (Hayashi et al., 1994) E88K GAG4AAG melanoma 7 ND + (Ohta et al., 1994; Pollock et al., 1995) V95A GTG4GCG NSCLC + 96+11 + (Hayashi et al., 1994) H98Y CAC4TAC glioma + 113+19 + (Kyritsis et al., 1996) G101W GGG4TGG familial melanoma + 66+4 7 (Hussussian et al., 1994; Kamb et al., 1994; Whelan et al., 1995; Sou®r et al., 1998) InsR113 Ins 3 bp familial melanoma 7 ND ND (Borg et al., 1996; Platz et al., 1997) A118T GCT4ACT familial melanoma + 88+12 7 (Harland et al., 1997) G122S GGC4AGC ampillary tumour + 95+13 + (Yoshida et al., 1995) A127S GCA4TCA esophageal/bladder/lung/ + 138+26 + (Okamoto et al., 1994; Gonzalez-Zulueta et al., 1995; germline Suzuki et al., 1995) A132P GCG4CCG NSCLC + 92+11 + (Hayashi et al., 1994; Marchetti et al., 1997)

The variant forms of p16INK4a are identi®ed by the residue number ¯anked by the wild-type and variant amino acids in single letter code. Thus, L16P refers to the substitution of proline for leucine at position 16. In InsR113 an additional arginine codon has been inserted between R112 and L113. The corresponding nucleotide alterations are indicated under Base change. CDK binding refers to the data illustrated in Figure 1 and in references (Parry and Peters, 1996; Harland et al., 1997). Residual CDK4 binding activity at 378C is expressed as a percentage of the binding activity at 308C and was calculated from the data shown in Figure 2. The values represent the averages from at least three independent experiments. Growth arrest was determined from the proliferation of retrovirally infected TIG-ER cells as described in Figure 5. + means indistinguishable from wild-type, 7 means indistingushable from the vector only control and + indicates an intermediate e€ect, ND, not determined. NSCLC, non-small cell lung cancer. HNSCC, head and neck squamous cell carcinoma Function and structure of p16 mutants MRuaset al 5426

INK4a INK4a Figure 2 Temperature sensitivity of p16 variants. His6-tagged p16 proteins were expressed in E. coli and puri®ed on nickel-charged Sepharose. 7.5 ng of each preparation was assayed for binding to 35S-labelled CDK4 at the indicated temperatures. The labelled CDK4 co-precipitating with p16INK4a was analysed by SDS ± PAGE and visualized by autoradiography (a). The amount of CDK4 in each precipitate was quanti®ed by phosphorimaging and expressed as a percentage of the input CDK4 (b). Note that whereas a single representative assay is shown in (a), the quantitative data in (b) were calculated from triplicate assays

between individual components in vitro might not viruses were produced by transient transfection of the accurately re¯ect the in vivo properties of the BOSC-23 packaging cell line, and the supernatants proteins. To try to address this possibility, we were added directly to so called TIG-ER cells, the explored alternative assays based on ectopic expres- TIG-3 strain of human ®broblasts engineered to sion of p16INK4a sequences in eukaryotic cells. Several express the receptor for ecotropic retroviruses procedures have been described that use two types of (Serrano et al., 1997; McConnell et al., 1998). readout: the ability of p16INK4a to cause a G1 cell cycle Infection eciencies were between 50 and 80% as arrest in transfected cells (Koh et al., 1995; Lukas et judged by the percentage of cells surviving selection in al., 1995; Shapiro et al., 1995; Parry and Peters, 1996; puromycin. The levels of ectopic p16INK4a RNA Arap et al., 1997) and the interaction of p16INK4a with expression in the di€erent cell pools were compared either endogenous or exogenous CDKs (Reymond and at 6 days post infection by Northern blot analysis using Brent, 1995; Shapiro et al., 1995; Yang et al., 1995; a probe speci®c for the HA-tag (Figure 3a). Relative to Castellano et al., 1997b). One example of the latter the signal from g-actin, equivalent amounts of ectopic approach would be the yeast two hybrid system, which p16INK4a expression were achieved (+20%). Similarly, has been used to some e€ect both in the initial immunoblotting with two di€erent monoclonal anti- identi®cation of p16INK4a (Serrano et al., 1993) and in bodies against p16INK4a con®rmed that the cell pools the functional evaluation of mutants (Reymond and expressed equivalent levels of HA-tagged p16INK4a Brent, 1995; Yang et al., 1995). However, we elected to (Figure 3b). Importantly, the level of HA-tagged base our assays on mammalian cells since they are p16INK4a was quite similar to that of the endogenous more likely to re¯ect the environment in which p16INK4a p16INK4a expressed in these `middle-aged' ®broblasts normally operates. (*40 population doublings). With the exception of An inherent problem with such assays is to control A20P, all the variant p16INK4a proteins were expressed for the levels of expression of the di€erent variants. To at roughly equivalent levels. Note that the levels of try to circumvent the variability associated with DNA A20P protein did not match those of the A20P RNA transfection, we used recombinant retroviruses to (compare Figure 3a and b) suggesting that this variant express the variant p16INK4a sequences in primary may be intrinsically unstable. human diploid ®broblasts. Use of a primary cell strain has the attraction that the cell cycle machinery Interaction between exogenous p16INK4a and endogenous is presumably intact and should not have sustained the CDKs alterations necessary to evolve an immortal cell line. Moreover, we recently showed that retroviral delivery The availability of these infected cell pools enabled us of p16INK4a into human diploid ®broblasts can induce to ask whether the variant proteins were capable of facets of the senescence phenotype, including G1 cell interacting with endogenous CDKs in TIG-ER cells. cycle arrest, accumulation of hypophosphorylated pRb Cell lysates were immunoprecipitated with antibodies and the expression of senescence associated b- against CDK4 and CDK6 and the amounts of co- galactosidase activity (McConnell et al., 1998). precipitating p16INK4a determined by Western blotting. The variant p16INK4a sequences were tagged with two Again, a clear distinction could be drawn between the copies of the HA epitope, to allow discrimination exogenous HA-tagged variants and the endogenous between exogenous and endogenous p16INK4a. Infectious p16INK4a, the latter serving as an internal control. Function and structure of p16 mutants MRuaset al 5427 Although there was some variability in the signals, all experiment whereas cells expressing wild-type p16INK4a of the p16INK4a variants that retained CDK binding in expanded much more slowly, at most increasing vitro were also able to interact with CDK4 and CDK6 around 3 ± 6-fold (Figure 5). In terms of the variant in vivo (Figure 4). As expected, the A20P variant was not detected in the CDK4 or CDK6 immunoprecipi- tates, presumably due to a combination of reduced binding anity and stability. In general, however, variants that failed to bind CDKs in vitro also failed to associate with CDK4 and CDK6 in transfected or infected cells (additional data not shown).

E€ects of p16INK4a variants on TIG-ER cell proliferation Since the retrovirally infected TIG-ER cells appeared to be expressing equivalent and physiologically relevant levels of p16INK4a, we next asked what e€ects the variants might have on cell behaviour. Wild-type p16INK4a has been shown to inhibit the phosphorylation of pRb in these cells and to cause a severe growth arrest in a simple proliferation assay (McConnell et al., 1998). Pools of TIG-ER cells expressing variant Figure 4 Interaction of p16INK4a variants with endogenous CDK4 p16INK4a proteins were monitored for a period of 8 and CDK6. TIG-ER cells were infected with p16INK4a-encoding days following the initial puromycin selection. Cells retroviruses and 6 days post-infection, lysates were prepared for immunoprecipitation with antibodies against CDK4 (a) or CDK6 infected with the empty vector grew rapidly, increasing (b). The precipitates were fractionated by SDS ± PAGE and 20 ± 25-fold in number (between four and ®ve immunoblotted with the JC8 antibody against human p16INK4a population doublings) during the course of the (equivalent results were obtained with the DCS50 antibody)

Figure 3 Ectopic expression of p16INK4a variants in TIG-ER cells. TIG-ER cells at population doubling 40 ± 42 were infected with recombinant ecotropic viruses encoding HA-tagged p16INK4a and selected in puromycin. (a) Six days after infection, total RNA was prepared and the levels of 26HA-p16INK4a transcripts were compared by Northern blotting. Fifteen mg of RNA were loaded in each lane and the relative amounts of exogenous p16INK4a RNA were calculated by reference to the signal from g-actin. (b) Cell lysates prepared at the same time points were analysed for p16INK4a protein expression by direct immunoblotting with either the JC8 or DCS50 monoclonal antibodies. Fifty mg of total protein were loaded in each lane. The positions of the exogenous (26HA-p16) and endogenous forms of p16INK4a are indicated on the right Function and structure of p16 mutants MRuaset al 5428 p16INK4a proteins, three distinct types of behaviour were phorylated (upper band) to hypophosphorylated forms observed. With variants of p16INK4a that were unable to of pRb (lower band). There was also an overall bind to CDKs in vitro, such as A20P, cell proliferation decrease in total amount of pRb in the cells (compare was essentially unimpaired. Interestingly, two of the ts lanes 1 and 2 in Figure 6). This phenomenon has been mutants, G101W and A118T, also fell into this observed in previous studies but the underlying category (not shown). Among the variants that mechanism remains obscure (Sandig et al., 1997; Fang retained CDK binding, H98Y and A127S caused a et al., 1998; Frizelle et al., 1998; McConnell et al., severe growth arrest akin to that of wild-type p16INK4a, 1999). The non-binding mutant A20P caused no whereas N71S, G122S and A132P gave intermediate discernible change in the pRb pattern (lane 3) whereas e€ects on cell proliferation. The results obtained with the V95A, H98Y and A127S variants were again the other variants in this study are summarized in indistinguishable from wild-type p16INK4a (lanes 5, 6 Table 1. and 8). In contrast, the e€ects observed with the N71S, Importantly, the same distinctions were manifest G122S and A132P variants were less pronounced, with when the phosphorylation status of pRb in these cells only a partial reduction in the phosphorylation of pRb was monitored by direct immunoblotting (Figure 6). (Figure 6, lanes 4, 7 and 9). As expected, expression of wild-type p16INK4a caused a dramatic reduction in the proportion of the hyperphos- Discussion

Relative merits of functional assays for p16INK4a In this study, we set out to devise a functional assay or set of assays for p16INK4a that would potentially discriminate between partially impaired variants and innocuous polymorphisms, as well as resolving some of the inconsistencies that occur in the published literature. For these purposes, we analysed a panel of 17 tumour-speci®c variants of p16INK4a whose back- ground and properties are summarized in Table 1. After evaluating a number of di€erent approaches, including measurements of kinase inhibition in baculovirus based systems, G1 arrest of transfected U20S cells, and protein-protein interactions and turnover in transfected COS-1 cells (Parry and Peters, 1996; and data not shown), we conclude that two assays, in vitro binding to CDK4 and CDK6 and the proliferation of retrovirally infected ®broblasts, provide the most reliable information. Indeed, the cell proliferation assay alone may suce, but since it is more time consuming than in vitro binding, the latter still remains our ®rst line assay of choice. In its simplest form, it avoids the problems associated with protein puri®cation and renaturation but its major weakness is that it can only be reliably interpreted in Figure 5 Inhibition of TIG-ER cell proliferation by p16INK4a variants. Six days after retroviral infection, TIG-ER cells were terms of all or nothing e€ects and does not give a transferred into 24-well plates and followed for 8 days. The quantitative readout of binding anity (see Figure 1). relative numbers of viable cells were determined by staining with Nevertheless, mutants that fail to bind CDKs in vitro crystal violet and measurement of absorbance at 590 nm (Serrano are generally impaired in all other functional tests. et al., 1997). Values were normalized against cell number at day 0. Each time point was determined in triplicate and similar results The in vitro assay can also be adapted to identify ts were obtained in at least two independent experiments variants but among the variants tested here, only N71S, G101W and A118T were considered to be ts (Figure 2). Thus far, ts behaviour appears restricted to germline variants associated with familial melanoma, although we have not conducted a systematic search. Since the assay requires the production and careful titration of recombinant protein, it has proved to be rather temperamental and we would not recommend its general application unless improved expression systems can be developed to produce soluble and correctly folded p16INK4a fusion proteins. We currently place Figure 6 Inhibition of pRb phosphorylation by p16INK4a more faith in the impaired function of these variants in variants. Samples (50 mg) of total protein from TIG-ER cells 6 the cell proliferation assay (Table 1). days after retroviral infection were fractionated by SDS ± PAGE Given the vagaries of the di€erent assay systems and in an 8% gel and directly immunoblotted with an antiserum the inconsistencies in the literature, how can we be sure against human pRb (G3-245, PharMingen). The positions of hyper (ppRb) and hypophosphorylated (pRb) forms of the that the cell proliferation assay is more reliable than protein are indicated on the right any of the others? Theoretically, an in vivo assay has Function and structure of p16 mutants MRuaset al 5429 the advantage that it can integrate the di€erences that shape and having only a limited hydrophobic core (see might occur in the ability of variant p16INK4a to interact Figure 7). The convex face of the INK4 protein binds with other proteins, in its subcellular distribution or in some distance from the active site of the CDK such its inherent stability, without necessarily identifying the that the b hairpins lie across the top of the CDK N- details of each de®ciency. Unlike permanent cell lines, terminal lobe with the base of the a helical region human diploid ®broblasts bene®t from having no making contacts with the C-terminal lobe (Brotherton genetic alterations that might compromise their et al., 1998; Russo et al., 1998). response to p16INK4a. However, these cells have the The most straightforward variants to interpret are disadvantage that the endogenous levels of p16INK4a those a€ecting core amino acids, de®ned here as increase with population doublings, so that their residues whose side chains are 57% accessible. For usefulness may diminish with time. Ectopic expression example, both L16 and A20 of helix 1 are conserved in of wild-type p16INK4a has been shown to induce all INK4 proteins and are buried within the molecule senescence-like phenotypic e€ects in some established where they contribute to the hydrophobic core and cell lines (Shapiro et al., 1995; Fueyo et al., 1996; ensure a close packing of adjacent helices. Replacing Spillare et al., 1996; Castellano et al., 1997a; Uhrbom either residue with proline would introduce a kink in et al., 1997; Vogt et al., 1998) but our preliminary the helix, by disrupting its mainchain hydrogen bonds ®ndings suggest that the levels of p16INK4a achieved by and thereby perturbing helical packing. A reciprocal retroviral delivery may not be sucient to arrest change occurs in the case of P48L. Residue P48 in immortal lines lacking endogenous p16INK4a. p16INK4a is replaced by alanine in the three other INK4 proteins and lies at the beginning of a loop region that in the canonical ankyrin repeat would be expected to Correlation of functional data with the structure of constitute an a helix. The absence or shortening of p16INK4a helix-3 is characteristic of the INK4 proteins and One way to judge the reliability of the functional data results in the splaying of the ®nger-like b-turns that is to consider the e€ects that the various sequence link the adjacent repeats. According to the structural alterations might have on the three dimensional data, P48 is buried and there is unlikely to be enough structure of the protein. The availability of the space to accommodate the bulkier leucine sidechain crystallographic coordinates for p16INK4a and related without distorting the adjacent loop region. The INK4 proteins bound to CDK6 (Brotherton et al., predicted e€ects of these three mutations are therefore 1998; Russo et al., 1998) has enabled us to classify the entirely consistent with complete loss of function. various p16INK4a mutants into those a€ecting buried A118 is also in the hydrophobic core of the core residues, surface exposed residues, or residues molecule, in helix 7, and the structure suggests that located in the INK4-CDK interface. As previously in the A118T variant, there should be sucient space discussed (McDonald and Peters, 1998), p16INK4a to accommodate the g oxygen of the threonine residue belongs to the ankyrin superfamily whose members which would form two additional hydrogen bonds to typically contain multiple (at least four) copies of the the carbonyl of V115 and amide of E119. This so-called ankyrin repeat (Luh et al., 1997; Byeon et al., prediction is consistent with the fact that A118T 1998; Venkataramani et al., 1998). Each repeat unit retains some capacity to bind to CDK4 and CDK6 comprises a helix ± loop ± helix motif ¯anked by b (Harland et al., 1997) but is at odds with the data in strands and the units are linked together by b hairpins Table 1 that show A118T to be functionally impaired formed by strands from adjacent repeats. The result is in the cell proliferation assay. Since A118T was an asymmetric structure resembling an extended `L' identi®ed in the germline of a melanoma kindred

Figure 7 The ®gure on the left depicts the electrostatic surface of a model of p16INK4a derived from the crystallographic coordinates of p18INK4c (McDonald and Peters, 1998; Venkataramani et al., 1998). Red indicates negative charge and blue positive charge. The CDK-binding interface faces towards the viewer and is centred approximately between M53 and D84 (Brotherton et al., 1998; Russo et al., 1998). On the right is a worm representation of the p16INK4a model in the same view. Selected p16INK4a mutations discussed in the text are highlighted and labelled. The ®gure was drawn using the programme GRASP (Nicholls et al., 1991) Function and structure of p16 mutants MRuaset al 5430 (Harland et al., 1997) and shows ts properties in the The other partially impaired variant, A132P, was CDK binding assay (Figure 2), there is little doubt that included in the study because it maps within the fourth the functional data are correct. ankyrin repeat and in previous studies, truncation of The other ts variants identi®ed here, N71S and p16INK4a at residue 131 yielded contradictory ®ndings G101W, can be classi®ed structurally as alterations to (Yang et al., 1995; Lilischkis et al., 1996). At the non-core residues that are solvent accessible. The most primary sequence level, there is little homology interesting is N71S which occurs in the germline of a between the INK4 proteins downstream of residue melanoma family but was judged to have only low 132 of p16INK4a, and the closely related p15INK4b protein penetrance (Hussussian et al., 1994). Previous func- terminates at the equivalent of residue 137. With only tional assays were equivocal and led to the conclusion 3.5 ankyrin repeats, p15INK4b may comprise the that this variant could represent a rare polymorphism minimum number required for the stability of the (Ranade et al., 1995; Reymond and Brent, 1995). The structure. In helix 8, the amide of A132 forms a p16INK4a structure shows that N71 does not participate hydrogen bond with the carbonyl of R128 and in the CDK interaction but forms two intramolecular insertion of a proline would disrupt this helix. It is hydrogen bonds, one of which is to the main chain possible that the loss of this hydrogen bond is enough amide of R103 and e€ectively crosslinks two adjacent to adversely a€ect the stability of the protein although ankyrin motifs. Substitution with a serine residue currently available structural information sheds little would remove this hydrogen bond and could poten- light on the likely impact of mutations that map close tially destabilize the protein. Signi®cantly, N71S to the carboxy terminus of p16INK4a (see compilation in behaved as if partially impaired in the cell prolifera- Ruas and Peters, 1998). tion assay and we would therefore conclude that it is The three remaining variants in the study are also in indeed a loss of function mutation. G101, on the other this indeterminate category in terms of structural hand, is one of several glycine residues that are critical considerations and all three behaved as wild-type in for the formation of the ankyrin repeat structure since the functional assays. A127S was already considered to their ability to adopt a positive phi torsion angle is be a polymorphism because of its presence in necessary for the sharp turn between adjacent helices. una€ected tissues (Okamoto et al., 1994; Gonzalez- Substitution with any other residue would be Zulueta et al., 1995; Suzuki et al., 1995) and from energetically unfavourable. The same argument ap- functional data (Koh et al., 1995). Since A127 is on the plies to the G122S variant which despite its ability to surface of the protein, away from the CDK interacting bind to CDK4 and CDK6 in vitro is clearly domains, addition of an oxygen atom, as in A127S, functionally compromised (Figure 5). would be readily tolerated. Three other variations have The third structural category comprises residues been recorded at this residue, A127V (Kyritsis et al., involved in the INK4-CDK interface, including D84, 1996), A127E (Konishi et al., 1996) and A127P E88 and M53 (see Figure 7). Five di€erent mutations (Betticher et al., 1997) and the same arguments have been recorded at D84, changing the encoded apply. In the case of V95, which is not conserved in amino acid to H, N, Y, E or V (Hayashi et al., 1994; INK4c and INK4d, the structure indicates that it is Mori et al., 1994; Zhang et al., 1994; Gonzalez-Zulueta partly accessible on the surface of the molecule and et al., 1995; Komiya et al., 1995; Liu et al., 1995; that substitution with alanine would be unlikely to Tamimi et al., 1996; Marchetti et al., 1997; Olshan et have an a€ect. Curiously, V95 occurs in a 20 amino al., 1997) and previous functional assays on D84 acid synthetic peptide, corresponding to residues 84 ± variants have given inconsistent results (Koh et al., 103 of p16INK4a, that is capable of binding to and 1995; Enders et al., 1996; Tevelev et al., 1996; inhibiting the function of CDK4 (FaÊ hraeus et al., Yarbrough et al., 1996). However, all three of the 1996). Substituting V95 with alanine signi®cantly D84 variants tested here proved to be functionally reduced the e€ectiveness of the peptide (FaÊ hraeus et impaired (Table 1). D84 forms a salt bridge with R87 al., 1996) but this does not appear to be the case in the and directly contacts CDK6, suggesting that it is full length protein. Thus, although this substitution critical to maintain the negative charge on the was recorded in a survey of non-small cell lung cancers aspartate residue to neutralise the positively charged (Hayashi et al., 1994), our data suggest that V95A is arginine (Figure 7). The D84N, D84H and D84Y either a polymorphism or a sequencing error. mutations would e€ectively remove the negative charge Similarly, our functional data imply that H98Y is and dramatically alter the electrostatic character of the not a mutant and the same conclusion was drawn CDK-binding surface of p16INK4a. Conversely, R87P, recently by Arap et al. in a survey of glioma-associated which is a well documented loss of function mutation sequence variants (Arap et al., 1997). This contrasts associated with familial melanoma (Hussussian et al., with H98P which we and others have shown to be 1994; Koh et al., 1995; Ranade et al., 1995; Reymond functionally defective (Koh et al., 1995; Parry and and Brent, 1995; Lilischkis et al., 1996; Parry and Peters, 1996; Zhang and Peng, 1996). H98 is not Peters, 1996; Zhang and Peng, 1996), would also a€ect strictly conserved among the INK4 proteins, being the charge distribution of this region of p16INK4a.In replaced by V and L in p18INK4c and p19INK4d contrast, M53 makes multiple Van der Waals contacts respectively. These residues are generally buried in which individually are likely to contribute less binding helix 6 and the structure suggests that insertion of a energy than interactions between charged residues. The proline is likely to disrupt the helix. However, insertion M53I alteration presumably interferes with these of an aromatic side chain, as in H98Y, should have contacts. Interestingly, this variant behaved as only little e€ect, although the tyrosine may no longer be partially impaired in the cell proliferation assay (Table completely buried. Point mutations are extremely rare 1) although we previously scored it as non-functional in primary gliomas (Ruas and Peters, 1998) and H98Y in the in vitro binding assay (Harland et al., 1997). is one of several that were identi®ed by sequencing Function and structure of p16 mutants MRuaset al 5431 individual clones of PCR-products from di€erent focal manufacturer's protocols (Stratagene). All the mutations areas (Kyritsis et al., 1996). Since none of these were veri®ed by DNA sequencing. The respective p16INK4a variants appear functionally compromised (Arap et coding sequences, ¯anked by BamHI and EcoRI restriction al., 1997), they were most likely generated during the sites, were then transferred into the pRSET vector (Invitrogen) to facilitate in vitro translation of His -tagged PCR. 6 proteins and the production of recombinant proteins in E. In conclusion, therefore, there is good agreement coli. The coding sequences were also transferred into the between the behaviour of the p16INK4a variants in the pcDNA3 vector to permit high level expression in COS-1 cell proliferation assay and the likely e€ects that these cells. Finally, the sequences were inserted into the changes would have on the structure of the protein. pBABEpuro retrovirus vector (Morgenstern and Land, Alterations a€ecting core residues are generally 1990) to allow infection of primary ®broblasts. Two copies associated with complete loss of function, the one of the HA tag were incorporated at the amino terminus of exception being a suspected sequencing artefact, the wild-type p16INK4a construct in pBABEpuro by ligating whereas alterations to solvent exposed residues or to complementary synthetic oligonucleotides: 26HA-FOR the CDK interface can have more subtle e€ects. 5'-gatctatgctagcatacccgtacgatgttcctgactatgccggctatccctatgacgt- Despite these correlations, however, structural model- cccggactatg-3' 26HA-REV 5'-gatccatagtccgggacgtcataggga- tagccggcatagtcaggaacatcgtacgggtatgctagcat-3'. The variant ling is likely to remain an adjunct to rather than a p16INK4a sequences were subsequently transferred into the substitute for the biological assays since the precise retroviral vector as BamHI-EcoRI fragments. e€ect of each alteration cannot be determined with any certainty. For example, although the structure has identi®ed residues involved in the CDK interface, their Preparation of recombinant p16INK4a relative contributions in terms of binding energy The variant forms of p16INK4a in the pRSETA vector were remain unknown. Predictions from the structure will transfected into the BL21(DE3)pLysS strain of E. coli. and invariably have to be con®rmed by biological assays individual colonies tested for production of recombinant and in this regard, we have developed a relatively protein after addition of IPTG. Typically, 400 ml cultures convenient and reliable test that can discriminate were used to prepare stocks of p16INK4a by exploiting the between severely impaired, partially impaired or amino terminal His6 tag. The recombinant protein was functionally normal variants of p16INK4a. recovered from bacterial inclusion bodies by solubilization in 20 mM Tris HCl (pH 8.0), 0.5 M NaCl, 5 mM imidazole, and 6 M urea and puri®ed directly on nickel-charged chelating Sepharose (Pharmacia). Bound proteins were washed and eluted using the same bu€er containing 50 mM Materials and methods and 1 M imidazole, respectively. Puri®ed proteins were dialysed stepwise against TBS (25 mM Tris HCl, pH 7.6, Cell culture, DNA transfection and retroviral infection 140 mM NaCl, 3 mM KCl) containing 3, 2, or 1 M urea, and All cell stocks used in this study were maintained at 378C and concentrated by a ®nal dialysis step against TBS containing 50% glycerol. Protein concentrations were determined using 5% CO2 in Dulbecco modi®ed Eagle's medium (DMEM) supplemented with 10% foetal calf serum (FCS). For the BCA protein assay reagent (Pierce), and the purity of transient transfections of COS-1 cells, 10 mg of the each preparation was assessed by SDS ± PAGE in a 12% appropriate plasmid DNA was added to 16106 cells using polyacrylamide gel and staining with PAGE Blue-83 (Harlow lipofectamine (Gibco ± BRL). TIG-3 human diploid fibro- and Lane, 1988). Typical protein preparations were approxi- blasts expressing the cell surface receptor for mouse ecotropic mately 95% pure. retroviruses (TIG-ER cells) were generated as previously described (McConnell et al., 1998) and maintained in medium containing 300 mg/ml G418. To prepare retroviral stocks, In vitro binding assays 56106 BOSC-23 cells (Pear et al., 1993) were plated in a Protein-protein interaction assays were performed as pre- 10 cm dish and transfected the next day with 10 mgof viously described (Parry and Peters, 1996) using components pBABEpuro retroviral vector DNA (Morgenstern and Land, synthesized in vitro by coupled transcription and translation 1990) using standard calcium phosphate precipitation in the of plasmid DNAs. The CDKs were labelled by incorporation 35 presence of 25 mM chloroquine. Transfection was allowed to of S-methionine. In some experiments, recombinant His6- proceed for 7 ± 8 h and the cells were grown for a further 18 ± tagged p16INK4a was used instead of the in vitro translated 24 h before harvesting the virus. The volume of medium was protein. The protein complexes were immunoprecipitated reduced to 5 ml and the virus suspensions were recovered with a polyclonal antiserum against the carboxy terminal 24 h later. Cell debris was removed by passage through a region of p16INK4a, fractionated by SDS ± PAGE and the 0.45 mm nitrocellulose ®lter. labelled components were visualized by autoradiography. For retroviral infections, early passage TIG-ER cells (40 ± 42 population doublings) were plated at 26106 cells per Protein analysis 75 cm2 ¯asks and incubated overnight. The culture medium was replaced with 5 ml of viral supernatant from the BOSC- TIG-ER cells expressing HA-tagged p16INK4a were analysed 6 23 cells, together with 3 ml of fresh medium and the days after retroviral infection. For immunoprecipitations, equivalent of 8 mg/ml polybrene. After 24 h, the medium 36106 cells were plated into 15 cm dishes and lysates were was replaced and selection in medium containing 2.5 mg/ml prepared the following day. The cells were washed twice with puromycin was initiated on day 2 post-infection and phosphate bu€ered saline (PBS) and then placed in 2.5 ml of continued for a further 4 days. NP40-lysis bu€er (50 mM Tris HCl, pH 8.0, 150 mM NaCl, 1% Nonidet P40, 2 mg/ml aprotinin, 0.1 mM NaF, 0.1 mM sodium orthovanadate, 100 mg/ml PMSF). Samples contain- Construction of vectors encoding mutant forms of p16INK4a ing 0.3 mg of total protein were immunoprecipitated over- Selected missense mutations were introduced into the wild- night at 48C with 5 ml of the following antibodies: a-CDK4 type p16INK4a cDNA sequence using synthetic oligonucleo- (Santa Cruz H-22), a-CDK6 (LBO-1, Bates et al., 1994) and tides containing the required base substitution and the a-p16 (DPAR-12, Parry et al., 1995) plus 25 ml of protein A QuikChange mutagenesis procedure according to the beads (Pierce). The precipitates were washed four times in Function and structure of p16 mutants MRuaset al 5432 lysis bu€er and once in 50 mM Tris HCl (pH 8.0) and membrane (Amersham). The blots were prehybridized for 1 h resolved by SDS ± PAGE in a 12% polyacrylamide gel at 658Cin66SSPE (16SSPE is 0.18 M NaCl, 10 mM (Harlow and Lane, 1988). sodium phosphate, pH 7.7, 1 mM EDTA) containing 0.5% For immunoblotting, the proteins were transferred to SDS and 56Denhardt's reagent (16Denhardt's reagent is Immobilon-P (Millipore) and the membranes were placed in 0.02% ®coll, 0.02% polyvinylpyrrolidone and 0.02% BSA). blocking solution (5% milk powder, 0.2% Tween-20 in PBS) and hybridized overnight with labelled probe. The 26HA- for 30 min. Two monoclonal antibodies were used to detect REV oligonucleotide terminally labelled with g-32P-ATP was p16: DCS-50, recognizing the carboxy terminal domain used to detect HA-tagged exogenous p16INK4a transcripts. The (Lukas et al., 1995) and JC8, recognizing an epitope near blot was washed three times in 16SSPE/0.1% SDS at 658C the amino terminus of the protein (generously provided by J for 20 min and once in 0.16SSPE/0.1% SDS. Signals were Koh and E Harlow). Membranes were incubated for 1 h at visualized by autoradiography and quanti®ed by phosphor- room temperature with the respective tissue culture super- imaging. natants diluted 1 : 5 in blocking bu€er, then thoroughly washed in 0.2% Tween-20 in PBS. Sheep anti-mouse HRP (1/ Cell proliferation assays 2000) was used as secondary antibody (Amersham). Antibody binding was visualized using Amersham ECL reagents. TIG-ER cells were infected with recombinant retroviruses For the direct immunoblotting of pRb, 106 cells were encoding p16INK4a and 5 days post infection, 56103 plated in 10 cm dishes and incubated overnight. Cell lysates puromycin resistant cells were plated in 24-well plates. At were prepared using 200 ml of incomplete sample bu€er (3% various times thereafter, individual wells were washed twice SDS, 96 mM Tris HCl pH 6.8) and after protein quantifica- with PBS, ®xed in 10% formaldehyde and rinsed twice with tion using the BCA protein assay reagent (Pierce), the water. Cells were stained with 0.1% crystal violet for 30 min samples were diluted 2 : 1 with 28% glycerol, 7% mercap- and washed four times with water. Intracellular stain was toethanol, 0.6% bromophenol blue to achieve the equivalent extracted with 1 ml of 10% acetic acid, diluted with an equal of 16 Laemmli sample bu€er (Harlow and Lane, 1988). volume of water, and the optical density at 590 nm was Samples corresponding to 50 mg of total protein were determined. Values were normalized to that obtained at day 0 resolved by SDS ± PAGE in an 8% polyacrylamide gel for cells expressing each p16INK4a variant. The day 0 value was (Harlow and Lane, 1988) and transferred to Immobilon-P determined immediately after cells had adhered to the tissue membrane (Millipore). The blot was probed as above using a culture plate (generally 4 ± 5 h after plating). Each time point 1 : 1000 dilution of a mouse monoclonal antibody against was assayed in triplicate. pRb (G3-245, PharMingen) as the primary antibody.

RNA analysis Acknowledgements Five days after retroviral infection, 106 cells TIG-ER cells We are grateful to Ronen Marmorstein and Ernest Laue were plated in 10 cm dishes and incubated overnight. RNA for supplying crystallographic coordinates and thank was prepared using the RNeasy kit as recommended by the Anna-Marie Frischauf and Malcolm Parker for comments manufacturer (Qiagen). Samples (15 mg) of total RNA were on the manuscript. Margarida Ruas is supported by a fractionated by electrophoresis through a 1.2% agarose gel Praxis XXI fellowship from FundacËaÄ oparaaCieÃnciae containing 2.2 M formaldehyde and transferred to Hybond N Tecnologia.

References

Alcorta DA, Xiong Y, Phelps D, Hannon G, Beach D and Della Ragione F, Russo GL, Oliva A, Mercurio C, Barrett JC. (1996). Proc.Natl.Acad.Sci.USA,93, Mastropietro S, Della Pietra V and Zappia V. (1996). J. 13742 ± 13747. Biol. Chem., 271, 15942 ± 15949. Arap W, Knudsen ES, Wang JYJ, Cavanee WK and Huang Dracopoli NC and Fountain JW. (1996). Cancer Surv., 26, H-JS. (1997). Oncogene, 14, 603 ± 609. 115 ± 132. Bates S, Bonetta L, MacAllan D, Parry D, Holder A, Enders GH, Koh J, Missero C, Rustgi AK and Harlow E. Dickson C and Peters G. (1994). Oncogene, 9, 71 ± 79. (1996). Oncogene, 12, 1239 ± 1245. Betticher DC, White GRM, Vonlanthen S, Liu X, Kappeler FaÊ hraeus R, Paramio JM, Ball KL, Laõ nSandLaneDP. A, Altermatt HJ, Thatcher N and Heighway J. (1997). Int. (1996). Curr. Biol., 6, 84 ± 91. J. Cancer, 74, 556 ± 562. Fang X, Jin X, Xu H-J, Liu L, Peng H-Q, Hogg D, Roth JA, Boice JA and Fairman R. (1996). Protein Sci., 5, 1776 ± 1784. Yu Y, Xu F, Bast Jr RC and Mills GB. (1998). Oncogene, Borg AÊ , Johansson U, Johannsson O, HaÊ kansson S, 16, 1±8. Westerdahl J, MaÊ sbaÈ ck A, Olsson H and Ingvar C. FitzGerald MG, Harkin DP, Silva-Arrieta S, MacDonald (1996). Cancer Res., 56, 2497 ± 2500. DJ, Lucchina LC, Unsal H, O'Neill E, Koh J, Finkelstein Brotherton DH, Dhanaraj V, Wick SLB, Domaille PJ, DM, Isselbacher KJ, Sober AJ and Haber DA. (1996). Volyanik E, Xu X, Parisini E, Smith BO, Archer SJ, Proc. Natl. Acad. Sci. USA, 93, 8541 ± 8545. Serrano M, Brenner SL, Blundell TL and Laue ED. Flores JF, Pollock PM, Walker GJ, Glendening JM, Lin AH- (1998). Nature, 395, 244 ± 250. T, Palmer JM, Walters MK, Hayward NK and Fountain Byeon I-JL, Li J, Ericson K, Selby TL, Tevelev A, Kim H-J, JW. (1997). Oncogene, 15, 2999 ± 3005. O'Maille P and Tsai M-D. (1998). Mol. Cell, 1, 421 ± 431. FrizelleSP,GrimJ,ZhouJ,GuptaP,CurielDT,GeradtsJ Castellano M, Gabrielli BG, Hussussian CJ, Dracopoli NC and Kratzke RA. (1998). Oncogene, 16, 3087 ± 3095. and Hayward NK. (1997a). J. Invest. Dermatol., 109, 61 ± FueyoJ,Gomez-ManzanoC,YungWKA,ClaymanG,Liu 68. T-J, Bruner J, Levin VA and Kyritsis AP. (1996). Castellano M, Pollock PM, Walters MK, Sparrow LE, Oncogene, 12, 103 ± 110. Down LM, Gabrielli BG, Parsons PG and Hayward NK. (1997b). Cancer Res., 57, 4868 ± 4875. Function and structure of p16 mutants MRuaset al 5433 Gonzalez-Zulueta M, Shibata A, Ohneseit PF, Spruck III Musgrove EA, Swarbrick A, Lee CSL, Cornish AL and CH, Busch C, Shamaa M, El-Baz M, Nichols PW, Sutherland RL. (1998). Mol. Cell. Biol., 18, 1812 ± 1825. Gonzalgo ML, MalmstroÈ m P-U and Jones PA. (1995). J. Nicholls A, Sharp K and Honig B. (1991). Proteins, 11, 281 ± Natl. Cancer Inst., 87, 1383 ± 1393. 296. Guan K-L, Jenkins CW, Li Y, O'Keefe CL, Noh S, Wu X, OhtaM,NagaiH,ShimizuM,RasioD,BerdD, Zariwala M, Matera AG and Xiong Y. (1996). Mol. Biol. Mastrangelo M, Singh AD, Shields JA, Shields CL, Cell., 7, 57 ± 70. Croce CM and Huebner K. (1994). Cancer Res., 54, Hall M, Bates S and Peters G. (1995). Oncogene, 11, 1581 ± 5269 ± 5272. 1588. Okamoto A, Demetrick DJ, Spillare EA, Hagiwara K, Hara E, Smith R, Parry D, Tahara H, Stone S and Peters G. Hussain SP, Bennett WP, Forrester K, Gerwin B, Serrano (1996). Mol. Cell. Biol., 16, 859 ± 867. M, Beach DH and Harris CC. (1994). Proc. Natl. Acad. HarlandM,MeloniR,GruisN,PinneyE,BrookesS,Spurr Sci. USA, 91, 11045 ± 11049. NK, Frischauf A-M, Bataille V, Peters G, Cuzick J, Selby Olshan AF, Weissler MC, Pei H, Conway K, Anderson S, P, Bishop DT and Newton Bishop J. (1997). Hum. Mol. Fried DB and Yarbrough WG. (1997). Oncogene, 14, Genet., 6, 2061 ± 2067. 811 ± 818. Harlow E and Lane D. (1988). Antibodies. A laboratory Parry D, Bates S, Mann DJ and Peters G. (1995). EMBO J., manual. Cold Spring Harbor Laboratory: Cold Spring 14, 503 ± 511. Harbor, New York. Parry D and Peters G. (1996). Mol. Cell. Biol., 16, 3844 ± Hayashi N, Sugimoto Y, Tsuchiya E, Ogawa M and 3852. Nakamura Y. (1994). Biochem. Biophys. Res. Commun., Pear WS, Nolan GP, Scott ML and Baltimore D. (1993). 202, 1426 ± 1430. Proc. Natl. Acad. Sci. USA, 90, 8392 ± 8396. Hayward NK. (1996). Curr. Opin. Oncol., 8, 136 ± 142. Platz A, Hansson J, MaÊ nsson-Brahme E, LagerloÈ f B, Linder Hussussian CJ, Struewing JP, Goldstein AM, Higgins PAT, S, Lundqvist E, Sevigny P, InganaÈ s M and Ringborg U. Ally DS, Sheahan MD, Clark Jr WH, Tucker MA and (1997). J. Natl. Cancer Inst., 89, 697 ± 702. Dracopoli NC. (1994). Nature Genet., 8, 15 ± 21. Pollock PM, Yu F, Qiu L, Parson PG and Hayward NK. Kamb A, Shattuck-Eidens D, Eeles R, Liu Q, Gruis NA, (1995). Oncogene, 11, 663 ± 668. Ding W, Hussey C, Tran T, Miki Y, Weaver-Feldhaus J, Ranade K, Hussussian CJ, Sikorski RS, Varmus HE, McClureM,AitkenJF,AndersonDE,BergmanW, Goldstein AM, Tucker MA, Serrano M, Hannon GJ, Frants R, Goldgar DE, Green A, MacLennan R, Martin Beach D and Dracopoli NC. (1995). Nature Genet., 10, NG, Meyer LJ, Youl P, Zone JJ, Skolnick MH and 114 ± 116. Cannon-Albright LA. (1994). Nature Genet., 8, 22 ± 26. Reymond A and Brent R. (1995). Oncogene, 11, 1173 ± 1178. Kinzler KW and Vogelstein B. (1997). Nature, 386, 761 ± 763. Rezniko€ CA, Yeager TR, Belair CA, Savelieva E, Koh J, Enders GH, Dynlacht BD and Harlow E. (1995). Puthenveettil JA and Stadler WM. (1996). Cancer Res., Nature, 375, 506 ± 510. 56, 2886 ± 2890. Komiya A, Suzuki H, Aida S, Yatani R and Shimazaki J. Ruas M and Peters G. (1998). BBA Rev. Cancer, 1378, 115 ± (1995). Jpn. J. Cancer Res., 86, 622 ± 625. 177. Konishi N, Hiasa Y, Tao M, Matsuda H, Nakamura M, Russo AA, Tong L, Lee J-O, Je€rey PD and Pavletich NP. Yane K and Kitahori Y. (1996). Int. J. Oncol., 8, 549 ± 554. (1998). Nature, 395, 237 ± 243. Kyritsis AP, Zhang B, Zhang W, Xiao M, Takeshimi H, Sandig V, Brand K, Herwig S, Lukas J, Bartek J and Strauss Bondy ML, Cunningham JE, Levin VA and Bruner J. M. (1997). Nature Med., 3, 313 ± 319. (1996). Oncogene, 12, 63 ± 67. Serrano M, Hannon GJ and Beach D. (1993). Nature, 366, Lilischkis R, Sarcevic B, Kennedy C, Warlters A and 704 ± 707. Sutherland RL. (1996). Int. J. Cancer, 66, 249 ± 254. Serrano M, Lin AW, McCurrach ME, Beach D and Lowe Liu Q, Neuhausen S, McClure M, Frye C, Weaver-Feldhaus SW. (1997). Cell, 88, 593 ± 602. J, Gruis NA, Eddington K, Allalunis-Turner MJ, Shapiro GI, Park JE, Edwards CD, Mao L, Merlo Am, Skolnick MH, Fujimura FK and Kamb A. (1995). Sidransky D, Ewen ME and Rollins BJ. (1995). Cancer Oncogene, 10, 1061 ± 1067. Res., 55, 6200 ± 6209. Loughran O, Malliri A, Owens D, Gallimore PH, Stanley Sherr CJ. (1994). Cell, 79, 551 ± 555. MA, Ozanne B, Frame MC and Parkinson EK. (1996). Sherr CJ and Roberts JM. (1995). Genes Dev., 9, 1149 ± 1163. Oncogene, 13, 561 ± 568. Sou®r N, Avril M-F, Chompret A, Demenais F, Bombled J, Luh FY, Archer SJ, Domaille PJ, Smith BO, Owen D, Spatz A, Stoppa-Lyonnet D, Group TFFMS, Be nard J BrothertonDH,RaineARC,XuX,BrizuelaL,Brenner and Bressac-de Paillerets B. (1998). Hum. Mol. Genet., 7, SL and Laue ED. (1997). Nature, 389, 999 ± 1003. 209 ± 216. Lukas J, Parry D, Aagaard L, Mann DJ, Bartkova J, Strauss Spillare EA, Okamoto A, Hagiwara K, Demetrick DJ, M, Peters G and Bartek J. (1995). Nature, 375, 503 ± 506. Serrano M, Beach D and Harris CC. (1996). Mol. Mahony D, Parry DA and Lees E. (1998). Oncogene, 16, Carcinogen., 16, 53 ± 60. 603 ± 611. Sun S, Pollock PM, Liu L, Karimi S, Jothy S, Milner BJ, Marchetti A, Buttitta F, Pellegrini S, Bertacca G, Chella A, Renwick A, Lassam NJ, Hayward NK, Hogg D, Narod Carnicelli V, Tognoni V, Filardo A, Angeletti CA and SA and Foulkes WD. (1997). Int. J. Cancer, 73, 531 ± 536. Bevilacqua G. (1997). J. Pathol., 181, 178 ± 182. Suzuki H, Zhou X, Yin J, Lei J, Jiang HY, Suzuki Y, Chan T, McConnell BB, Gregory FJ, Stott FJ, Hara E and Peters G. Hannon GJ, Mergner WJ, Abraham JM and Meltzer SJ. (1999). Mol. Cell. Biol., 19, 1981 ± 1989. (1995). Hum. Mol. Genet., 4, 1883 ± 1887. McConnell BB, Starborg M, Brookes S and Peters G. (1998). Tamimi Y, Bringuier PP, Smit F, van Bokhoven A, Curr. Biol., 8, 351 ± 354. Debruyne FMJ and Schalken JA. (1996). Br.J.Cancer, McDonald NQ and Peters G. (1998). Nat. Struct. Biol., 5, 74, 120 ± 122. 85 ± 88. Tevelev A, Byeon I-JL, Selby T, Ericson K, Kim H-J, Morgenstern JP and Land H. (1990). Nucl. Acids Res., 18, Kraynov V and Tsai M-D. (1996). Biochemistry, 35, 3587 ± 3596. 9475 ± 9487. Mori,T,MiuraK,AokiT,NishihiraT,MoriSand Uhrbom L, Niste r M and Westermark B. (1997). Oncogene, Nakamura Y. (1994). Cancer Res., 54, 3396 ± 3397. 15, 505 ± 514. Function and structure of p16 mutants MRuaset al 5434 Venkataramani R, Swaminathan K and Marmorstein R. Yarbrough WG, Aprelikova O, Pei H, Olshan AF and Liu (1998). Nat. Struct. Biol., 5, 74 ± 81. ET. (1996). J. Natl. Cancer Inst., 88, 1489 ± 1491. Vogt M, Haggblom C, Yeargin J, Christiansen-Weber T and Yeager TR, De Vries S, Jarrard DF, Kao C, Nakada SY, Haas M. (1998). Cell Growth Di€., 9, 139 ± 146. Moon TD, Bruskewitz R, Stadler WM, Meisner LF, Walker GJ, Hussussian CJ, Flores JF, Glendening JM, Gilchrist KW, Newton MA, Waldman FM and Rezniko€ Haluska FG, Dracopoli NC, Hayward NK and Fountain CA. (1998). Genes Dev., 12, 163 ± 174. JW. (1995). Hum. Mol. Genet., 4, 1845 ± 1852. YoshidaS,TodorokiT,IchikawaY,HanaiS,SuzukiH, Weinberg RA. (1995). Cell, 81, 323 ± 330. Hori M, Fukao K, Miwa M and Uchida K. (1995). Cancer Whelan AJ, Bartsch D and Goodfellow PJ. (1995). New Eng. Res., 55, 2756 ± 2760. J. Med., 333, 975 ± 977. Zhang B and Peng Z. (1996). J. Biol. Chem., 271, 28734 ± Wick ST, Dubay MM, Imanil I and Brizuela L. (1995). 28737. Oncogene, 11, 2013 ± 2019. Zhang S-Y, Klein-Szanto AJP, Sauter ER, Shafarenko M, Yang R, Gombart AF, Serrano M and Koe‚er HP. (1995). Mitsunaga S, Nobori T, Carson DA, Ridge JA and Cancer Res., 55, 2503 ± 2506. Goodrow TL. (1994). Cancer Res., 54, 5050 ± 5053.