Oncogene
SEARCH     advanced search my account e-alerts subscribe register
Journal home
Advance online publication
Current issue
Archive
Press releases
For authors
For referees
Contact editorial office
About the journal
For librarians
Subscribe
Advertising
naturereprints
Contact NPG
Customer services
Site features
NPG Subject areas
Access material from all our publications in your subject area:
Biotechnology Biotechnology
Cancer Cancer
Chemistry Chemistry
Dentistry Dentistry
Development Development
Drug Discovery Drug Discovery
Earth Sciences Earth Sciences
Evolution & Ecology Evolution & Ecology
Genetics Genetics
Immunology Immunology
Materials Materials Science
Medical Research Medical Research
Microbiology Microbiology
Molecular Cell Biology Molecular Cell Biology
Neuroscience Neuroscience
Pharmacology Pharmacology
Physics Physics
Browse all publications
 
27 September 2001, Volume 20, Number 43, Pages 6152-6163
Table of contents    Previous  Article  Next   [PDF]
Original Paper
Role of LXCXE motif-dependent interactions in the activity of the retinoblastoma protein
Ho Man Chan, Linda Smith and Nicholas B La Thangue

Division of Biochemistry and Molecular Biology, Davidson Building, University of Glasgow, Glasgow, G12 8QQ, UK

Correspondence to: N B La Thangue, Division of Biochemistry and Molecular Biology, Davidson Building, University of Glasgow, Glasgow, G12 8QQ, UK. E-mail: n.lathangue@bio.gla.ac.uk

Abstract

Cell cycle control by pRb requires the integrity of the pocket domain, which is a region necessary for interactions with a variety of proteins, including E2F and LXCXE-motif containing proteins. Through knowledge of the crystal structure of pRb we have prepared a panel of pRb mutant derivatives in which a cluster of lysine residues that demark the LXCXE peptide binding domain were systematically mutated. One of the mutant derivatives, Rb6A, exhibits significantly reduced LXCXE-dependent interactions with HPV E7, cyclinD1 and HDAC2, but retained LXCXE-independent binding to E2F. Consistent with these results, Rb6A could down-regulate E2F-1-dependent activation of different E2F responsive promoters, but was compromised in Rb-dependent repression. Most importantly, Rb6A retained wild-type growth arrest activity, and colony forming activity similar to wild-type pRb. It is compatible with these results that directly targeting HDAC2 to E2F responsive promoters as an E2F/HDAC hybrid protein failed to effect cell cycle arrest. These results suggest that LXCXE-dependent interactions are not essential for pRb to exert growth arrest. Oncogene (2001) 20, 6152-6163.

Keywords

pRb; E2F; HDAC; LXCXE; tumour suppressor

Introduction

The retinoblastoma tumour suppressor protein pRb plays an important role in regulating early cell cycle progression through its interaction with and control of key transcription factors that promote cell cycle progression, such as E2F (Dyson, 1998; Harbour and Dean, 2000). In turn, the activity of pRb is governed at the post-translational level by cyclin-dependent kinases, principally cyclinD/cdk4 and cyclinE/cdk2, which sequentially phosphorylate pRb as cells progress towards S phase (Mittnacht, 1998).

Whilst pRb regulates the activity of E2F, multiple mechanisms appear to be involved in the control of E2F responsive genes. In this respect, it is recognized that some E2F target genes, such as cyclinE, contain E2F sites that function in transcriptional activation, and become inactive upon the interaction between E2F and pRb (Ohtani et al., 1995; Botz et al., 1996). In contrast, other genes, such as B-myb and E2F-1, contain E2F sites that can act to dominantly repress constitutive transcription (Tommasi and Pfeifer, 1995; Liu et al., 1996; Zwicker et al., 1996). The physical interaction between pRb and E2F requires residues located in the E2F transcriptional activation domain, and it is likely that the ability of pRb to physically hinder the interaction of the E2F activation domain with co-activators and basal transcription factors, such as p300/CBP and TFIIA respectively (Trouche and Kouzarides, 1996; Shikama et al., 1997; Dyson, 1998; Ross et al., 1999), is involved in pRb-mediated E2F transcriptional inactivation.

On the other hand, transcriptional repression by pRb is believed to involve histone deacetylases (HDACs) together with components of the SWI/SNF chromatin-remodelling complex, which form a physical complex with pRb (Brehm et al., 1998; Luo et al., 1998; Magnaghi-Jaulin, et al., 1998; Zhang et al., 2000). HDACs are a family of enzymes that act to remove acetyl groups from nucleosomal histone N-terminal tails and thereby favour nucleosome condensation (Hassig and Schrieber, 1997). The enzyme activity of HDACs opposes the action of histone acetyltransferases (HATs), such as p300/CBP co-activators (Shikama et al., 1997; Goodman and Smolik, 2000), which acetylate lysines in histone tail regions, and thereby are believed to loosen chromatin to favour transcriptional activation (Grunstein, 1997; Kingston and Narlikar, 1999). It has been proposed that during cell cycle progression cyclinD/cdk4 phosphorylation of pRb frees HDAC, and subsequent phosphorylation by cyclinE/cdk2 prompts the release of pRb from E2F (Harbour et al., 1999). Furthermore, different E2F genes may be regulated by pRb/HDAC/SWI/SNF and pRb/SWI/SNF complexes (Zhang et al., 1999).

The pRb pocket was originally identified as a region required to bind viral oncoproteins, such as adenovirus E1A and SV40 large T antigen (Moran, 1993). The pocket is divided into two domains, A and B, separated by a spacer region (Mulligan and Jacks, 1998). The integrity of the pocket is required for binding to E2F, and mutant Rb alleles that arise in tumour cells disrupt the pRb pocket and prevent the interaction between pRb and E2F (Moran, 1993; Mulligan and Jacks, 1998). Viral oncoproteins, including adenovirus E1A, SV40 large T antigen and HPV E7, share a conserved protein motif, LXCXE, which is required for the physical interaction with pRb (Moran, 1993).

The three dimensional structure of the pRb pocket bound to the LXCXE peptide taken from E7 indicated that the LXCXE peptide binds to a hydrophobic groove in the B domain (Lee et al., 1998). This peptide binding site is highly conserved among pRb homologues from different species, and between the pRb-related proteins p107 and p130, and several conserved amino acid residues in the groove contact the LXCXE peptide (Liu et al., 1996). In contrast, E2F binds to an extended region that does not overlap the peptide binding site (Liu et al., 1996).

A variety of cellular proteins possess either the LXCXE or closely related motifs, including cyclinD1, HDAC1 and 2, and BRG1 (Dowdy et al., 1993; Ewen et al., 1993; Dunaief et al., 1994; Singh et al., 1995; Brehm et al., 1998). Both HDAC 1 and 2, and BRG1 can bind to pRb, and some evidence suggests that the LXCXE motif is involved in these interactions. Alternatively, it has been proposed that HDACs can associate with pRb indirectly, by forming a complex with the pRb binding protein RBP1 (Albert et al., 1999).

In this study, we have investigated the consequences of mutating the LXCXE binding domain in pRb, by altering key residues in pRb that are involved in binding to the LXCXE motif. We identified a mutant derivative of pRb, referred to as Rb6A, that exhibits significantly reduced binding to LXCXE-motif containing proteins, such as HDAC2, cyclin D1 and HPV E7 but retained LXCXE-independent binding, to E2F. Consistent with these results, Rb6A could down-regulate E2F-dependent activity, but failed to cause efficient transcriptional repression. Most importantly, Rb6A was effective in G1 arrest to a similar efficiency as wild-type pRb. Furthermore, directly targeting HDAC2 to E2F responsive genes failed to cause cell cycle arrest. Overall, LXCXE-dependent interactions, such as binding to HDAC, do not appear to be essential for pRb to exert growth control.

Results

Mutational analysis of the conserved lysine patch in the B domain of pRb

The strategy for preparing mutant derivatives of pRb was focused on a cluster of six conserved lysine residues in the B pocket, flanked by Lys713 and Lys765 (Figure 1a). This lysine cluster is highly conserved across metazoans, apart from Drosophila Rbx where an Arg replaces Lys at several positions (Figure 1a), a substitution that maintains the overall basic charge. The three dimensional crystal structure of the pRb pocket bound to an LXCXE motif-peptide derived from HPV E7 indicates that this cluster of Lys residues is scattered around the LXCXE hydrophobic binding grove, and might function in recognizing the LXCXE peptide (Figure 1b). In particular, Lys713 and Lys765 make contact with the backbone of the LXCXE peptide, and Lys720, 722, 729 and 740 contribute a cloud of positive charge which facilitates binding of the LXCXE motif to pRb (Lee et al., 1998). We anticipated that mutant derivatives prepared in this Lys rich region may lose the ability to bind to LXCXE-motif containing targets, like HDAC, but retain the ability to bind to the non-LXCXE motif target E2F. In turn, such mutants would allow us to assess the relevance of the Lys-rich region, and therefore LXCXE motif proteins, in pRb function.

In order to test the importance of these Lys residues we prepared a panel of mutants in which we had systematically altered each residue in the cluster from Lys to Ala, and thereafter measured their stability after transfection into 293 cells. Since each Lys residue is located on the surface of pRb (Lee et al., 1998), a substitution to Ala would be unlikely to cause a dramatic conformational alteration. In this respect, all the mutant derivatives were expressed to a similar level in various cell-types, including 293 cells (Figure 1d, and data not shown).

The pRb6A mutant derivative is compromised in binding to the LXCXE motif

We performed a variety of biochemical and cell-based assays to determine the properties of the pRb mutants, and present here relevant examples of the data. In the first series of experiments, each pRb mutant derivative was in vitro translated (Figure 2a) and thereafter assessed for binding to different GST-fusion proteins derived from protein targets that are known to interact with pRb, namely HPV E7, cyclinD1 and E2F-1. HPV E7 contains a classical LXCXE motif that allows high affinity pRb binding (Dyson, 1998), cyclinD1 contains a related motif, LLCCE, that has been implicated in pRb binding (Dowdy et al., 1993; Ewen et al., 1993), and E2F binds to pRb in an LXCXE-independent fashion (Lee et al., 1998).

There was a gradual decrease in the binding efficiency between GST-E7 and the in vitro translated pRb mutant derivatives. Whereas pRb1A1 and 2A exhibited a modest decrease in binding, pRb6A was significantly reduced in its binding efficiency to GST-E7 (Figure 2b). As a control, in vitro translated luciferase failed to interact with GST-E7 (Figure 2b). A similar analysis performed on the binding between pRb and cyclinD1 found that the binding activity was much less efficient than the E7/pRb interaction (Figure 2c). Nevertheless, a similar trend was observed, as pRb6A bound least efficiently to GST-cyclinD1 (Figure 2c). When the interaction between pRb and E2F-1 was analysed, whilst there was a modest decline in binding efficiency, pRb6A retained a good level of binding for E2F-1 at about 50% of wild-type pRb activity (Figure 2d). None of the pRb derivatives exhibited any binding to GST alone (Figure 2e).

As a complementary approach, we expressed each of the pRb mutant derivatives as a GST fusion protein and thereafter employed the purified protein (Figure 3a) in a binding assay performed in HeLa cell extracts to explore the effect of the mutations on the interaction between pRb and endogenous HeLa HDAC2 or E2F-1, in which HDAC2 represents a protein with a motif related to LXCXE, namely IACCE (Brehm, 1998). In a similar fashion to the other LXCXE-dependent interactions, HDAC2 exhibited a markedly reduced interaction with pRb6A (Figure 3b). Consistent with the results shown earlier, Rb6A continued to exhibit efficient binding to E2F-1, retaining about 50% wild-type pRb binding activity (Figure 3c).

In conclusion, the analysis of the binding properties of the pRb mutant derivatives provides support for the idea that the cluster of Lys residues demarked by Lys713 and Lys 765 plays an important role in LXCXE motif recognition. For example, pRb6A had lost the ability to interact with HDAC and, whilst reduced binding was observed between pRb and E2F-1, pRb6A could interact with significant levels of E2F-1.

The Rb6A mutant can down-regulate E2F-1-dependent transcription

The ability of pRb to regulate E2F activity is believed to be important for pRb tumour suppression (Dyson, 1998). However, it is not clear how important the contribution of pRb-dependent inactivation of E2F relative to pRb-dependent repression is for pRb tumour suppressor activity. To explore this question, we employed the pRb mutant derivatives described above in functional assays that measure E2F-dependent transcriptional activation and pRb-dependent repression, and thereafter studied their growth-regulating properties.

To measure E2F-dependent transcriptional activation we used the cyclinE promoter, which is regulated by E2F (Ohtani et al., 1995; Botz et al., 1996), and an artificial reporter construct, p3´WT, which contains three activating E2F binding sites taken from the adenovirus E2A promoter (Zamanian and La Thangue, 1992). Both reporter constructs were efficiently activated by exogenous E2F-1 in C33A cells, and could be subsequently down-regulated upon co-expression of wild-type pRb (Figure 4a,b). We compared the effect of Rb1A1 and Rb6A to wild-type pRb, and found that both mutant derivatives could inactivate E2F-1-dependent transcription on either promoter and, whilst there was a moderate decrease in their activity, both Rb1A1 and Rb6A possessed very similar activity to that exhibited by wild-type pRb (Figure 4a,b). These results therefore imply that the Lys rich cluster required for LXCXE binding does not impact on the ability of pRb to down-regulate E2F activity.

Next, we investigated the ability of the pRb mutant derivatives to repress transcription, since several studies suggest that transcriptional repression is relevant to pRb-dependent growth control (Weintraub et al., 1995; Chow and Dean, 1996; Brehm et al., 1998). To this end, representative mutants were fused to the Gal4 DNA binding domain and tested for their ability to repress the constitutive level of transcription from two promoters containing Gal4 binding sites, namely pG4-Ad ML-luc and pG4-TK-luc. We found on both promoters that Gal4-pRb repressed transcription effectively; in comparison, however, Gal4-Rb6A possessed a much reduced level of repression activity (Figure 4c,d). Similar results were also obtained from SAOS2 and U2OS cells (data not shown). These results imply that Rb6A has lost some of the repressive activity associated with wild-type pRb, a result which is compatible with the reduced ability of Rb6A to undergo LXCXE-dependent interactions, and the presence of the LXCXE-motif in HDAC1 and 2, both of which are believed to contribute to pRb-dependent repression (Brehm et al., 1998; Luo et al., 1998).

Rb6A cannot bind HDAC, but retains E2F binding activity in cells

To compare the interaction between wild-type pRb and the mutant derivatives, we transfected expression vectors encoding the Rb proteins together with either HDAC2 or E2F-1 into SAOS2 cells, and thereafter performed immunoprecipitation to assess the level of interaction between the exogenous proteins. The Rb proteins were expressed at equivalent levels (Figure 5a) and, as expected, wild-type pRb co-immunoprecipitated with HDAC2 (Figure 5b). In contrast, whilst an interaction with Rb1A1 was apparent, the ability of Rb6A to bind HDAC2 was very much reduced (Figure 5b). A similar result was seen for the interaction with HDAC1, which exhibited reduced binding to Rb6A (data not shown).

A similar analysis was performed to assess the interaction between the Rb derivatives and E2F-1. The binding of Rb1A1 and Rb6A to E2F-1 in cells was similar to the results derived from the in vitro biochemical binding assays, since both mutant derivatives continued to bind efficiently to E2F-1 (Figure 5c). Overall, therefore, the observations made in cells reflect the biochemical binding activity of the mutants, and argue strongly that Rb6A binds poorly to LXCXE-motif containing proteins, but retains the ability to interact with E2F-1.

The Rb6A mutant can induce G1 arrest

Since the Rb6A mutant was compromised in its ability to undergo LXCXE-motif dependent interactions, and had reduced repression activity, we used Rb6A as a means towards assessing the importance of pRb-dependent repression and LXCXE-motif binding in growth control mediated by pRb. In this respect, a widely used assay for measuring pRb-dependent growth arrest employs the over-expression of wild-type pRb in Rb-deficient SAOS2 human tumour cells, which undergo G1 arrest upon the introduction of wild-type pRb (Dick et al., 2000). When wild-type pRb was compared to Rb6A after expression in SAOS2 cells we found that both Rb proteins caused efficient cell cycle arrest; in fact, frequently we observed that Rb6A caused marginally greater levels of cell cycle arrest relative to wild-type pRb (Figure 6a). These results therefore imply that in SAOS2 cells the integrity of the Lys cluster which is mutated in Rb6A is not required for cell-cycle arrest. In turn, the results suggest that LXCXE-dependent interactions may not be essential for pRb-dependent growth arrest in SAOS2 cells.

As a further indication of the ability of Rb6A to cause cell cycle arrest, we performed a colony forming assay in SAOS2 cells where we compared the activity of wild-type pRb to Rb6A after drug selection of cells in G418. There was a clear and significant reduction in the number of colonies when transfected wild-type pRb was compared to the control empty vector, usually in the range of 80% reduction. Colony forming activity was apparent when Rb6A was assayed, which exhibited similar colony forming activity to wild-type pRb (Figure 6b). Moreover, both wild-type pRb and Rb6A protein were expressed at similar levels in colonies derived from the selection procedure (Figure 6c). Overall, these two different assays that measure pRb-dependent growth control, namely flow cytometry and colony forming activity, strongly suggest that pRb6A has similar growth regulating activity to wild-type pRb.

HDAC and cell cycle arrest

Whilst the results indicate that pRb can cause cell cycle arrest in the absence of efficient LXCXE-dependent interactions, it was nevertheless possible that certain LXCXE-containing targets such as HDAC, were themselves capable of causing cell cycle arrest through the regulation of E2F-dependent transcription. In previous studies, a hybrid protein in which the pRb pocket region was fused to E2F could down-regulate E2F target genes and cause cell cycle arrest (Sellers et al., 1995). It was of interest therefore to establish whether a hybrid protein where HDAC, rather than pRb, was fused to E2F could affect cell cycle progression. To pursue this question, we prepared two HDAC hybrids, the first with HDAC2 fused to an E2F protein lacking the C-terminal region (residue 1-368) and the second with HDAC fused to the same region of E2F, but containing a mutation at residue 132 which inactivates E2F DNA binding activity (Figure 7a; Sellers et al., 1995). In transfected cells, E2F wt/HDAC and E2F mt/HDAC hybrid proteins were expressed at similar levels (Figure 7b), and E2F wt/HDAC was capable of repressing transcription of an E2F-dependent reporter, in contrast to the E2F DNA binding domain (E2F DBD) which, as previously reported (Qin et al., 1995; Zhang et al., 1999), activated transcription (Figure 7c). In order to assess further the integrity of the hybrid proteins, we determined whether they could physically interact with wild-type pRb, and compared the level of interaction to that occurring with the naturally-occurring pocket mutant pRbDelta22. Both E2F wt/HDAC and E2F mt/HDAC exhibited specific and equivalent binding to GST-Rb, and decreased binding to GST-RbDelta22, as expected for an LXCXE-dependent interaction (Figure 7d).

Thus, having established the integrity of the E2F-HDAC hybrids, we investigated any effects on cell cycle progression by flow cytometry. As noted previously (Sellers et al., 1995), the E2Fwt/pRb hybrid caused effective cell cycle arrest when introduced in SAOS2 cells (Figure 7e). In contrast, we failed to observe significant effects on cell cycle progression by E2F wt/HDAC (Figure 7e), even though the expression of the HDAC hybrid was at similar levels to endogenous HDAC (Figure 7b). These results therefore suggest that targeting HDAC to E2F-dependent genes is not sufficient to cause cell cycle arrest. In turn, the results concur with and corroborate the observations made earlier on Rb6A, which caused efficient cell cycle arrest in the absence of efficient LXCXE-dependent interactions, and therefore imply that the interaction between pRb and LXCXE-motif containing proteins may not be essential for pRb-dependent cell cycle arrest.

Discussion

The LXCXE-motif binding domain in pRb

By mutating key lysine residues in the LXCXE-motif binding domain of pRb, we created a panel of mutant derivatives in pRb, and from this panel identified Rb6A, that is severely compromised in binding to LXCXE-motif containing targets, including HPV E7, HDAC2 and cyclinD1. Although Rb6A could effectively down-regulate E2F activity, it failed to exhibit significant levels of transcriptional repression activity when assayed in the context of a Gal4-pRb hybrid. The most likely explanation for the reduced level of repression is that of a diminished interaction with an important LXCXE-motif containing cellular target that contributes to pRb-dependent transcriptional repression. In this respect, likely candidates include certain members of the HDAC family, notably HDAC1 and 2, which possess an LXCXE-like motif, and have been implicated in repression mediated by pRb (Brehm et al., 1998; Luo et al., 1998; Zhang et al., 2000). It is possible therefore that pRb6A is a loss-of-function mutant that has reduced capacity to undergo LXCXE-dependent interactions.

In this respect, there is a considerable body of evidence which implicates E2F as an important target in pRb-dependent growth control. For example, a good correlation exists with E2F binding to pRb, and pRb-dependent growth arrest (Dyson, 1998). Increased levels of E2F can overcome pRb-dependent growth arrest (Hiebert et al., 1992), and hypo-phosphorylated pRb, which is active in growth control, associates with E2F in non-cycling arrested cells (Mittnacht, 1998). These studies, and others, argue strongly that pRb-dependent tumour suppression involves the regulation of E2F activity.

Transcriptional repression by pRb

The role of transcriptional repression in pRb-dependent tumour suppressor activity is less clear. pRb can bind to different groups of proteins involved in modulating chromatin, including HDACs, RBP1, and components of the SWI/SNF complex (Brehm et al., 1998; Luo et al., 1998; Magnaghi-Jaulin et al., 1998; Woitach et al., 1998; Zhang et al., 2000), and actively repress for example enhancer driven transcription (Weintraub et al., 1995). Less is known about the physiological importance of pRb repression activity, although some studies have suggested that repression may be relevant in cell cycle control (Weintraub et al., 1995; Zhang et al., 1999). For example, a dominant-negative version of E2F-1 which lacks the transcriptional activation domain can overcome pRb-dependent growth arrest (Qin et al., 1995). However, it is not known whether cells rescued in this fashion are liberated to undergo normal cell cycle progression, and if so whether this truly occurs in the absence of transcriptional activation by physiological E2F.

The results presented in this study derived from analysing the Rb6A mutant bare on the role of LXCXE-binding domain in pRb for pRb-dependent growth control. Our results indicate that whilst Rb6A cannot bind to HDAC2, both in vitro and in vivo, and exhibits reduced repression activity, it retains E2F binding activity. Furthermore, Rb6A could both down-regulate E2F-dependent transcription and, most importantly, effect cell-cycle arrest in SAOS2 cells in two distinct assays that measure pRb-dependent growth control. The ability of Rb6A to cause G1 arrest occurred as efficiently as that observed for wild-type pRb and, similarly the colony-forming activity of Rb6A was seen to be comparable to wild-type pRb. Taken together, Rb6A is a mutant derivative that dissociates LXCXE-dependent effects from E2F regulation and cell cycle arrest. Because Rb6A can effectively impede cell cycle progression these results imply that LXCXE-dependent interactions are not essential for cell cycle arrest.

Several other studies have addressed the role of LXCXE-binding in pRb (Chen and Wang, 2000; Dahiya et al., 2000; Dick et al., 2000). At a general level the results presented here are consistent with previous reports, namely that LXCXE-dependent interactions are not essential for pRb to exert cell cycle arrest. At a specific level, however, the nature and properties of the pRb mutant derivatives analysed in this study exhibit considerable differences to those described earlier. For example, Dahiya et al. (2000) described a series of pRb mutants that failed to undergo LXCXE-dependent interactions and, whilst competent to stimulate cell cycle arrest, the arrest was reported to be transient and not apparent in longer term assays, such as colony forming activity. Furthermore, Dick et al. (2000) mutated the LXCXE-binding domain to obtain mutant derivatives that retained interactions with HDAC1, were able to actively repress transcription, and cause cell cycle arrest. The mutants described in our study differ both in the residues that have been altered, together with the properties described in the biochemical and cell-based assays. As such, this study represents new information and provides a further level of resolution on the importance and role of residues within the LXCXE-binding domain in pRb.

HDAC and pRb

Our results raise the possibility that LXCXE-dependent interactions and transcriptional repression are not necessary for Rb-mediated cell cycle arrest, at least within the limits of the assays employed here. This view is supported by the results derived from studying the E2F/HDAC hybrid proteins, which failed to exhibit cell cycle arrest activity. In contrast to pRb, targeting HDAC to E2F-dependent genes would appear not to be sufficient to alter cell cycle progression. This result is consistent with the properties of Rb6A, which is compromised in HDAC2 binding, yet retains the ability to cause cell cycle arrest. It is possible that HDAC binding to pRb contributes to other physiological outcomes of pRb activity. In this respect, it is noteworthy that pharmaceutical intervention of HDAC activity can alter cellular differentiation (Hoshikawa et al., 1994; Chen and Wang, 2000; Han et al., 2000).

Perhaps the role of HDAC in the pRb complex is related to attaining an irreversible state of differentiation. Analysis of the phenotype of Rb-/- knockout mice suggests that pRb is required for differentiation in diverse cellular lineages, including the haematopoietic and central nervous systems (Mulligan and Jacks, 1998). A model that explains our results suggests that LXCXE-dependent interactions are not essential for pRb-dependent cell cycle arrest. Rather, subsequent physiological stimuli, perhaps those that act in the G1 phase and promote differentiation, may require interactions between pRb and LXCXE motif containing proteins, such as HDAC, possibly resulting in an outcome that maintains irreversible and stable changes in gene expression.

Materials and methods

Plasmids

pcDNA-HA-E2Fwt-Rb, pcDNA-HA-E2Fmut-Rb, pE2F DBD, pCMV-CD20, pCMV-betagal, pCMV-CycE-Luc, pGEX-Rb (379-929), pGEX-RbDelta22 (379-928), pFLAG-HDAC2, pcDNA3-9E10 Rb, and pCMV-HA-E2F-1, pG4-TK-luc, pG4-AdML-luc and p3´WT-luc reporters have previously been described (Zamanian and La Thangue, 1992; Qin et al., 1995; Sellers et al., 1995; Laherty et al., 1997; Lee et al., 1998; Shikama et al., 2000). To construct the E2F-HDAC2 chimeras, HDAC2 was amplified by PCR from FLAG-HDAC2 using the primers (5'-TAATGAATTCGCGTACAGTCAAGGAGGCGGC-3', 5'-GAGATCTAGATCAAATTCAAGGGTTGCTGAGTTG-3') and the incorporated EcoRI/XbaI sites were used to subclone HDAC2 into pM2 (Weintraub et al., 1995). E2F (1-368)-Rb (379-792) or E2F-1 (1-368; mt132)-Rb (379-792) were used as the templates to amplify a 250 bp E2F1 3' - fragment (5'-TTTCAGATCTCCCTTAAGAGCAAAC-3'; 5'-GTGAATTCGGGAGCCCGCAGGCTG-3') which was cloned upstream of the HDAC2 in the pM2 vector at the BglII/EcoRI site. The chimeras were subsequently DNA sequenced. Quickchange site-directed mutagenesis kit (Stratagene) was used to create the panel of pRb point mutants with pcDNA3-9E10Rb as the template. Primers used were: K713A (5'-GTGTTCCATGTATGGCATATGCGCAGTGAAG
AATATAGACC-3'; 5'-GGTCTATATTCTTCACTGCGCATATGCCATA
CATGGAACAC-3'), K765A (5'-CGGTCTTCATGCAGAGACTGGCAACAAATA
TTTTGCAGTATGC-3' 5'-GCATACTGCAAAA TAT TTG TTGC CAGT CTCTG CAT GAAGAC CG-3'), K740A (5'-GCTGTTCAGGAGACATTCGCACGTGTTTTG
ATCAAAGAAGAG-3'; 5'-CTCTTCTTTGATCAAAACACGTGCGAATGTCTCCTGA
ACAGC-3'), K720/722A (5'-GCAGTGAAGAATATAGACCTTGCATTCGCAA
TCATTGTAACAGCATACAAGG-3'; 5'-CCTTGTATGCTGTTACAA TGATTGCGAATG CAAGG TCTATATTCTTCACTGC-3'), and K729A (5'-CATTGTAACAGCATACGCGGATCTTCCTCA
TGCTGTTCAG-3';
5'-CTGAACAGCATGAGGA AGATCC GC GTATGCTG TTACA ATG-3'). To generate the pG4-Rb constructs, the EcoRI fragments from the pcDNA3-9E10-Rb or mutant derivatives were sub-cloned into the XhoI site of pG4polyMII backone (Weintraub et al., 1995). All pG4-Rb constructs included the pRb sequence from amino acid 300-928.

Cell culture and transfection

SAOS2, U2OS and C33A cells were grown in DMEM containing 10% foetal calf serum. For transfection, 4´106 C33A cells and 106 SAOS2 or U2OS cells were transfected as described (Shikama et al., 2000). For the reporter assays, 1 mug of the cyclinE reporter or 3´WT-luc reporter was used, together with 100 ng of pCMV-E2F-1 and 10 mug of 9E10-pcDNA3-pRb or mutant derivatives were transfected into the cells. For the Gal4-pRb assay, 10 mug of pG4-Rb or mutant derivative expression vector together with 1 mug of pG4-AdML-luc or 100 ng of pG4-TK-luc reporter plasmids were transfected into cells. All transfections included pCMV-beta-gal (1 mug) as an internal control.

Immunoblotting, immunoprecipitation and binding assays

In transfected cells, the HA-tagged E2F hybrids were detected using Mab HA.11 (Cambridge Bioscience) and endogenous HDAC2 detected using HDAC2 antibody H54 (Santa Cruz). For the E2F-HDAC binding assay with GST-Rb, E2Fwt-HDAC2 and E2Fmt-HDAC2 were in vitro transcribed and translated (Promega), and equal amounts of 35S labelled E2F-HDAC protein was incubated with 2 mug of purified GST-Rb (379-928) or GST-RbDelta22(379-928), in TNN buffer (50 mM Tris pH 7.4, 0.5% NP40, 250 mM NaCl, 5 mM EDTA, 10% glycerol) at 4°C for 2 h. The GST beads were then washed three times with TNN buffer (150 mM NaCl) and analysed by SDS-PAGE.

GST-Rb and mutant derivatives (5 mug) were incubated with a HeLa nuclear extract (1 mug) and anti-HDAC2 antibody (Santa Cruz) or KH95 (Santa Cruz) were used to detect HDAC2 and E2F-1 respectively. In the reverse assay, pRb and mutant derivatives were in vitro transcribed and translated and equal amounts of 35S labelled polypeptide incubated with the appropriate GST fusion protein (about 5 mug). For both assays, the incubation was carried out in IPH buffer (50 mM Tris-pH8.0, 150 mM NaCl, 5 mM EDTA, 0.5% NP-40, protease inhibitor cocktail) at 4°C for 2 h, and washed three times in IPH buffer. Quantitation of binding activities was carried out using a Kodak Scientific Imaging System (Kodak Digital Science).

Flow cytometry and colony forming activity assay

106 SAOS2 cells seeded in a 10 cm dish were transfected with at least 10 mug of the indicated expression vectors together with 8 mug of pCMV-CD20. At 48 h after transfection cells were harvested in Cell Dissociation Buffer (Sigma) and incubated with FITC-conjugated CD20 antibody (Becton-Dickinson) to identify the transfected cell population. The cells were then washed in PBS and fixed overnight at 4°C in 50% ethanol. Cells were washed again and treated with RNaseA (25 U/ml) and propidium iodide solution (10 ng/ml) for 20 min at 4°C. Cells were sorted on a FACScan cell sorter (Becton-Dickinson) and analysed using the Cell Quest Software package.

For the colony forming activity assay, about 106 SAOS2 cells were transfected with 10 mug of the indicated expression vectors. At 48 h after transfection, cells were trypsinized, and 2´104 or 4´104 transfected cells were mixed with about 106 untransfected cells and replated. The cells were selected in media containing G418 (0.8 mug/ml) for 16 days until colonies became visible. Cells were fixed and stained in crystal violet. All experiments were performed in duplicate, and counting was performed double blind.

Acknowledgements

We thank Marie Caldwell for help in preparing this manuscript. This work was supported by the Medical Research Council, Wellcome Trust, Leukaemia Research Fund and Cancer Research Campaign.

References

Albert L, Lee JM, Yang W-M, DeCaprio JA, Kaelin Jr WG, Seto E, Branton PE. (1999). Mol. Cell. Biol. 19, 6632-6641. MEDLINE

Botz J, Zerfass-Thome K, Spitkovsky D, Delius H, Vogt B, Eilers M, Hatzigeorgiou A, Jansen-Dürr P. (1996). Mol. Cell. Biol. 16, 3401-3409. MEDLINE

Brehm A, Miska EA, McCance DJ, Reid JL, Bannister AJ, Kouzarides T. (1998). Nature 391, 597-601. Article MEDLINE

Chen T-T, Wang JYJ. (2000). Mol. Cell. Biol. 20, 5571-5580. MEDLINE

Chow KNB, Dean DC. (1996). Mol. Cell. Biol. 16, 4862-4868. MEDLINE

Dahiya A, Gavin MR, Luo RX, Dean DC. (2000). Mol. Cell. Biol. 20, 6799-6805. MEDLINE

Dick FA, Sailhamer E, Dyson NJ. (2000). Mol. Cell. Biol. 20, 3715-3727. MEDLINE

Dowdy SF, Hinds PW, Louie K, Reed SI, Arnold A, Weinberg RA. (1993). Cell. 73, 499-511. MEDLINE

Dunaief JL, Strober BE, Guha S, Khavari PA, Alin K, Luban J, Gegemann M, Crabtree GR, Goff SP. (1994). Cell. 79, 119-130. MEDLINE

Dyson N. (1998). Genes Dev. 12, 2245-2262. MEDLINE

Ewen ME, Sluss HK, Sherr CJ, Matushime H, Kato J, Livingston DM. (1993). Cell. 73, 487-497. MEDLINE

Goodman RH, Smolik S. (2000). Genes Dev. 14, 1553-1577. MEDLINE

Grunstein M. (1997). Nature 389, 349-352. Article MEDLINE

Han JW, Ahn SH, Park SH, Wang SY, Bae GU, Kwon HK, Hong S, Lee HY, Lee YW, Lee HW. (2000). Cancer Res. 60, 6068-6074. MEDLINE

Harbour JW, Luo RX, Dei S, Postigo AA, Dean DC. (1999). Cell. 98, 859-869. MEDLINE

Harbour JW, Dean DC. (2000). Genes Dev. 14, 2393-2409. Article MEDLINE

Hassig CA, Schrieber SL. (1997). Curr. Opin. 1, 300-308.

Hiebert SW, Chellappan SP, Horowitz JM, Nevins JR. (1992). Genes Dev. 6, 177-185. MEDLINE

Hoshikawa Y, Kwon HJ, Yoshida M, Horinouchi S, Beppu T. (1994). Exp. Cell. Res. 214, 189-197. MEDLINE

Kingston RE, Narlikar GJ. (1999). Genes Dev. 13, 2339-2352. Article MEDLINE

Laherty CD, Yang WM, Sun JM, Davie JR, Seto E, Eisenman RN. (1997). Cell. 89, 349-356. MEDLINE

Lee CW, Sørensen TS, Shikama N, La Thangue NB. (1998). Oncogene 16, 2695-2710. MEDLINE

Lee JO, Russo AA, Pavletich NP. (1998). Nature 391, 859-865. Article MEDLINE

Liu N, Lucibello FC, Zwocker J, Engeland K, Müller R. (1996). Nucleic Acids Res. 24, 2905-2910. MEDLINE

Luo RX, Postigo AA, Dean DC. (1998). Cell. 92, 463-473. MEDLINE

Magnaghi-Jaulin L, Groisman R, Naguibneva I, Robin P, Lorain S, Le Villain JP, Troalen F, Trouche D, Harel-Bellan A. (1998). Nature 391, 601-605. Article MEDLINE

Mittnacht S. (1998). Curr. Opin. Genet. Dev. 8, 21-27. MEDLINE

Moran E. (1993). Curr. Opin. Genet. Dev. 3, 63-70. MEDLINE

Mulligan G, Jacks T. (1998). Trends Genet. 14, 223-229. Article MEDLINE

Ohtani K, DeGregori J, Nevins JR. (1995). Proc. Natl. Acad. Sci. USA 92, 12146-12150. MEDLINE

Qin X-Q, Livingston DM, Ewen M, Sellers WR, Arany Z, Kaelin Jr WG. (1995). Mol. Cell. Biol. 15, 742-755. MEDLINE

Ross JF, Liu X, Dynlacht BD. (1999). Mol. Cell. 3, 195-205. MEDLINE

Sellers WR, Rodgers JW, Kaelin Jr WG. (1995). Proc. Natl. Acad. Sci. USA 92, 11544-11548. MEDLINE

Shikama N, Lyon J, La Thangue NB. (1997). Trends Cell. Biology 7, 230-236.

Shikama N, Chan HM, Krstic-Demonacos M, Smith L, Lee CW, Cairns W, La Thangue NB. (2000). Mol. Cell. Biol. 20, 8933-8943. MEDLINE

Singh P, Coe J, Hong W. (1995). Nature 374, 562-565. MEDLINE

Tommasi S, Pfeifer GP. (1995). Mol. Cell. Biol. 15, 6901-6913. MEDLINE

Trouche D, Kouzarides T. (1996). Proc. Natl. Acad. Sci. USA 93, 1439-1441. Article MEDLINE

Weintraub SJ, Chow KNB, Luo RX, Zhang SH, He S, Dean DC. (1995). Nature 375, 812-815. MEDLINE

Woitach JT, Zhang M, Niu C-H, Thorgeirsson SS. (1998). Nat. Genet. 19, 371-374. Article MEDLINE

Zamanian M, La Thangue NB. (1992). EMBO J. 11, 2603-2610. MEDLINE

Zhang H, Gavin M, Dahiya A, Postigo AA, Ma D, Luo RX, Harbour JW, Dean DC. (2000). Cell. 101, 79-89. MEDLINE

Zhang HS, Postigo AA, Dean DC. (1999). Cell. 97, 53-61. MEDLINE

Zwicker J, Liu N, Engeland K, Lucibello FC, Muller R. (1996). Science 271, 1595-1597. MEDLINE

Figures

Figure 1 Strategy for mutagenesis of the pRb pocket lysine patch. (a) Sequence alignment of pRb B box from various species, together with the Rb consensus sequence at the bottom. The six conserved lysines mutated in the study are indicated in red, with their numerical location (in human pRb) indicated above. (b) The three dimensional structure of the cluster of conserved lysine residues around the LXCXE-peptide binding groove. This picture was constructed by using Rasmol with the co-ordinates published (Lee et al., 1998), and shows the E7 LXCXE peptide bound to pRb. Lys713 and Lys765 (blue) make contact with the LXCXE peptide, whereas the other lysines (red) may facilitate binding to the motif. Purple in the LXCXE peptide indicates conserved residues. (c) Nomenclature of the mutant derivatives generated in this study. In all cases, lysine residues (K) were altered to alanine residues (A) (d) An immunoblot showing the expression levels of the 293 HEK cells. The indicated pRb expression vectors (10 mug) were transfected into 293 HEK cells, and 100 mug of the total cell extract was used for the immunoblot with anti-pRb monoclonal antibody to measure Rb levels. Molecular weight standards are indicated, and * indicates a non-specific polypeptide

Figure 2 Binding properties of Rb mutant derivatives. (a) The lanes show the input of wild-type pRb, mutant derivatives and luciferase protein generated by in vitro transcription and translation radiolabelled with 35S methionine, which was used in the following binding studies. (b) GST-E7 (5 mug; bottom panel shows the Coomassie blue stained purified protein) was used in a binding assay with the indicated in vitro translated Rb proteins and luciferase. (c) GST-cyclin D1 (5 mug; bottom panel shows the Coomassie blue stained purified protein) was used in a binding assay with the indicated in vitro translated Rb proteins and luciferase. (d) GST-E2F-1 (5 mug; bottom panel shows the Coomassie blue stained purified protein) was used as a binding assay with the indicated in vitro translated Rb proteins and luciferase. (e) GST (5 mug; bottom panel shows the Coomassie blue stained purified protein) was used as a binding assay with the indicated in vitro translated Rb proteins and luciferase

Figure 3 Binding of pRb and mutant derivatives to HeLa cell HDAC2 and E2F-1. (a) Coomassie blue stain of purified GST, GST-pRb and mutant derivatives (as indicated by *). (b) and (c) The indicated GST proteins (5 mug) were incubated with a HeLa cell extract (1 mg), and the amount of HDAC2 (b) and E2F-1 (c) bound to the GST fusion proteins assessed by immunoblotting with either anti-HDAC2 (b) or anti-E2F-1 (c) as indicated

Figure 4 Functional properties of pRb and mutant derivatives. (a) and (b) Expression vectors for pRb, RbA1 and Rb6A together with (10 mug) E2F-1 (100 ng) were co-transfected into C33A cells as indicated, and the reporter activity from pCycE-luc (a) and p3´WT-luc (b) assayed. The values shown represent the ratios (luc/beta-gal). The activity of E2F-1 alone was assigned the arbitrary value of 100. (c) and (d) Expression vectors for Gal4-Rb and Gal4-Rb6A proteins (10 mug) were expressed in C33A cells, and the reporter activity from pG4-AdML-luc (c) and pG4-TK-luc (d) was assayed as indicated. The Gal4 control transfection (lane 1), which represented the unrepressed state of the reporter, was normalized to 100% of activity, and the values were calculated as described for (a). The data shown in (a), (b), (c) and (d) are representative of at least three different experiments

Figure 5 Binding of Rb mutant derivatives to HDAC and E2F in cells. SAOS2 cells were transfected with expression vectors for wild-type pRb, Rb1A1 or Rb6A tagged with the myc epitope (10 mug) together with HDAC2 (100 ng) or E2F-1 (100 ng) expression vectors. Immunoprecipitation was performed with anti-myc monoclonal antibody (9E10) followed by immunoblotting with anti-Rb (a), anti-HDAC2 (b) or anti-E2F-1 (c). Tracks 1 and 5 represent the empty vector transfected cells

Figure 6 Growth-regulating properties of Rb mutant derivatives. (a) SAOS2 cells were transfected with expression vectors for pRb or pRb6A (10 mug), together with pCMV-CD20 (10 mug). Cells were fixed and stained for CD20 and propidium iodide and analysed by flow cytometry as described. The percentage change of cells (compare to a mock transfected sample with pcDNA3) in different phases of the cell cycle is shown. (b) About 1´106 SAOS2 cells were transfected with expression vectors for pRb or pRb6A (10 mug). After 48 h of transfection, about 2´104 or 4´104 transfected cells were mixed with 1´106 untransfected SAOS2 cells, and selected for colony growth with G418 (0.8 mug/ml) for 16 days. Colonies were stained with crystal violet and counted. The percentage of colonies relative to the pcDNA3 empty vector control (10 mug) treatment is shown. (c) Immunoblot of extracts from colonies derived from the colony assay described in (b) transfected with either empty vector (track 1), pRb (track 2) or Rb6A (track 3) is shown

Figure 7 Role of HDAC in cell cycle control. (a) Schematic diagram of the E2F-HDAC2 chimeric proteins. E2Fmt/HDAC contains a mutation that alters residue 132 in the E2F DNA binding domain and prevents DNA binding. (b) Immunoblot of E2Fwt-HDAC2 and E2Fmut-HDAC2 chimeras after transfection in C33A cells, probed with anti-HDAC2 (tracks 2, 4 and 6) and anti-HA (tracks 1, 3 and 5) antibodies as indicated. Tracks 1 and 2 show untransfected cell extracts. (c) C33A cells were transfected with the indicated expression vectors (1 or 3 mug), namely E2F DBD or E2Fwt/HDAC, together with the reporter p3´WT-luc. The values shown represent the ratio (luc/beta-gal), and are representative of at least three different experiments. (d) In vitro binding analysis of S35-labelled E2F-HDAC2 chimeras to the purified GST, GST-Rb or GST-RbDelta22 proteins indicated. (e) Cell cycle analysis of the chimeric E2F proteins in SAOS2 cells. SAOS2 cells were transfected with the indicated chimeric expression vectors (10 mug) namely E2Fwt/Rb, E2Fmt/Rb, E2Fwt/HDAC or E2Fmt/HDAC, and the cell cycle profile analysed by FACS. The percentage change of cells (compared to mock transfected cells with pcDNA3) in different phases of the cell cycle is shown

Received 26 April 2001; revised 23 June 2001; accepted 5 July 2001
27 September 2001, Volume 20, Number 43, Pages 6152-6163
Table of contents    Previous  Article  Next    [PDF]