Cancer is a complex lethal disease that is only now being understood on the molecular level. Primarily due to the variety and complexity of tumors, cancer treatment has previously been approached only through direct therapies aimed at blocking proliferation or inducing apoptosis of any rapidly dividing cell population. However, rapid advances in cancer research have increasingly revealed many of the mechanisms of the tumorigenic process. This knowledge is now being exploited in order to design targeted therapies that simultaneously leave normal cells relatively unaffected while inhibiting growth through interference in specific mechanisms required for tumorigenesis.
Carcinogenic mutations occur in two predominant categories of either oncogenic and/or tumor suppressor mutation. Oncogenes such as Ras, myc, Erb-B2, Bcl-2, and Abl are derived from normal cellular growth and survival genes (proto-oncogenes), that when mutated may produce abnormal cellular division. Conversely, tumor suppressor genes such as p53 are often inactivated through mutation or suppression, both post-transcriptionally and translationally, in cancer. This is further complicated when considering the fact that malignancies often display multiple mutations, harbor epigenetic alterations, and contain chromosomal abnormalities, aneuploidy, and loss of heterozygosity at numerous loci.
Many tumors rely on continual activation of certain cancer-promoting genes, termed "oncogene addiction" (Weinstein and Joe, 2006). Analysis of signaling networks involved in tumor maintenance, while confirming complexity, has also shown that the tools for their own destruction are stored within. Studies in transgenic mice, that overexpress a particular oncogene in a specific target tissue, have shown that activation can lead to tumor development, but when subsequently switched off the tumor cells stop dividing, display differentiation, and apoptosis (Felsher and Bishop, 1999; Huettner et al., 2000; Fujita et al., 1999; Jackson et al., 2001). Drug design has largely exploited this dependence with inhibitors, which have already proven to be effective in treatment, such as the EGFR-targeted drugs gefitinib (Iressa) and erlotinib (Tarceva) in non-small-cell lung carcinoma (NSCLC), pancreatic cancer, and glioblastoma (Grunwald and Hidalgo, 2002). Although oncogene addiction is true in many cases, cancers may break free of this dependence likely due, at least in part, to their genomic instability. Therapy directed at multiple targets has proven more effective than use of individual agents; thus producing synergistic or additive effects upon treatment.
Restoring the loss of a tumor suppressor, either transcriptionally or translationally, is usually an attractive target for combination with oncogene suppression. p53 is necessarily one of the most appealing as it is the "guardian of the genome" (Lane, 1992). Therefore, understanding the ways in which p53 may be affected upon interaction with its transcriptional effecters and just as important, the characteristics of these interaction(s), is important in the process of developing ways to restore wild type phenotypes.
A. Tumor Suppressor p53
p53 is a 53-kDa-tumor suppressor mutated in about half of all human tumors (Lomax et al., 1998; Bardeesy and DePinho, 2002) with over 80% of these mutations falling within the DNA-binding domain (Hainaut, 2002). However, several somatic and germ-line mutations have been observed in the tetramerization domain (Lomax et al., 1997; Varley et al., 1996) that could potentially affect tetramerization and thus biological activity (Chene, 2001). p53 plays a central role in a complex signaling network evolved to mediate a variety of cytotoxic and genotoxic stressors that have the potential to compromise genomic stability and promote neoplastic transformation in the absence of action through the p53-signaling network.
p53 was first identified in 1979 in complex with simian virus 40 (SV40) large T-antigen and under normal conditions is maintained at low levels by continuous ubiquitylation and subsequent degradation by the 26S proteasome (Haupt et al., 1997; Kubbutat et al., 1997). When activated by stress signals such as DNA damage, hypoxia, oncogene expression, viral infection and/or ribonucleotide depletion, p53 levels stabilize and exert its role of "guardian of the genome" (Lane, 1992), acting as a transcription factor coordinating a program eventually leading to cell cycle arrest, senescence or apoptosis (Vousden and Lu, 2002).
This view of p53 activity is overly simplified when considering activities where p53 conducts transcriptionally independent functions within the apoptotic response and is involved in cellular processes of metabolism, autophagy, DNA repair and the antioxidant response to increased ROS levels (Crighton et al., 2006; Sablina et al., 2005; Zhou et al., 2001). For example, p53 is over-expressed under conditions of oxidative stress, in an attempt to counteract DNA damage until the cell is no longer viable; it then activates the apoptotic response through induction of genes such as Bax, Bcl2, and IGF2. Once activated, p53 recognizes multiple promoters with pivotal roles in growth arrest, repair, apoptosis, senescence, and differentiation (Sphyris and Harrison, 2005).
The central role of p53 in tumor suppression is further highlighted given that direct inactivation of this gene is the most common mutation in human cancer, occurring in more than 50% of malignancies (Bardeesy and DePinho, 2002). Its knockdown promotes genome instability by uncoupling chromosome duplication from the cell cycle at the G1-S phase checkpoint (Canman et al., 1994). For example, in Li Fraumeni syndrome where one p53 mutant allele is germ-line inherited, afflicted individuals show a 25-fold increase in the occurrence of developing early-onset cancers compared with the general population (Evans and Lozano, 1997). Additionally, other components of the p53 network frequently are altered in wild type p53 tumors; thus it is not unreasonable to contend that the p53 pathway is compromised to some extent in all tumors. For these reasons, p53 is an ideal target in the study of pathways active in carcinogenesis and the molecules that affect its activities through their association.
1. Restoration of p53
Restoration of p53 has been shown to be effective in tumor cell growth inhibition. For example, p53 conditional knockout mice that were chemically knocked-in led to restoration of p53 expression and subsequent rapid inhibition of tumor growth. Other ways in which the p53 pathway can be reactivated include, Nutlins, compounds that bind MDM2 in the p53 binding pocket with high selectivity and release p53 from negative regulation leading to effective stabilization of p53 and activation of the p53 pathway in vitro and in vivo (Thompson et al., 2004). DNA-methyl transferase inhibitors such as azacytidine that reverse epigenetic silencing genes participating in the p53 pathway (Fang et al., 2004), and restoration of wild type characteristics have been achieved with small molecules PRIMA-1 and CP-31398 (Bykov et al., 2002; Demma et al., 2004; Rippin et al., 2002; Wang et al., 2006). However, exactly how p53 carries out its tumor suppressive function seems to differ according to the tumor type and niche.
For example, restoring p53 function in p53-deficient lymphomas rapidly induces apoptosis (Martins et al., 2006; Ventura et al., 2007). Whereas, p53 restoration in solid soft tissue sarcoma and hepatocellular carcinoma induces a potent growth arrest characterized by the hallmarks of cellular senescence (Ventura et al., 2007; Xue et al., 2007). It is therefore unclear which features determine whether a tumor's response to p53 activation is apoptosis or senescence. Notably, p53 activation occurred only in cancer cells and not in normal cells that likely lack the specific environment necessary to activate the p53 pathway (Ventura et al., 2007). Consequently, it seems that p53 restoration can be effective in therapies directed toward selectively killing tumor cells relative to their normal counterparts.
2. Regulation of p53
What stimuli lead to the activation of p53 activity and which modifications and interactions are essential for p53 function? What are the mechanisms by which p53 specifically selects different subsets of target genes ultimately leading to distinct cellular outcomes? These are some of many interesting questions which future research will undoubtedly focus on in order to achieve greater insight into the effects that p53 has on cellular biology.
Depending on the stimulus, p53 undergoes a series of post-translational modifications at specific residues affecting its stability, sub-cellular localization, ability to interact with different co-factors, and bind to its target promoters (Gostissa et al., 2003; Vousden and Lu, 2002). Therefore, p53 activity is tightly regulated through posttranslational mechanisms, most notably through Mdm2 cytosolic degradation, and at the pre- and posttranscriptional levels. Posttranslational regulation of p53 primarily occurs through binding of numerous effecter molecules in the intrinsically disordered N and C-terminal domains.
Although p53 is mutated in more than 50% of all human tumors and leads to direct inactivation of this gene (Lomax et al., 1998; Bardeesy and DePinho, 2002), it cannot be ignored that there are still mutated variations that continue to be expressed. However, there are currently no published accounts of epigenetic inactivation through hyper-methylation of mutant p53 promoters. Therefore, this would indicate other, as of yet unknown, distinct functions for p53 mutants.
3. p53 transcription factor family
The p53 family, p53 and its homologues p63 and p73, belong to a family of transcription factors that have high structural similarity (Kaghad et al., 1997; Yang et al., 1998). All three contain a central DNA binding domain (DBD), N-terminal transactivation domain (TAD) and tetramerization domain (TD). The apparent homology shown by p53, p63 and p73 was at first suggested to convey similar or even redundant function(s). However, prior to discovery of p53 isoforms, it was thought that only p63 and p73 could be expressed as different variants. p63 and p73 are transcribed from two major promoters that may be expressed with or without the N-terminal TA domain (Stiewe et al., 2002). However, differences in the regulation and activity of each suggest that they may be critical regulators of distinct processes. In particular, p53 is mainly involved in the DNA-damage response while p63 and p73 play key roles in development.
Phylogenetic analysis demonstrated that the primordial ancestor of all three genes is actually much more related to p63 and p73, while p53 appears as the most recently evolved member (Yang et al., 2002). Therefore, it is likely that p53 lost some ancestral function(s), currently conducted by its homologues, while gaining its tumor-suppressive abilities. This seems to be confirmed by evidence that p53 is mutated in almost 50% of all human malignancies, while mutations in the p63 and p73 genes are rare (Irwin and Kaelin, 2001; Nomoto et al., 1998).
On the one hand, p63/73 contribute to p53 mediated tumor suppression, as it has been shown that mice heterozygous for mutations in both p53/63 or p53/73, lead to a more aggressive tumor phenotype (Flores et al., 2005) and that p53 requires at least one of its homologues to function properly (Flores et al., 2002). On the other hand, p63/73 might have a specific role in tumor suppression as their loss or knockdown leads to tumorigenesis in specific tissues (Ahomadegbe et al., 2000; Park et al., 2000; Park et al., 2004; Puig et al., 2003; Urist et al., 2002) with a spectrum differing from that of p53 and reflects their pattern of mostly epithelial tissue expression.
Only recently, the human TP53 gene has been established to have a structure similar to p63 and p73 genes (Bourdon et al., 2005). p53 gene transcription can be initiated from two distinct sites that have the potential to encode at least nine different p53 isoforms named according to p63/p73 nomenclature (Ghosh et al., 2004). In addition, p53 splice variants are expressed in a tissue-dependent manner, indicating that the internal promoter and alternative splicing of p53 can be regulated (Bourdon et al., 2005). Potentially, deficiency in regulation of these isoforms may play a role in tumor formation/progression and, in particular, the expression or loss of certain p53 isoforms could impair function in cells not harboring inactivating mutations of the parental gene. At the moment, it is unclear what the biological function of the individual p53 isoforms may be, but it seems likely that interplay between isoforms on specific targets may play a role in controlling p53 activity under normal and transformed conditions.
4. Structure of the p53 protein
The p53 protein, like many other transcription factors, has a modular structure characterized by the presence of evolutionarily conserved functional domains that is common to the other p53 family members p63 and p73 (Kaghad et al., 1997; Yang et al., 1998). In p53, only the DNA binding and tetramerization domains are highly structured (Cho et al., 1994; Clore et al., 1995), which suggests the ability of the domains of the N-terminus and C-terminal regulatory domain to allow for binding of numerous transcriptional effecters.
4.1. Transactivation Domain.
Early experiments already demonstrated that the N-terminal acidic domain of p53 is responsible for transcriptional activity (Fields and Jang, 1990). This portion contains two transactivation domains (TAD1 aa 1-42 and TAD2 aa 43-62) that are independently sufficient to activate transcription when fused to a heterologous DNA-binding domain (DBD) (Unger et al., 1992) and that interact with components of the basal transcriptional machinery as the TATA-binding protein (TBP) and TBP associated proteins (TAFs) as well as with p300/CBP (Liu et al., 1993). These two domains are both necessary for p53 full transcriptional activity, but with different functions. The TAD1 is in fact required for inducing p21 transcription and G1 arrest while the TAD2, together with the proline rich domain (PRD), is essential for the apoptotic response (Sabbatini et al., 1995; Zhu et al., 2000). The transcriptional activation domain is finely regulated both by post-translational modifications and by interaction with protein partners such as p300/CBP and Mdm2. Mdm2, the most critical negative regulator of p53, interacts with residues 17-27 of TAD1 and mediates the ubiquitylation of p53 C-terminal lysine residues. Moreover, Mdm2 together with MdmX prevents the binding of transcriptional co-activators such as p300 and interferes with the transactivation of TAD1-dependent target genes.
4.2. Proline Rich Domain.
C-terminal to the transactivation domain, exists a proline-rich domain (PRD, aa 61-94) containing five repeats of the amino acid motif PXXP (Walker and Levine, 1996). This region may be significant for p53 regulation as it could serve as a binding region for Src homology 3 (SH3) domain-containing proteins and can modulate signal transduction (Kay et al., 2000). Additionally, the p53 PXXP motifs may contribute to interactions with the transcription co-activator p300 (Dornan et al., 2003a); thus influencing p53 acetylation. Potentially, acetylation could reduce Mdm2 binding and influence p53 stability (Wulf et al., 2002; Zacchi et al., 2002); which is consistent with other evidence indicating that the PRD modulates Mdm2 binding (Berger et al., 2001; Dumaz et al., 2001). Although the PRD may play some role in regulation of p53, p53?P mice (p53 sans PRD) display increased Mdm2-mediated degradation and decreased transactivation capacity (Toledo et al., 2007). Therefore, PXXP motifs are essential for p53 function, as their deletion leads to increased degradation and decreased p53 mediated cell cycle control.
4.3. DNA-binding Domain.
The sequence-specific DNA binding domain (DBD) comprises the central third of p53 (aa 100-300) required for p53 to function as a transcriptional activator (el-Deiry et al., 1992). The canonical p53-response element (p53RE) contains two decamers of PuPuPuC (A/T)(A/T) GPyPyPy in which a p53 monomer binds the pentameric half sequence and the p53 tetramer the full consensus site (Cho et al., 1994; Ma et al., 2005). When presented in linear conformation, these sites are poorly recognized by p53, but while in stem-loop conformation, are bound with high affinity. It has also been shown that p53 specifically binds to CTG_CAG trinucleotide repeats that undergo topological alterations (Walter et al., 2005) or to cruciform-forming sequences (Jett et al., 2000) with no resemblance to the p53RE consensus. The crystal structure of the p53-DBD bound to DNA has revealed that the conserved regions are crucial for the p53-DNA interaction (Cho et al., 1994).
The importance of sequence-specific DNA binding for p53 is highlighted by the fact that 97% of tumor-associated mutations cluster in this domain (Sigal and Rotter, 2000). Mutants at some positions (R273H) lose direct contacts with DNA and are therefore "contact mutants" (Bullock et al., 1997; Cho et al., 1994). While other mutants, termed "conformational mutants" (R175H, G245S, R249S, R282W) show reduced binding due to conformational destabilization (Bullock et al., 2000; Wong et al., 1999). Still other mutations (R248W) result in breaking DNA-protein contacts as well as introducing extensive structural changes (Wong et al., 1999). Any of which could result in disruption of p53 tetramerization. However, the DBD requires flexibility within the tetramer so as to form contacts with the nucleotides of a consensus site which also appears to facilitate the function of some p53 mutants to bind and inactivate p63 and p73, as well as to force wt p53 into a mutant conformation (Bensaad et al., 2003; Di Como et al., 1999; Gaiddon et al., 2001). Partially active tetramers may lead to new phenotypes via generation of abnormal binding surface(s), which may change the panel of p53 transcriptional effecters. At least three DNA-binding defective molecules of p53 are needed in order to effectively inactivate the tetramer (Chan et al., 2004; Shieh et al., 1999). Alternatively, mutations in the DBD may change the relative affinity for target genes resulting in loss of binding or recognition of novel sites with consequent alteration in the transcriptional activity of p53.
4.4. Oligomerization Domain.
The tetramerization required for high-affinity DNA binding and transcriptional activation is mediated through this domain (aa 326-356). Tetramerization appears to be essential for biological activity as p53 is unable to bind DNA in vitro if the core that stabilizes the dimerization is disrupted. The TD is highly conserved in the p53 family, but each shows distinct properties preventing heterotetramerization. This is significant for the function of these proteins since p53, p63, and p73 are likely to be expressed simultaneously, at least in some tissues (Chene, 2001). Interestingly, mutant forms of p53 are able to heterotetramerize with p73 thus inhibiting its apoptotic functions (Strano et al., 2000). Moreover, the heterotetramerization of wild-type and mutant p53 is likely causative of the dominant-negative activities of mutant p53. It will be interesting to learn how heterotetramerization of full-length p53 and other isoforms could affect activity and if this activity is different from that of a homotetramer.
4.5. C-terminal Regulatory Domain.
The basic C-terminal regulatory domain (CT, aa 364-393) has the ability to influence p53 activity due to the number of post-translational modifications that occur here upon receiving a stress response signal. This domain is intrinsically disordered and is able to interact directly with DNA and RNA (Ayed et al., 2001; Lee et al., 1995). This conformation allows it to interact with multiple partners. For example, the CT is able to bind ssDNA ends, mismatches, recombination intermediates and g-irradiated DNA in vitro; thus implying an ability to recognize damaged DNA and DNA repair intermediates in vivo (Bakalkin et al., 1995; Lee et al., 1995; Zotchev et al., 2000). Additionally, it has been demonstrated that the CT is important for binding non-linear DNA (Fojta et al., 2004; McKinney and Prives, 2002; Palecek et al., 2004) and is involved in p53 linear diffusion on DNA (McKinney et al., 2004). These studies also revealed that the CT is required for efficient promoter activation by p53.
4.6. Nuclear localization and export signals.
p53 is known to shuttle between the nucleus and the cytoplasm. Since tetrameric p53 is too large to passively diffuse across the nuclear membrane, necessarily its nucleo-cytoplasmic shuttling is facilitated by nuclear import and export signals (Middeler et al., 1997). The C-terminus of p53 contains a cluster of three nuclear localization signals (NLS) that mediate migration into the nucleus.
NLSI (aa 316-322), the most active domain, is highly conserved in genetically diverged species and shares perfect homology with consensus NLS sequences found in other nuclear proteins (Shaulsky et al., 1990). This NLS has a bipartite structure (Liang and Clarke, 1999a) consisting of a canonical (aa 315-322) and basic sequence (aa 305-306) separated by a spacer (Liang and Clarke, 1999b). The basic residues at the N- and C-termini of the NLS (K305, R306, and K319/320/321) are necessary and sufficient for the complete nuclear localization of a cytoplasmic reporter protein (Liang and Clarke, 1999a). NLS II and III (aa 370-384) appear to be less effective and conserved (Shaulsky et al., 1990).
Additionally, p53 has two putative nuclear export signals NES, in the N-terminus and oligomerization domain (aa 11-27 and 340-351 respectively). As the NLS is adjacent to and the NES contained within the TD, it has been suggested that oligomerization of p53 may determine nuclear export by affecting the accessibility of the NLS and/or NES to their respective receptors (Liang and Clarke, 2001; Stommel et al., 1999).
B. The cellular surface glycoprotein MUC1
MUC1 is a large, type I transmembrane protein normally expressed on the apical surface of ductal epithelia (Hollingsworth & Swanson, 2004) and is usually aberrantly expressed in tumor cells. During translation, MUC1 undergoes auto-proteolytic cleavage resulting in a large extra cellular extensively glycosylated and smaller transmembrane subunit that remain non-covalently associated at the cell surface (Levitin et al., 2005). The transmembrane subunit consists of a 58-amino acid extracellular domain, 28-amino acid transmembrane domain, and a 72-amino acid cytoplasmic tail (Merlo et al., 1989) that contains 21 potential phosphorylation sites. With 212 potential phosphorylation states, differential phosphorylation of the MUC1 cytoplasmic tail (MUC1CT) suggests the ability to function as a molecular rheostat modulating function of a wide array of signaling pathway factors through associations based solely upon phosphorylation status.
Although expressed at the cell surface, MUC1CT is able to internalize and translocate to the nucleus. Characteristics of which remain largely unanswered as the precise mechanism(s) is unknown in addition to the absence of a traditional NLS within this domain. However, the presence of an RRK motif, while not fitting the canonical monopartite NLS sequence either, would be the most likely way in which translocation is mediated through direct binding of importin (Kau et al., 2004). However, nuclear localization has been observed through binding of MUC1CT with Nup62 while dimerized through the preceding CQC motif (Leng et al., 2007). While this is one way in which MUC1CT may be translocated to the nucleus, it is likely not unique in that MUC1CT may come in contact with and modulate the activity of its transcriptional machinery targets.
The restoration of tumor suppressor pathways, either transcriptionally or translationally, is appealing in the treatment of cancer. Several studies have associated the MUC1CT with signaling events that influence processes ranging from apoptosis to invasion and metastasis (Gao et al., 2009). For example, the MUC1CT has previously been detected in association with the tumor suppressor p53 following phosphorylation by the Met receptor tyrosine kinase (Singh, 2008). Although the role MUC1 plays in modification of tumor suppressor activity is not well understood and only now being revealed, the 212 potential phosphorylation states cannot be ignored in MUC1CT ability to bind and modulate the activities of important factors in numerous pathways.
5.1 MUC1 phosphorylation
While there is great interest in identification of the downstream effects of MUC1CT phosphorylation, the kinases involved are only now becoming known. However, studying the effects of phosphorylation at different sites is not always straight-forward as regulation of the in vivo phosphorylation of MUC1CT is not well characterized. Even so, some of the kinases acting upon MUC1CT are known, such as PDGF-R, GSK-3, c-Src, and EGF-R to name a few (Li et al. 1998, 2001, 2001 and Singh et al. 2007 respectively). Additionally, there is evidence that the Met receptor tyrosine kinase interacts with MUC1 and catalyzes phosphorylation of the YHPM motif in response to cellular stimulation with its ligand hepatocyte growth factor (HGF). This stimulation increases the relative amount of MUC1CT found interacting with p53, enables p53-mediated suppression of AP1 transcriptional activity and decreases MMP1 expression (Singh et al. 2008).
Phosphorylated proteins, at physiological pH ~7.4, carry a negative charge on the deprotonated phosphate group. Potentially this negatively charged modification could participate in a protein-protein interaction with the lysine and arginine residues present driving association through favorable charge-charge interactions.
Just outside of the p53-binding region in MUC1CT, FGF1 stimulation induces c-Src-dependent phosphorylation of the YEKV motif, inhibiting GSK-3 action, and enhances the binding of -catenin to MUC1 (Li et al., 2001; Ren et al., 2001 and 2006). MUC1CT has been identified as a phosphorylation target of GSK3 (Li et al., 1998), EGFR (Li and Ren et al., 2001), PDGFR (Singh et al., 2007), FGFR/c-Src (Li and Kuwahara et al., 2001; Ren et al., 2006), and c-Met (Singh, 2008). It is currently known that GSK-3 phosphorylates a threonine 16 amino acids C-terminal to the CQC motif, PDGFRb, c-Met, FGFR, and TrkA phosphorylate the adjacent tyrosine, IR and c-Abl the tyrosine 26 amino acids C-terminal and PDGFRb tyrosine 32 in the MUC1CT binding region of p53 (Singh, 2007) (Figure1). C-Src phosphorylates the tyrosine located in the YEKV motif just outside of the p53-binding site, inhibiting GSK-3 action, and enhances the binding of -catenin to MUC1 (Li et al, 2001). This is of significant interest as each phosphorylation site is located within the binding region of p53 and participates in MUC1 translocation to the nucleus.
AimsIn this study, we have observed the inability of an unphosphorylated MUC1CT to interact with p53. Phosphorylation on YHPM (Y14) of MUC1CT previously was identified to increase association with p53 (Singh, 2008). Association of MUC1CT, bearing a negatively charged phosphate, with p53 may well stimulate electrostatic effects with positively charged amino acid side chains by neutralizing some of the phosphate negative charges. On the other hand, the neutralization of phosphate charges by cations and water in solution may result as well. Still, the molecular forces involved remain unclear.
1. To identify the oxidation state of the Cys residues of the CQC motif in MUC1 CT.
2. To compare the necessity of p53 association with MUC1CT for nuclear translocation in 76R30 breast cancer cells, S2-013 and Panc1 cells; with decreased and wild type levels of p53 expression respectively.
3. To determine if the CQC motif participates in a possible cleavage event of the MUC1CT when bound to p53, -catenin, and a host of other proteins allowing translocation to the nucleus.
The current hypothesis implicates reactive oxygen species (ROS) as significant regulators for the nuclear localization of MUC1CT bound to p53. Under hyperoxic conditions, the Cysteine thiol side chains of the CQC motif would become thionylated and potentially facilitate the cleavage of MUC1CT from the transmembrane region such that it may be translocated to the nucleus. with -catenin, p53 and a host of other factors in complex from the membrane to its intended destination; characterized translocation thus far is to the nucleus. ROS may also affect the phosphorylation status in the p53 binding region allowing for dissociation and alternative mitochondrial translocation.
It is currently known that GSK-3 phosphorylates a threonine 16 amino acids C-terminal to the CQC motif, PDGFRb, c-Met, FGFR, and TrkA phosphorylate the adjacent tyrosine, IR and c-Abl the tyrosine 26 amino acids C-terminal and PDGFRb tyrosine 32 in the MUC1CT binding region of p53 (Singh, 2007) (Figure1). C-Src phosphorylates the tyrosine located in the YEKV motif just outside of the p53-binding site, inhibiting GSK-3 action, and enhances the binding of -catenin to MUC1 (Li et al, 2001) (Figure1). This is of significant interest as each phosphorylation site is located within the binding region of p53 and participates in MUC1 translocation to the nucleus. If oxidative stressors disrupt phosphorylation, -catenin and p53 association and consequent nuclear translocation could be disrupted; thus providing a mechanism by which MUC1 is internalized to the mitochondria to prevent apoptosis rather than to the nucleus for transcriptional activation.
The site of interaction of MUC1CT in the C-terminus of p53 has been previously reported due to an absence of binding when deleted (14) and confirmed here through positive interaction of MUC1CT with only the C-terminus and full-length p53. Characterization of binding, studied through isothermal titration calorimetry, surface plasmon resonance, and nuclear magnetic resonance, show that MUC1CT does not bind only monomeric p53 but rather dimerized and tetramerized p53. We provide a comprehensive characterization of the conditions that contribute to transcriptional regulation of p53 target genes through MUC1CT association. This information will assist in the development of potential treatments for inhibition of MUC1CT binding p53 and the concomitant alteration of invasion and metastatic potential of tumor cells.
MUC1 binding of p53 is associated with tumorigenesis and is suggested to act as a negative regulator of the p53 pathway. MUC1CT and p53 association may have numerous effects on cellular processes including growth, motility and invasion of cancer cells; thus making it an important interaction in cancer progression.
To date, a phosphorylation event at YHPM of MUC1CT has been implicated in binding with p53 (Singh, 2008). For these reasons, phosphorylation of MUC1CT, which precedes p53 binding and alters the invasive potential of 76r-30 cells, can be thought of as a gain of function association. The binding site of MUC1CT in the C-terminus of p53 has been reported due to an absence of binding when deleted (Wei et al, 2005). Here we confirm a positive interaction of MUC1CT with the C-terminus and full-length p53 through GST pull down of endogenously expressed MUC1CT in tumor lines 76r-30 and S2-013 (Figure 1). Additional binding studies, through isothermal titration calorimetry and surface plasmon resonance, confirmed that MUC1CT does not bind p53 unless phosphorylated and that binding is phosphorylation specific.
Although there is an increase in the relative amount of MUC1CT found in association with p53 upon stimulation with HGF, we can assume that phosphorylation of MUC1CT on YHPM and/or YEKV exists at baseline levels and that stimulation only increases the amount of MUC1CT able to bind p53. This is attributable to our ability to co-immunoprecipitate p53 bound to MUC1CT in the absence of stimulation as well, but at a reduced amount.
MUC1CT phosphorylation is necessary for binding p53. Mono-phosphorylation of YHPM and YEKV were both able to interact with the C-terminus of p53. YEKV phosphorylation also enables some interaction with the N-terminus of p53. However, this association is only minor as compared with C-terminal and full-length p53, but unlike phosphorylated YHPM, there is a greater amount of binding to full length as compared with C-terminal p53 alone. The increase in response units bound with phosphorylated YEKV for full length as compared with C-terminal p53 can be attributed to the fact that there is minimal binding to the N- as well as C-terminus. Given that there isn't interaction with N-terminal p53 for phosphorylated YHPM the maximal number of response units is roughly equivalent in C-terminal and full length p53 (figure 4). When considering that both phosphorylated YHPM and YEKV MUC1CT interact with p53, it could be suggested that phosphorylation of any of the 21 available phosphorylation sites may participate in p53 binding. However, any phosphorylation event of MUC1CT is not a factor in binding p53. For example, phosphorylation of Y29 (YVPP) was unable to interact with any of the recombinant p53 constructs in a similar manner to unphosphorylated MUC1CT.
Therefore, even from the sites of MUC1CT phosphorylation that are able to stimulate interaction with p53, there is an obvious preference for certain sites when compared with others in strength of association. Differential strength in binding may be an evolved mechanism through which effects as a transcriptional co-repressor/activator may be modulated. There may be other sites or combinations of phosphorylation that would contribute to this apparent molecular rheostat mode of function for MUC1CT on p53. Naturally, if two different sites of phosphorylation are able to facilitate p53 interaction, YHPM and YEKV, it could be assumed that a diphosphorylated peptide would interact similarly, but with a greater rate and/or affinity than a monophosphorylated MUC1CT.
The region of MUC1CT bound by p53 is known, but the region of p53 that is binding MUC1CT is not as clear. HGF facilitated MUC1CT interaction differentially modulates the invasive potential of cells harboring distinct mutations of p53. The implicated phosphorylation site involved in this protein-protein interaction has been mapped to the YHPM motif of MUC1CT (Singh, 2008). Binding does not occur in the absence of MUC1CT phosphorylation. Therefore, we propose that MUC1 plays a role as a gain of function modifier for mutant p53 in reference to metastasis.
MUC1CT association with p53
In this study, we have observed the inability of an unphosphorylated MUC1CT to interact with p53. Phosphorylation on YHPM (Y14) of MUC1CT previously was identified to increase association with p53 (Singh, 2008). Association of MUC1CT, bearing a negatively charged phosphate, with p53 may well stimulate electrostatic effects with positively charged amino acid side chains by neutralizing some of the phosphate negative charges. On the other hand, the neutralization of phosphate charges by cations and water in solution may result as well. Still, the molecular forces involved remain unclear.
Over expression of MUC1 is associated with a multitude of adenocarcinomas, particularly the majority of invasive ductal carcinomas. The cytoplasmic tail of MUC1 (MUC1CT) is differentially phosphorylated allowing for participation in many intracellular signaling mechanisms. We report here that MUC1CT is able to bind the p53 tumor suppressor, specifically interacting with a C-terminal portion comprising the tetramerization and regulatory domains. This interaction was further characterized through determination of binding kinetics using a combination of spectroscopic and calorimetric techniques. We show that unphosphorylated MUC1CT does not bind p53, binding is phosphorylation specific, and stabilizes the disordered p53 C-terminal regulatory domain. These findings are consistent with MUC1CT acting as a co-regulator of p53 activity.
MUC1 is a cell surface glycoprotein expressed primarily on the apical surface of ductal epithelia that is usually aberrantly expressed in tumor cells. MUC1 undergoes co-translational cleavage into a large extracellular extensively glycosylated and a smaller transmembrane subunit that remain non-covalently associated at the cell surface (Levitin et al., 2005). The transmembrane subunit includes a 58-amino acid extracellular domain, 28-amino acid transmembrane domain, and a 72-amino acid cytoplasmic tail (Merlo et al., 1989) containing 21 potential phosphorylation sites. With 212 potential phosphorylation states, differential phosphorylation of MUC1CT suggests the ability to function as a molecular rheostat in the co-regulation of many macromolecules.
In this study, we have observed the inability of an unphosphorylated MUC1CT to interact with p53. Phosphorylation on YHPM (Y14) of MUC1CT previously was identified to increase association with p53 (Singh, 2008). Association of MUC1CT, bearing a negatively charged phosphate, with p53 may well stimulate electrostatic effects with positively charged amino acid side chains by neutralizing some of the phosphate negative charges. On the other hand, the neutralization of phosphate charges by cations and water in solution may result as well. Still, the molecular forces involved remain unclear.
Our current interest is in elucidating the kinetic and thermodynamic properties resulting from phosphorylated MUC1CT association with p53. In this work, we used a variety of physical techniques to determine the binding properties of MUC1CT with p53. Complete thermodynamic profiles indicate that the favorable formation of 3 out of 4 complexes result from the characteristic compensation of a favorable enthalpy with an unfavorable entropy term. Whereas the other interaction yields both unfavorable enthalpy and entropy terms that are likely due to p53 aggregation in the final complex formation. This is compensated with a larger favorable enthalpy in the final interaction site and therefore suggests driving of p53 aggregation by phosphorylated MUC1CT. Phosphorylated MUC1CT interaction with the C-terminal regulatory domain of p53 may affect pivotal roles played by p53 in growth arrest, repair, apoptosis, senescence, and differentiation (Sphyris and Harrison, 2005) as the ability of p53 to bind target promoters is likely modulated.
- Kau, T. R., Way, J. C., and Silver, P. A. (2004) Nat. Rev. Cancer 4, 106-117
- Gao, J., McConnell, M. J., Yu, B., Li, J., Balko, J. M., Black, E. P., Johnson, J. O., Lloyd, M. C., Altiok, S., Haura, E. B. (2009). MUC1 is a downstream target of STAT3 and regulates lung cancer cell survival and invasion. Int. J. Oncol. 35(2), 337-45
- Raina, D., Ahmad, R., Joshi, M. D., Yin, L, Wu Z, Kawano T, Vasir B, Avigan D, Kharbanda S, Kufe D. (2009) Direct targeting of the mucin 1 oncoprotein blocks survival and tumorigenicity of human breast carcinoma cells. Cancer Res. 69(12), 5133-41
AIM OF THE THESIS
The p53 tumor suppressor often plays an integral role in preventing cancer development through regulation of its many target genes; underscored by the fact that it is compromised to some extent in almost all human tumors. Significant advances have been made in understanding p53 functions, yet a greater insight is needed in the specificity of p53 response. The aim of this thesis is in elucidating the kinetic and thermodynamic properties resulting from phosphorylated MUC1CT association with p53 and what effects this association may have. The first part of this study will focus on the in vivo association and effects upon mutant p53 variants from binding MUC1CT. In the second part, the functional interaction between p53 and its interaction partner MUC1CT will be characterized, as the presence of phosphorylation and evidences from the literature make it a promising candidate for modulating the p53 pathway at the transcriptional level. These studies are conducted in light of achieving greater knowledge on the regulation of the p53 pathway by MUC1CT that might be of relevance for understanding the tumorigenesis process and could be exploited in the design of new targeted therapies.
In this work, we used a variety of biophysical techniques to determine the binding properties of MUC1CT with p53. Complete thermodynamic profiles indicate that the favorable formation of 3 out of 4 complexes result from the characteristic compensation of a favorable enthalpy with an unfavorable entropy term. Whereas the other interaction yields both unfavorable enthalpy and entropy terms that is likely due to p53 aggregation in the final complex formation. This is compensated with a larger favorable enthalpy in the final interaction site and therefore suggests driving of p53 aggregation by phosphorylated MUC1CT. Phosphorylated MUC1CT interaction with the C-terminal regulatory domain of p53 may affect pivotal roles of growth arrest, repair, apoptosis, senescence, and differentiation (Sphyris and Harrison, 2005) as the ability of p53 to bind DNA is likely modulated.
Cells and mediaWell-differentiated SUIT-2 sub clone S2-013 (Taniguchi et al., 1992) MUC1 over-expressing MIF16 and Neo cells were grown in RPMI supplemented with 7% fetal bovine serum. 76r-30 cells were grown in DFCI-1 medium as described previously (Band et al., 1991 and Boyer et al., 1996).
MUC1CT Peptides.Phosphorylated peptides N'-NYGQLDIFPARDTYHPMSEYPTYHTHGRYVPPSSTDRSPYEKVSAGNGGSSLSYTNPAVAAASANL-C' with masses of 7150 Daltons for phosphorylated Y14 and Y29 and 7136 Daltons for unphosphorylated (GeneMed Synthesis, Inc.) were used in in vitro binding studies. Peptide sequence was confirmed by amino acid analysis and MALDI-TOF mass spectrometry.
Recombinant p53 expression and purification. Transformed DH5a E. coli cells containing p53: pGEX 4t-1 expression constructs were incubated with 100 ug/mL ampicillin at 37oC until OD660 reached 0.8. The culture was cooled and induced with 1mM IPTG at 16oC for 16 additional hours. Cells were harvested by centrifugation at 4 oC, 3000xG for 20 minutes, washed with ice-cold 1X PBS and lysed in 10 mL ice cold STE buffer (10 mM Tris-HCl pH 8.0, 1 mM EDTA, 150 mM NaCl) with 100 ul of freshly prepared lysozyme solution (10 mg/mL) and 200 ul of Pierce Inclusion Body Buffer. Cells were sonicated and 100 ul of 1 M DTT and 1.4 mL of 10% Sarkosyl in STE buffer added. Cell debris was removed by centrifugation at 16000xG for 20 minutes, discarded, and the volume brought to 20 mL with STE Buffer.
Lysate was added to a 1 mL bed of 50% Glutathione Sepharose slurry in 1XPBS and nutated at RT for 1 hour. Beads were pelleted at 100 x G and washed with ice-cold 1XPBS. Recombinant protein was eluted with 10 x 1 mL fractions of Elution Buffer (50 mM Tris-HCl pH 9.0, 20 mM GSH) and dialyzed into 150mM NaCl, 30mM Tris HCl, pH 7.9. Dialyzed protein was applied to a Superdex 200 size exclusion HPLC column at flow rate of 0.5mL/minute and 0.5 ml fractions were collected into 150mM NaCl, 30mM Tris HCl, pH 7.9.
GST Pull-Down Assay.Recombinant GST-p53 expressed in BL21 Escherichia coli cells (Amersham Biosciences) containing pGEX 4t-1:p53fusion plasmids (G. E. Healthcare) kindly provided by the laboratory of Vimla Band. p53 constructs comprising amino acids 1-101, 292-393, and 1-393 were used for binding 76r-30 lysate proteins (Band and Sager, 1989). 280 ug each of 76r-30, MP1P, and MP3P cellular lysate was incubated in a 1:1 ratio with each of N-terminal (1-101), C-terminal (292-393), and full length (1-393) GST-p53. Incubation was conducted at 4 degrees C for 12 hours and binding determined by western blot. After washing thoroughly, protein complexes were eluted in 2x Laemmli sample buffer, electrophoresed by SDS-PAGE and transferred to a polyvinylidene fluoride (PVDF) membrane. Detection MUC1CT was performed by standard Western blot procedures using CT2 (1:100 diluted in blocking solution).
Surface Plasmon Resonance.Goat anti-GST antibody was diluted to 30 ug/mL in immobilization buffer (10 mM sodium acetate, pH 5.0) and amine coupled according to standard procedures to all 4 flow cells of a CM5 sensor chip. Bacterial expressed human GST-p53 (amino acid constructs 1-101, 292-393, and 1-393) were diluted to 10 ug/ml with HBS-EP buffer (10 mM HEPES, 150 mM NaCl, 3 mM EDTA, and 0.005% surfactant P-20, pH 7.5) and injected into the flow cells in the order of recombinant GST in flow cell 1, 1-101 in flow cell 2, 292-393 in flow cell 3, and 1-393 in flow cell 4. For affinity measurements, binding and dissociation were monitored with a Biacore 3000 (GE Healthcare). The sensorgrams were corrected by subtracting the signal of the reference cell 1 and the data fitted using the BIAevaluation software version 4.1 (GE Healthcare). Binding data were globally fit to the 1:1 Langmuir binding model that assumes binding homogeneity. Values for the statistical closeness of fit, ?2, were always above 10, indicating that the simple 1:1 model of interaction doesn't correctly describe the experimental data.
Isothermal titration calorimetry.Titration experiments were performed at 25degrees C in Tris buffer (30 mM Tris-HCl, 150 mM sodium chloride, and pH 7.8) using an ITC Microcal Omega titration calorimeter (MicroCal LLC, Northampton, MA). The measuring cell contained an effective volume of 1.4093 mL of protein solution, the reference cell was filled with deionized water, and the instrument calibrated by a known standard electrical pulse. Before starting experiments, the baseline stability was verified and a spacing of 600 s between injections was applied in order to allow the system to reach thermal equilibrium after each injection. A control experiment was carried out for each titration by adding the titrant into the buffer alone, under identical conditions; heats of dilution were negligible.
In a set of experiments aimed at determining the binding properties of MUC1CT (800 um) with p53 (30 um), a 100-uL syringe filled with a solution of the titrant (800 um) was used to titrate 2 mL of p53 solution (30um). p53 was titrated with an initial 1 uL injection followed by 29 X 5 uL injections of 800 um MUC1CT. Complete mixing is achieved by continuous stirring of the syringe paddle at 400 rpm. The area of each injection peak is proportional to the heat of interaction, Q, that when corrected for titrant dilution heat and normalized by concentration of bound titrant, is equal to the binding enthalpy (?Hb) at a particular degree of binding; the precision of each injection heat is about 0.5 ucal.
The KD value was calculated using the approximation KD = 1/ KA. The three parameters ?Hb (average binding enthalpy), Kb (macroscopic equilibrium constant) and "n" (non-interacting binding sites) were determined by interactively fitting the calorimetric binding isotherm using the Origin software (version 5.0) (Microcal Software, Inc.). Alternatively, the resulting n values from the complex stoichiometry obtained from Job plots were also used to fit the other two parameters. All procedures yielded similar results.
Binding constants and thermodynamic parameters provided do not take into account possible events of proton transfer linked to metal binding, or the presence in solution of complexes between metal ions and the buffer. This treatment is beyond the scope of the present study.
Circular Dichroism spectroscopy. Far-UV CD spectra were recorded in a stopped flow circular dichroism spectrometer model 202SF (Aviv Instruments, Inc.) equipped with a Peltier temperature control unit, flushed with N2. Protein samples were prepared at optical density 1 (O.D.) in Tris buffer (30 mM Tris-HCl, 150 mM sodium chloride, and pH 7.8). The ellipticity was recorded in 1 nm increments using free-strained quartz cuvettes with path lengths of 1.0 cm between 200 and 400 nm to minimize signal noise and the reported spectra correspond to the average of at least three scans in order to achieve an appropriate signal-to-noise ratio. The spectra were corrected for buffer contributions.
Stoichiometries (Job plot).MUC1CT:p53 stoichiometry was checked according to previously published protocols (Job, 1928; Soto et al, 2002). Absorbances at 220 and 280 nm of several solutions with similar concentrations (~30 um) were measured as a function of the mole fraction of MUC1CT.
p53 aggregation.UV absorbance of recombinant p53 in Tris buffer (30 mM Tris-HCl, 150 mM sodium chloride, pH 7.8) was checked using a Perkin-Elmer Lambda 10 spectrophotometer with concentrations of 1 mM, 25 mM, 50 mM, 100 mM, 200 mM, 400 mM, and 800 mM was measured between 200 and 400 nm. Absorbance at 220 nm was plotted as a function of mole fraction of p53 to determine concentration dependent aggregation.
Kinetic analysis of MUC1CT association with p53
The MUC1CT is a 66 amino acid peptide that plays a role in the modification of p53 transcriptional activity. Co-immunoprecipitation, with CT2 (Croce et al., 2003), of MUC1CT with p53 is increased upon stimulation of MUC1 expressing cells with hepatocyte growth factor (HGF) (Singh et al., 2008). The p53 region bound by MUC1CT has been reported to be in the C-terminal regulatory domain (Wei et al., 2005).
Phosphorylated MUC1CT binding region was confirmed through GST pull down. 1:1 incubations of GST tagged full length, N-terminal (Transactivation and Apoptotic activation domains) or C-terminal p53 (Tetramerization and Regulatory domains), with HGF stimulated lysates from p53 null (76r-30), over expressing ?N239 p53 (MP1P) or N247I (MP3P) (Wazer et al., 1994) were analyzed by western blot to determine which recombinant construct(s) bound endogenously expressed MUC1CT. Binding of MUC1CT with C-terminal p53 containing amino acids 292-393 and full-length p53 were clearly observed, but not with the N-terminus (Figure 1C).
Phosphorylation of the YHPM motif in MUC1CT has been implicated and previously reported in binding p53 (Singh et al., 2008), but other phosphorylation sites located within the p53-binding domain may also facilitate interaction. Phosphorylated YHPM (Y14), YVPP (Y29) and unphosphorylated MUC1CT were assayed by surface plasmon resonance (SPR) to determine whether binding was phosphorylation specific (Figure 1A,B,D).
Unphosphorylated and phosphorylated Y29 MUC1CT were unable to interact with p53, but phosphorylated Y14 MUC1CT bound and characteristics of binding are summarized (Table 1). Interaction occurred with the C-terminal and full length p53 peptides when phosphorylated at Y14, but not the N-terminus. Therefore, phosphorylation of MUC1CT is clearly specific and required for binding the C-terminus of p53.
Thermodynamic analysis of MUC1CT association with p53
Thermodynamic comparison between phosphorylated Y14 and unphosphorylated MUC1CT association with p53 was investigated using isothermal titration calorimetry (ITC). ITC measurements obtained through successive titration of phosphorylated Y14 MUC1CT into a Tris buffered solution of C-terminal p53 yielded exothermic peaks from each injection, but negligible dilution heats into buffer alone. Fitting of the isotherm, using Origin software version 5.0 produced a 4-interaction site model, as simpler models did not adequately represent the data. When applying the concentrations of titrant and solution, a stoichiometry of 1.5 was obtained, indicating a 3:2 molecular ratio of MUC1CT to p53 (Figure 2A). Saturation of p53 binding sites occurred following 18 injections of MUC1CT where a molar ratio of 1.7 was achieved and then further titrated to 2.0 to achieve the heat of dilution produced by p53 aggregation.
Dissociation rates (KD) were calculated for phosphorylated Y14 MUC1CT from measured association rates (KA) with the approximation of KD = 1/ KA (Table 2); all sites are similar in measured rate of association. However, favorable formation of 3 of 4 complexes result from the characteristic compensation of a favorable enthalpy with an unfavorable entropy term, but the remaining site had an unfavorable endothermic enthalpy and was thus entropically driven (Table 2). The enthalpy and entropy values were calculated with Gibbs' free energy (DG = DH-TDS) and standard Gibbs free energy of formation equations (DG = -RTlnK) (Weber, 1995; D'auria et al., 1997). Values for the KA and KD are comparable to those established by SPR for binding of phosphorylated Y14 MUC1CT with p53 and are consistent with a role of MUC1CT as a transcriptional effecter of p53.
Continuous variation experiments were used to determine complex stoichiometries (figure 3A). At these two wavelengths, the lines intercepted at mole fractions of 0.5974 (220 nm) and 0.4766 (280 nm), which correspond to molar ratios of 1.4838 and 0.9106 respectively. This is in agreement with the ITC obtained molar ratio of ~1.5 and confirmed a 3:2 complex stoichiometry formation of phosphorylated Y14 MUC1CT:p53.
MUC1CT induced p53 aggregation
In its active conformation, p53 is tetrameric and one domain - the tetramerization domain - permits the oligomerization (Chene, 2001). Protein aggregation was assayed through reading UV absorbance in successive variations of concentration and plotting as a function of mole fraction in solution. Absorbance readings of p53 solutions at 220 nm demonstrate that aggregates do not begin formation until concentrations of 25-50 mM and tetramerization occurs at a concentration of ~200 mM (figure 3B).
Circular Dichroism spectroscopy
The CD spectra of each complex at 25degrees C are shown for titration of p53 with phosphorylated Y14 MUC1CT. CD spectrum for this titration is in contrast to that expected if additive (Figure 4A and B respectively). In the far UV spectrum (190-250 nm) both p53 and MUC1CT contain alpha helical secondary structures that upon titration, result in an increase in the magnitude of molar ellipticity and subtle shift . In the near UV spectrum (250-350 nm), there is also a shift away from baseline around the absorbance of all aromatic amino acids (280 nm), in particular tyrosine. The shift in the molar ellipticity around the absorbance of the aromatic amino acids, suggests stabilization of secondary and tertiary structures for both participants and the comparatively larger shift around tyrosine indicates that it is important in binding of MUC1CT to p53. However, the relative increase in magnitude in the molar ellipticity around the peptide bond absorbance suggests the formation of p53 aggregates.
Co-immunoprecipitation analysis of MUC1 complexes with CT2 (Croce et al., 2003), demonstrate HGF induced phosphorylation of Y14, by Met, increases the relative amount of MUC1CT bound by p53 (Singh et al., 2008). As a phosphorylated protein would necessarily carry a negative charge at physiological pH, it is possible that in a protein-protein interaction the presence of lysine and arginine residues would drive association through favorable charge-charge interactions.
Phosphorylated proteins, at physiological pH ~7.4, carry a negative charge on the deprotonated phosphate group. Since interaction of MUC1CT with the C-terminus of p53 occurs when phosphorylated, this suggests potential ionic interaction with positively charged amino acid side chains residing in this fragment. The C-terminal construct used in binding contains the regulatory domain, 23.1% positively charge amino acid side chains at neutral pH, in addition to the 18.4% in the tetramerization domain. This does not necessarily indicate that phosphorylated MUC1CT interaction with p53 is ionic, just that phosphorylation of Y14 precedes the interaction, as heats obtained from ITC would have been endothermic instead of exothermic.
The p53-binding region of MUC1CT resides between tyrosine residues YGQL and YEKV (Hattrup and Gendler, 2008). Located within this region are six tyrosine residues that when phosphorylated could potentially facilitate interaction with p53. Although known, the specific kinases that phosphorylate each site have not yet been fully characterized. For this reason, demonstrated phosphorylation of YHPM, catalyzed by PDGFR and c-Met (Singh et al., 2007; 2008), and YVPP, also catalyzed by PDGFR (Singh et al., 2007), were chosen for binding p53 on the Biacore chip. The inability of unphosphorylated and phosphorylated YVPP to interact with p53 and the simultaneous ability of phosphorylated Y14 MUC1CT to induce direct association with p53 demonstrated binding was phosphorylation specific. Although SPR measurements were statistically a good fit having a ?2 value < 10 (0.456), this instrument assumes a 1:1 Langmuir stoichiometry that does not fully describe all of the potential complexes that may result.
ITC experiments conducted with C-terminal p53 established a 3:2 stoichiometry for binding of MUC1CT to the C-terminal fragment, with KD ca. ~10 um, for each interaction site at room temperature (25oC). In the 3:2 complex (MUC1CT:p53) the KA values are all similar, as well as enthalpy values recorded for binding of the first, second, and forth sites, but enthalpy of the third site is different having endothermic rather than an exothermic term. It is not unreasonable to predict that this is the measurement of p53 aggregation.
Assaying the concentrations that elicit p53 oligomerization, the concentration of 30 um p53 used in ITC is below the 25 mM concentration where at a minimum dimerization begins. Therefore, phosphorylated Y14 MUC1CT binding necessarily induces any aggregation at this low of a concentration in order to form the 3:2 stoichiometry as p53 is predominantly monomeric. Dimerization may be induced by MUC1CT associating in complex or may occur and then stabilized by MUC1CT binding. In either case, the observance of exothermic heats after saturation of p53 binding sites combined with the absence of dilution heats in titration of buffer alone indicates aggregation in p53.
Saturation at a molar ratio of 1.7 at room temperature (25oC) indicated that, although MUC1CT displayed rather low affinity, it induced a conformational stabilization in p53. This association is not spontaneous, as evidenced by the endergonic free energies obtained for each site (Table 2), and therefore driven by increasing concentrations of phosphorylated Y14 MUC1CT. In particular, MUC1CT is involved in the oligomerization of p53, a process that would play a role in p53 biological activity.
MUC1CT binding of p53 has been reported due to an absence of binding upon deletion of the regulatory domain (Wei et al, 2005). Although compelling, deletion may affect p53 conformation, even though only the DNA binding and tetramerization domains are highly structured (Cho et al., 1994; Clore et al., 1995), indicating the observed MUC1CT stabilizing effects are on the C-terminal regulatory domain. Stabilization of this domain, which is just C-terminal to the tetramerization domain, likely provides the mechanism through which aggregation is achieved.
In a typical macromolecular interaction, the CD spectrum for each individual participant would shift left or rightward. However, an increase in the magnitude of molar ellipticity during phosphorylated Y14 MUC1CT titration of p53 suggests aggregation of at least one of the participants. While MUC1CT has been reported to dimerized in the nucleus, this is dependent upon the CQC motif (Leng et al., 2007), which is not present in the synthetic peptides used in all of these binding studies. Of significance, the large increase in the magnitude of molar ellipticity, upon addition of phosphorylated Y14 MUC1CT indicates aggregation of p53. Together with the endothermic measurements for the third of four binding sites and the measurement of binding heats even at saturation, these results indicate that the presence of phosphorylated Y14 MUC1CT is driving formation of higher order p53 complexes.
- Levitin, F., Stern, O., Weiss, M., Gil-Henn, C., Ziv, R., Prokocimer, Z.,Smorodinsky, N. I., Rubinstein, D. B., and Wreschner, D. H. (2005) J. Biol. Chem. 280.39, 33374-33386
- Merlo, G., Siddiqui, J., Cropp, C., Liscia, D. S., Lidereau, R., Callahan, R., and Kufe, D. W. (1989) Cancer Res 49, 6966-71
- Singh, P. K., Behrens, M. E., Eggers, J. P., Cerny, R. L., Bailey, J. M., Shanmugam, K., Gendler, S. J., Bennett, E. P., and Hollingsworth, M. A. (2008) J. Biol. Chem. 283.40, 26985-26995
- Sphyris, N., and Harrison, D. J. (2005) Oncogene. 24, 2184-2194
- Taniguchi, S., Iwamura, T., and Katsuki, T. (1992) Clin. Exp. Metastasis. 10, 259-266
- Band, V., De Caprio, J. A., Delmolino, L., Kulesa, V., and Sager, R. (1991) J. Virol. 65, 6671-6676
- Boyer, S. N., Wazer, D. E., and Band, V. (1996) Cancer Res. 56, 4620-4624
- Band, V., and Sager, R. (1989) Proc. Natl. Acad. Sci. USA 86, 1249-1253
- Job, P. (1928) Ann. Chim. (Paris). 9, 113-203
- Soto, A. M., Kankia, B. I., Dande, P., Gold, B., and Marky, L. A. (2002) Nucleic Acids Res. 30(14), 3171-80
- Croce, M. V., Isla-Larrain, M. T., Rua, C. E., Rabassa, M. E., Gendler, S. J., and Segal-Eiras, A. (2003) J. Hist. & Cytochem. 51(6), 781-788
- Wei, X., Xu, H., and Kufe, D. (2005) Cancer Cell. Feb; 7(2), 167-78
- Wazer, D. E., Chu, Q., Liu, X. L., Gao Q., Safaii, H., and Band, V. (1994) Mol Cell Biol. 14, 2468-2478
- Weber, G., (1995) J. Phys. Chem. 99(3), 1052-1059
- D'auria, S., Barone, R., Rossi, M., Nucci, R., Barone, G., Fessas, D., Bertoli, E., and Tanfani, F. (1997) Biochem. J. 323, 833-840
- Chene, P. (2001) Oncogene. 20(21), 2611-7
- Pandey, P., Kharbanda, S., and Kufe, D. (1995) Cancer Res. 55.18, 4000-3
- Hattrup, C. L. and Gendler, S. J. (2008) Annual Review of Physiology. 70, 431-457
- Singh, P. K., Wen, Y., Swanson, B. J., Shanmugam, K., Kazlauskas, A., Cerny, R. L., Gendler, S. J., and Hollingsworth, M. A. (2007) Cancer Res. 67, 5201-5210
- Cho, Y., Gorina, S., Jeffrey, P. D., and Pavletich, N. P. (1994) Science 265, 346 -355
- Clore, G. M., Ernst, J., Clubb, R., Omichinski, J. G., Kennedy, W. M., Sakaguchi, K., Appella, E., and Gronenborn, A. M. (1995) Nat. Struct. Biol. 2, 321-333
- Leng, Y., Cao, C., Ren, J., Huang, L., Chen, D., Ito, M., and Kufe, D. (2007) J. Biol. Chem. 282(27), 19321-30
Kinetic measurements of phosphorylated Y14 MUC1CT:p53
|Construct||KA (1/Ms)||KD (1/s)||Rmax (RU)||[Analyte] (um)||KA (1/M)||KD (M)||Req (RU)|
Table 1. Surface Plasmon Resonance binding measurements for phosphorylated YHPM MUC1CT with N-terminal (1-101), C-terminal (292-393) and full length p53 (1-393). Binding was conducted in HBS-EP buffer (0.01 M HEPES, 150 mM NaCl, 3 mM EDTA, 0.005% P20, pH 7.4) at 25oC, 5 ul/min, and 50 ul injection volume. (Chi2 = 0.456)
Thermodynamic measurements of phosphorylated Y14 MUC1CT:p53
|Interaction site||KA||KD (1/ KA)||?G||?Hcal (J)||?Hcal (Cal)||?S (J K-1)|
Table 2. Isothermal Titration Calorimetry free energy calculations, of 800 um phosphorylated Y14 MUC1CT interaction with 30 um C-terminal 53, made with equations ?G = ?H-T?S and ?G = -RTlnK. Free energy values are all endergonic indicating binding is non-spontaneous and therefore concentration dependent.
Figure 1. MUC1CT is only able to associate with the C-terminal portion and full-length p53 as determined by GST pull-down in 76r-30 cells stimulated with 100 ng/ml HGF. Surface Plasmon Resonance association of phosphorylated Y14 MUC1CT is observed with C-terminal and full length p53 only. Absence of binding occurred with phosphorylated Y29 and unphosphorylated MUC1CT peptides. Kinetic values are reported in the accompanying table (Table 1).
Figure 2. Isothermal titration of phosphorylated Y14 MUC1CT. Titration of 1x1-ul pre-injection followed by 29x5-ul injections of phosphorylated Y14 MUC1CT into ITC buffer containing 30-um C-terminal p53 (residues 292-393) resulted in a 4 interaction site model with molar ratio of 1.5 (A). Unphosphorylated MUC1CT titration into C-terminal p53 also yielded no dilution or association heats (B). After 5x5-ul dilution injections of 800-um phosphorylated Y14MUC1CT into ITC buffer (30mM Tris-HCl, 150 mM NaCl, pH 7.8), the dilution enthalpy was determined to be negligible (supplemental). The corresponding enthalpies, rates, and energy calculations for each site are shown (Table 2).
Figure 3. Continuous variation experiments (Job plot) for the determination of complex stoichiometry in 30 mM Tris-HCl, 150 mM NaCl buffer, pH 7.8, at 25degrees C. The wavelengths used are 220 and 280 nm (A). p53 aggregation experiments at 220 nm as a function of concentration; aggregates begin formation at ~50 mM.
Figure 4. Circular dichroism spectra for titration of p53 with phosphorylated Y14 MUC1CT. Mean CD spectra for three batches of phosphorylated Y14 MUC1CT and C-terminal p53 purified by SEC in Tris buffer used to obtain modification of the Far-UV spectrum of p53 upon phosphorylated Y14 MUC1CT binding. Far UV spectra of the apo forms and mixtures of 1 O.D. MUC1CT as either bound (A) or expected additive (B). The ellipticities shown in (A) are those of MUC1CT bound p53 after subtraction of buffer CD spectrum and those shown in (B) are what would be expected if phosphorylated Y14 MUC1CT and p53 were only additive with subtraction of the buffer CD spectrum.