New insights into cancer-related proteins provided by the yeast model

Authors


L. Saraiva, Laboratory of Microbiology, Faculty of Pharmacy, University of Porto, Rua Aníbal Cunha 164, 4050-047 Porto, Portugal
Fax: +351 222 003 977
Tel: +351 222 078 990
E-mail: lucilia.saraiva@ff.up.pt

Abstract

Cancer is a devastating disease with a profound impact on society. In recent years, yeast has provided a valuable contribution with respect to uncovering the molecular mechanisms underlying this disease, allowing the identification of new targets and novel therapeutic opportunities. Indeed, several attributes make yeast an ideal model system for the study of human diseases. It combines a high level of conservation between its cellular processes and those of mammalian cells, with advantages such as a short generation time, ease of genetic manipulation and a wealth of experimental tools for genome- and proteome-wide analyses. Additionally, the heterologous expression of disease-causing proteins in yeast has been successfully used to gain an understanding of the functions of these proteins and also to provide clues about the mechanisms of disease progression. Yeast research performed in recent years has demonstrated the tremendous potential of this model system, especially with the validation of findings obtained with yeast in more physiologically relevant models. The present review covers the major aspects of the most recent developments in the yeast research area with respect to cancer. It summarizes our current knowledge on yeast as a cellular model for investigating the molecular mechanisms of action of the major cancer-related proteins that, even without yeast orthologues, still recapitulate in yeast some of the key aspects of this cellular pathology. Moreover, the most recent contributions of yeast genetics and high-throughput screening technologies that aim to identify some of the potential causes underpinning this disorder, as well as discover new therapeutic agents, are discussed.

Abbreviations
Apaf-1

apoptosis-inducing factor 1

cyt c

cytochrome c

FASAY

functional analysis of separated alleles in yeast

IAP

inhibitor of apoptosis proteins

PFT

pifithrin

PKC

protein kinase C

ROS

reactive oxygen species

wt

wild-type

Introduction

Most of our knowledge about basic cellular processes has originated from model organisms [1]. The budding yeast Saccharomyces cerevisiae has provided a major contribution to fields as diverse as cell metabolism, DNA replication, recombination, cell cycle, cell death, protein folding, trafficking and organelle biogenesis [2,3]. Unpredictably, in recent years, this knowledge has been applied and expanded to an understanding of human diseases. Indeed, as a cell model of human diseases, yeast has provided insights into the basic processes underlying pathogenesis. However, as a unicellular organism, the obvious limitation of this cell system for the study of human diseases concerns an analysis of the disease aspects that rely on multicellularity and cell–cell interactions. Additionally, as a less complex system, some relevant genes involved in the pathology may not be present in the yeast genome. Although important aspects of human diseases lie beyond the reach of S. cerevisiae, this cell system has already proved its value as a first-line tool in the discovery of mechanistic processes involved in the disease. This has been possible because molecular interaction networks are unpredictably conserved from yeast to humans [2,4]. A remarkable example is the direct application of fundamental knowledge of cell cycle regulation, as uncovered in yeast, towards studies in human cancer biology [5]. Furthermore, with the finding that yeast can undergo an apoptotic cell death exhibiting phenotypic features and basic molecular machinery similar to those found in higher eukaryotes, the mechanisms of apoptosis could also be intensively addressed in yeast and the knowledge obtained transposed to human cells, providing clues towards an understanding of apoptosis-related diseases [6]. Yeast shows other undeniable advantages as a model organism as far as molecular studies are concerned. The yeast genome was the first eukaryotic genome to be sequenced [7] and, over the years, this knowledge has fueled whole-genome scale screening methods, including microarrays [8–10], two-hybrid analysis [11,12] and the use of deletion and overexpression libraries [13,14]. The yeast toolbox can also be used to unravel potential rescuing mechanisms because it is particularly well-suited to genetic suppressor isolation and chemical library screening [3,15]. In addition, yeast presents many technical advantages over other systems, such as a short generation time, ease of manipulation and a high amenability to genetic modifications.

The advantages presented above have fuelled the emergence of yeast models for several human diseases, including cancer. When establishing cancer-related protein models, different approaches are adopted depending on the degree of conservation of the protein under study. If the gene codifying for the protein is conserved in yeast (e.g. TOR1) [16], it is possible to directly study its function. If the gene has no orthologue in yeast, the heterologous expression of the human gene in this organism (the so-called ‘humanized yeast’) can still be highly informative because yeast may conserve protein interactions that give clues to its function and pathobiology. Yeast expressing the tumour suppressor p53 represents an example of this strategy [17]. The absence of orthologues of a protein or an entire pathway can sometimes even be advantageous because the protein can be studied in a simpler eukaryotic environment, without the interference of other proteins with similar or overlapping functions, as well as its endogenous regulators. For example, yeast has been used for the independent analysis of each isoform of the protein kinase C (PKC) family [17,18]. Typical approaches to study human proteins in yeast are presented in Fig. 1.

Figure 1.

 Typical approaches when studying human disease-related proteins in yeast. First, the gene of interest is cloned into a yeast vector with a constitutive or more often regulatable promoter because many disease-related proteins are toxic when expressed in yeast. Second, the cytotoxicity of a human protein is determined by optical density and colony-forming unit (cfu) counts. The optical density is a simple and fast method that can be used in genomic- and pharmacological-wide screens. The cfu counts allow distinction between decreased growth and increased cell death. When the human protein is involved in ageing or stress pathways, it may be needed to reproduce these conditions with yeast when assaying for protein function to uncover a phenotype. Third, whether the human protein-induced growth inhibition is the result of cell cycle arrest, stress or cell death can be assessed. Several techniques are available to distinguish between different types of cell death (apoptotic, necrotic or autophagic). Apoptosis can be identified for instance by assessing DNA fragmentation using terminal deoxynucleotidyl transferase dUTP nick end labelling, chromatin condensation upon staining with 4′,6-diamidino-2-phenylindole, and externalization of phosphatidilserine (PS) using Annexin V stainning. Propidium iodide, which only stains cells with ruptured cellular membranes, is used as a marker of necrosis. Autophagy can be assessed by monitoring the increase in mature Atg8p or the formation of autophagosomes by electron microscopy. Because many proteins affect vesicular trafficking, this pathway can be monitored by monitoring the uptake of the dye FM4-64. As a first-line model, all the discoveries made in yeast must be validated in more physiological models of the disease.

The present review focuses on the use of ‘humanized yeast’ as a model system for studying the major human proteins involved in cancer.

Caspase family members

It became evident that cellular and biochemical features of apoptosis are a consequence of the cleavage of a subset of proteins by proteases of the caspase family. Caspases are a conserved family of cysteine-dependent aspartate-specific proteases consisting of at least 15 members that can be divided into pro-apoptotic and pro-inflammatory subfamilies. The pro-apoptotic caspases can be further separated into activator or initiator caspases (caspase-2, -8, -9, -10 and -12) and executioner or effector caspases (caspase-3, -6 and -7) that are activated by the initiator caspases. It must be noted, however, that this classification is an oversimplification because there are situations where pro-apoptotic caspases (e.g. caspase-3) can mediate other responses than apoptosis, including cell differentiation and the activation of survival pathways [19]. Deregulations in the expression or activity of these proteases can lead to the development of several human apoptotic diseases, including cancer and neurodegenerative disorders. Hence, an understanding of the cellular function of caspase family members and the mechanisms behind their specificity and regulation has been the focus of extensive research [19,20].

The high complexity of mammalian caspase-signalling pathways led several research groups to investigate simpler eukaryotic systems as complementary cell models. Although yeast encodes a metacaspase (Yca1p) that shares structural homology and mechanistic features with mammalian caspases, major differences in the primary cleavage specificity have lead to the questioning of its classification as a ‘true’ caspase. This has been the subject of an intensive discussion [6,21]. Indeed, Yca1p has been implicated in the same cellular processes as mammalian caspases, namely in apoptosis [22]. However, although mammalian caspases specifically cleave their substrates after aspartic acid residues, metacaspases specifically cleave substrates after an arginine or lysine (basic residues) [23]. Although it still remains a controversial issue, recent studies have demonstrated that the activity of mammalian caspases in yeast (namely of caspase-2, -3, -4, -7, -8 and -10) is independent of Yca1p [24,25].

Yeast has therefore been widely exploited for the independent analysis of several members of the human caspase family with the aim of identifying their regulators and substrates. For example, a modified yeast two-hybrid system, based on the detection of β-galactosidase activity, was used to identify caspase-3 [26] and -7 [27] substrates. An approach to monitor the caspase activity was also developed in yeast using a reporter system consisting of a transcription factor linked by caspase cleavage sites to the intracellular domain of a transmembrane protein. Caspase activation induced the release of the transcription factor from the membrane, which in turn drove the transcriptional activation of a reporter gene, such as bacterial lacZ, therefore resulting in β-galactosidase activity dependent on caspase activation [28,29]. However, the most commonly used strategy for studying human caspases in yeast has relied on the fact that high expression levels of a caspase lead to a pronounced yeast growth arrest. Unlike caspase-1, -2, -4, -5, -8, -10 and -13, caspase-3, -6, -7 and -9 do not auto-activate when expressed in S. cerevisiae. Despite this, the activation of these caspases in S. cerevisiae could be achieved using different strategies as natural caspase activators, such as other caspases or the apoptosis-inducing factor (Apaf-1) (Fig. 2A). Indeed, the activation of caspase-3 in S. cerevisiae was successfully achieved by co-expression with caspase-8 or -10, which are two caspases that are efficiently processed and activated in yeast. Interestingly, these caspases were insufficient to catalyse the maturation of caspase-6. These results distinguished sequential modes of action for different caspases in vivo [30]. Additionally, although co-expression of caspase-9 and -3 in yeast had no cytotoxic effects, when the Apaf-1 was also expressed, a pronounced yeast cell death was obtained. This suggested the activation of caspase-9 by Apaf-1, which in turn activates caspase-3. Accordingly, the possibility of reconstituting the Apaf-1-activated pathway in yeast was demonstrated [31]. Auto-activation of caspases in yeast can also be obtained by using differently engineered auto-activated caspase variants (Fig. 2B–E). The caspase activity in yeast has been frequently associated with the induction of a cell death phenotype. Indeed, several studies have explored this feature for the functional analysis of human caspase-3, -8 and -10, as well as for the identification of yeast specific protein interaction partners and corresponding orthologues in human cells [24,25]. However, it was found that the phenotypic consequences of the expressed human caspase in yeast were highly dependent on the yeast strain used. This is probably a result of differences in the expression levels of endogenous caspase substrates and/or of other proteases with similar functions among the yeast strains [24,32]. Moreover, the strategy aiming to produce auto-activating forms of caspases was also referred to as having some impact on the final outcome of caspase activation in yeast. Indeed, different efficiencies on generating an active caspase can be related to different proteolytic activities. Although the co-expression of caspase-3 subunits as separate proteins caused yeast growth inhibition without the induction of cell death [33,34], the expression of the reverse-caspase-3 caused necrotic yeast cell death [24]. However, under our experimental conditions, the expression of a reverse form of human caspase-3 in yeast lead to a pronounced growth arrest [35] associated with an increase in DNA fragmentation and mitochondrial reactive oxygen species (ROS) production, as well as the maintenance of plasma membrane integrity (I. Coutinho and L. Saraiva, unpublished data). These results therefore indicate the induction of apoptotic (instead of necrotic) yeast cell death by human reverse caspase-3.

Figure 2.

 Different strategies of human caspase-3 activation in S. cerevisiae. (A) Natural caspase activators, namely caspase-8/-10 and Apaf-1. (B–E) Engineered auto-activated caspase variants that undergo spontaneous proteolytic processing or folding (also valid for caspase-6, -7 and -9). (B) Generation of reverse caspase, in which the small subunit precedes its prodomain and large subunit [27,30,31]. (C) Removal of the N-terminal prodomain from the caspase coding sequence [31]. (D) Separately co-expressing the large and small subunits of an active caspase [26]. (E) Joining in-frame the caspase cDNA to the coding regions for Escherichia coliβ-galactosidase (lacZ ) [36]. Activated caspases lead to yeast growth arrest, which can be abolished by co-expression with p35 or IAPs, and by the small molecule caspase-3 inhibitor Ac-DEVD-CMK. PD, prodomain; LS, large subunit; SS, small subunit.

Many aspects of the regulation of mammalian caspases by natural adapters could also be recapitulated in yeast. For example, it was shown that the cytotoxic effect of caspase-8 in S. cerevisiae [30] and of caspase-3 in Schizosaccharomyces pombe [32] was abrogated by the co-expression of the pan-caspase inhibitor baculovirus protein p35. Additionally, it was reported that inhibitors of apoptosis proteins (IAPs) in mammals, namely XIAP, suppressed the growth defect of S. cerevisiae expressing active human caspase-3 [34]. A subsequent study showed that the protective effect of IAPs could be abrogated by the co-expression of Drosophila pro-apoptotic proteins HID and GRIM or the mammalian protein DIABLO/Smac [31]. Taken together, these studies opened the possibility of using yeast-based caspase assays to screen for chemical and genetic inhibitors of caspase family members. Indeed, several yeast assays have recently been developed for the high-throughput screening of modulators of human caspases, using either chemical or cDNA libraries [29]. Native cell lysates from S. cerevisiae [24,30,32] and S. pombe [36] expressing a human caspase were used to establish a correlation between the effects of caspase expression on yeast growth and the proteolytic processing and enzymatic activity of these caspases. Accordingly, a yeast caspase-3 phenotypic assay, based on the measurement of yeast cell growth, was developed to search for small molecule inhibitors of this human caspase. In this assay, the correlation established between the reversion of caspase-3-induced yeast growth inhibition by a caspase-3 inhibitor (e.g. Ac-DEVD-CMK) and caspase-3 inhibition led to the discovery of new inhibitors of caspase-3 among a chemical library of vinyl sulfones [35]. The exploitation of these assays to screen for activators and inhibitors of individual caspase family members may help in the discovery of new therapeutic agents against cancer and neurodegeneration, respectively.

Bcl-2 family proteins

Bcl-2 family proteins include regulators of mitochondrial apoptotic signal transduction and, as such, have an important role in diseases such as cancer. The approximately 30 Bcl-2 family members can be classified, based on their different functions and corresponding BH domains, as both anti-apoptotic (e.g. Bcl-2, Bcl-xL and Mcl-1) and pro-apoptotic (e.g. Bax, Bak and BH3-only) proteins [37].

The co-existence of several members of the Bcl-2 family and the complex interactions and regulatory mechanisms in mammalian cells has hindered our knowledge about this family of proteins and, subsequently, about the emergence of associated diseases. In 1994, Sato et al. [38] found that the chimeric protein LexA-Bax was able to kill yeast, an effect that (as in mammals) could be prevented by the anti-apoptotic proteins Bcl-2 and Bcl-xL. Because Bcl-2 family proteins were able to conserve at least part of their functions when expressed in yeast, to overcome the problem of mammalian redundant pathways, this organism was selected by many groups for functional analysis. It is important to note that, although yeast was previously considered to be devoid of Bcl-2 family members, a Bcl-xL interacting protein harbouring a Bcl-2 homology (BH3) domain (Ybh3p) was recently identified in S. cerevisiae [121].

The observation that Bax also inserted into yeast mitochondrial membrane and caused cytochrome c (cyt c) release [39], a hallmark of Bax function in mammals, opened the way for the use of yeast in several studies concerning this process [40]. For example, several mitochondrial proteins, which were suspected to be required for Bax activity in mammals, have yeast orthologues and their roles were investigated in this model. Although the requirement of some proteins such as adenine nucleotide translocator is still considered controversial [41,42], F0F1-ATPase appears to be required for Bax toxicity in yeast [43,44]. An additional controversial issue concerns the type of Bax-induced yeast cell death (apoptotic versus autophagic). Some features in yeast cells expressing Bax, such as cyt c release from mitochondria [39], DNA fragmentation and phosphatidylserine exposure [45], supported an ‘apoptotic-like’ cell death involving the formation of an outer membrane mitochondrial apoptosis-induced channel, which allows cyt c release and is associated with oxidative stress and apoptosis induction [45,46]. Yet, the discovery that cyt c release was not essential for Bax-induced cell death, as well as the presence of autophagic markers (namely Atg8p maturation, vacular uptake of material and accumulation of autophagosomes), suggested a Bax-induced autophagic cell death [47,48]. Bax induction of autophagy was recently reported in mammalian cells [49]. The function of pro-apoptotic (as Bax and Bid) or of anti-apoptotic (as Bcl-xL and Bcl-2) proteins depends on their ability to translocate, oligomerize and insert into the mitochondrial membrane. Structural studies in yeast gave rise to the first descriptions of BH domains [50], as well as to the first insights into domains and residues that are critical for mitochondrial insertion and toxicity [51,52]. For example, although ablation of the Bax C-terminal domain was found to be dispensable for mitochondrial localization and cyt c release, the absence of the C-terminal domain of Bcl-xL prevented this protein from rescuing the cells from the effects of Bax [52]. Nevertheless, heterodimerization was not crucial for Bcl-xL inhibition of Bax toxicity [53,54]. Such studies, along with a previous study showing that Bcl-2 also prevented Bax-induced cell death independent of heterodimerization, provided the first evidence for heterodimerization-independent mechanisms of Bcl-2 family proteins [53,55].

The role of post-translational modifications in the regulation of Bcl-2 family members has been a subject of increased interest. Using yeast, it was observed that some phosphorylatable serine residues of Bax were able to affect its translocation to mitochondria, regulating the release of cyt c and a triggering of the apoptotic cascade [51]. Similarly, it was shown that the cytoprotective effect of Bcl-xL in yeast cells undergoing apoptosis was differently regulated by PKC isoforms through modulation of the Bcl-xL phosphorylated state [18] (as discussed below).

The cytotoxicity of Bax in yeast allowed the screening for toxicity suppressors as putative anti-apoptotic factors. The success of this approach was demonstrated by the identification of new mammalian apoptotic regulators, such as Bax inhibitor-1 [56], bifunctional apoptosis regulator [57] and Calnexin orthologue Cnx1 [58].

Overall, studies on Bcl-2 family proteins in yeast, mainly focussing on interactions among their members and structure–activity relationships, have provided important insights into the biology of these proteins. Despite the amenability of yeast to high-throughput screenings, only one study has used yeast for screening a library of BH3 peptides specific for Mcl-1 or Bcl-xL [59].

PKC family

The PKC family consists of at least 10 serine/threonine protein kinases grouped into three major PKC subfamilies according to their primary structure and the cofactors required for activation: classical (cPKCs; α, βI, βII and γ), novel (nPKCs; δ, ε, η and θ) and atypical (aPKCs; ζ and λ\ι). PKC isoforms are crucial regulators of cell proliferation and death. Consequently, it is not unexpected that the activity and expression of some of these kinases are altered in cancer. PKC isoforms therefore represent promising therapeutic targets for the treatment of this pathology [60,61].

A striking feature is that individual PKC isoforms can exert either similar or opposite effects in cell proliferation and death. PKCα, ε and ζ have been mostly described as promoters of cell proliferation and survival, and PKCδ as a cell death promoter. However, these roles have been widely questioned. Indeed, depending on the biological context, overlapping functions of a same PKC isoform in both promoting and inhibiting cell proliferation and death have been frequently reported [60,61]. The complexity of the PKC family (i.e. the coexistence of several PKC isoforms in a same cellular environment and the different expression profiles of PKC isoforms in different cell types) has contributed to such contradicting reports. An individual analysis of each PKC isoform would certainly contribute to a clear understanding about their biology and the identification of isoform-selective pharmacological modulators. This would help the development of new therapeutic strategies based on the selective modulation of these kinases. However, with respect to the limitations presented above, the exclusive use of mammalian cells would certainly hamper the achievement of such a goal.

In 1993, several studies performed by Riedel and colleagues demonstrated that yeast, (endogenous PKC, Pkc1p in S. cerevisiae, is a structural but nonfunctional homologue of mammalian PKC isoforms [62,63]), comprised a promising cell system for studying individual mammalian PKC isoforms [64–67]. These studies showed that mammalian cPKCα and βI were functionally expressed in yeast, leading to biological responses similar to those observed in mammalian cells. Indeed, the phorbol ester activation of PKCα and βI caused PKC down-regulation, uptake of extracellular Ca2+, Ca2+-dependence of cell viability, changes in cell morphology and an increase in the cell doubling time. Taken together, these findings indicated the conservation in yeast of a mammalian PKC signalling pathway. The observation that activation of a mammalian PKC isoform in yeast induced a specific phenotype (i.e. an increase in the cell doubling time that was proportional to the level of its enzymatic activity) led to the establishment of a yeast PKC phenotypic assay. As a sensitive and fast method for quantitatively measuring PKC activity, this yeast assay was successfully used for functional, molecular and pharmacological studies of mammalian PKC isoforms.

In a continuous effort to establish a PKC structure–function relationship, the yeast PKC assay has been extensively used. For example, using this assay, the role of cysteine-rich binding sites in the conserved C1 region of the PKC regulatory domain, as well as in the regulation of wild-type (wt) and mutants PKCα by several activators (e.g. 4β-phorbol l2-myristate 13-acetate, mezerein and indolactam V), was investigated [68,69]. Moreover, scanning mutagenesis studies performed in yeast helped in the identification of major regions within the PKCα regulatory domain that are important for phorbol binding and phorbol-dependent activation of the enzyme [70]. Additionally, deletion analysis of PKCα in yeast allowed the identification of a novel regulatory segment in the C2 region, described by amino acids 260–280, which, similar to the pseudosubstrate region, regulates PKCα activity by preventing its constitutive activation [71]. The observation that PKCα activity was not strictly regulated by the pseudosubstrate sequence led to other studies investigating the regions within the regulatory domain essential for PKCα auto-inhibition [72], which provided new data about the existence of intramolecular interactions between the regulatory and catalytic regions of PKCα that maintain the enzyme in its inactive conformation, thus regulating its catalytic activity and the physical access of PKC modulators to its target sites.

The yeast PKC phenotypic assay was also extensively used in the characterization of the potency and selectivity of known PKC activators [73] and inhibitors [74,75], as well as in the screening for new potent and selective PKC pharmacological modulators of individual mammalian PKC isoforms [74,76–78]. More recently, using yeast cells expressing PKCα, βI, δ, ε or ζ, a new potent and selective activator of nPKCδ and ε was discovered [79]. The applicability of this small molecule as an analytical probe for studying nPKCδ and ε cellular signalling pathways has been confirmed in mammalian cells [80]. Furthermore, the stimulation of an apoptotic pathway independent of anti-apoptotic PKC isoforms (α, βI and ζ), as revealed by a study in yeast [79], also supports the promising application of coleon U as an anticancer drug. Interestingly, it was also shown in yeast that, although 4β-phorbol l2-myristate 13-acetate activated nPKCδ and ε, inducing their translocation from the cytosol to the plasma membrane and G2/M cell cycle arrest, coleon U induced the translocation of these isoforms to the nucleus and metacaspase- and mitochondria-dependent apoptosis [79]. These results corroborated those obtained in mammalian cells showing that distinct stimuli can induce the translocation of a specific PKC isoform to distinct subcellular compartments, which was subsequently associated with distinct cellular responses [81]. Indeed, it was shown that the cellular localization of PKCδ regulates the survival/death pathway. Although the retention of PKCδ in the cytoplasm is compatible with cell survival, its nuclear retention is required for commitment to apoptosis [82,83]. Consistently, the nuclear targeting of kinases such as PKCδ is considered as a new and essential regulatory mechanism that directly influences the induction of apoptosis [82]. Thus, a broader outcome of this type of study was the validation of this cell model to unravel intra-organelle communication systems and their roles in the PKC isoform apoptotic signalling network.

The finding that yeast can undergo apoptosis with characteristics similar to mammalian cells [6] uncovered the possibility of exploiting this model organism for unravelling the role of individual mammalian PKC isoforms in the regulation of this cellular process, particularly of key apoptotic proteins such as Bcl-xL [18], Bax [84] and p53 [17,85].

Using yeast cells co-expressing an individual mammalian PKC isoform and the anti-apoptotic Bcl-xL protein, it was shown that PKC isoforms differently regulated Bcl-xL anti-apoptotic activity in acetic acid-induced yeast cell death by affecting its phosphorylated state. Although PKCδ had no effect on Bcl-xL activity, PKCα abolished its anti-apoptotic effect through an increase of the Bcl-xL phosphorylated form. On the other hand, PKCε and ζ enhanced the Bcl-xL activity, decreasing the Bcl-xL phosphorylated form. Consistent with the results obtained in mammalian cells, the results obtained in yeast showed that Bcl-xL phosphorylation disables its anti-apoptotic function (Fig. 3). Du et al. [86] proposed the existence of a kinase and phosphatase system in mammalian cells that may be operating in tandem, leading to a coordinated phosphorylation–dephosphorylation cycle that modulates Bcl-xL activity. However, the precise mechanisms for this modulation remain undetermined. The study in yeast provided new insights into the role of phosphorylation on the modulation of the Bcl-xL function, identifying individual PKC isoforms as regulators of the phosphorylation–dephosphorylation state of Bcl-xL (Fig. 3). Additionally, it corroborated several studies with respect to the identification of Bcl-2 family members as major apoptotic targets of PKC isoforms [87]. Indeed, it was recently demonstrated that PKCα also regulates Bax in yeast [84]. The pro-apoptotic activity of this Bcl-2 family protein depends on its ability to translocate, oligomerize and insert into the mitochondrial membrane after stress [40]. Using yeast cells co-expressing Bax and PKCα, it was shown that PKCα increased the translocation and insertion of Bax into the outer mitochondrial membrane. This was associated with an increase in cyt c release, ROS production, mitochondrial network fragmentation and autophagic cell death. Curiously, this PKCα effect was revealed to be independent of its kinase activity. This suggests that PKCα can promote apoptosis by a kinase-independent way [84].

Figure 3.

 Regulation of Bcl-xL activity and phosphorylated state by mammalian PKCα, ε and ζ. PKCα abolished the Bcl-xL anti-apoptotic activity by Bcl-xL phosphorylation, directly or through activation of other kinase. PKCε and ζ enhanced the Bcl-xL anti-apoptotic activity most likely through phosphorylation and, consequently, the activation of a phosphatase responsible for the decrease of the Bcl-xL phosphorylated form.

By contrast to human wt p53, which induces yeast growth arrest when expressed in yeast, in the absence of an exogenous activator (e.g. phorbol ester), PKCα, δ, ε and ζ do not interfere with yeast cell growth [22]. Using yeast cells co-expressing PKCα, δ, ε or ζ and wt p53, the role of individual PKC isoforms in the regulation of p53 activity was studied. The results obtained revealed a differential regulation of p53 activity and phosphorylation state by PKC isoforms [17,85]. In an initial study, it was shown that, although PKCα reduced p53-induced yeast growth inhibition/cell cycle arrest and PKCζ had no effect on p53 activity, nPKCδ and ε enhanced p53 activity through its phosphorylation at Ser376-378 [17]. Similar results were obtained in the presence of an apoptotic stimulus. In this case, although cPKCα and aPKCζ had no effect on p53 activity, nPKCδ and ε stimulated transcription-dependent and -independent p53-mediated apoptosis [85] (Fig. 4).

Figure 4.

 Regulation of transcription-dependent and -independent p53 mechanisms by mammalian nPKCδ and ε in yeast. (A) Human wt p53 expressed in yeast presents nuclear localization and induces growth inhibition associated with S-phase cell cycle arrest, which is abolished by the selective inhibitor of p53 transcriptional activity, PFT-α. Co-expression of mammalian nPKCδ or ε markedly increases the p53-induced growth arrest through p53 phosphorylation [17,85]. (B) Under cell death conditions induced by H2O2 (cellular stress is highlighted with a yellow flash), nPKCδ and ε stimulate the translocation of a fraction of p53 from the nucleus to the mitochondria, and an increase in mitochondrial ROS production, mitochondrial transmembrane potential (ΔΨm) loss and mitochondrial network fragmentation. This p53-mitochondrial apoptotic pathway is markedly reduced by the absence of respiration (using strains devoid of mitochondrial DNA, rho0 mutants) and by the selective inhibitor of mitochondrial p53 translocation, PFT-μ [85].

Indeed, important insights with respect to the mechanisms of apoptotic function of PKC isoforms, pharmacological modulators and physiologically relevant substrates have been provided by yeast PKC assays (Table 1). Concerning PKCα, although frequently described as an anti-apoptotic protein [61], contradictory results suggesting a pro-apoptotic activity have also been reported [88]. This pro-apoptotic function of PKCα was corroborated in yeast. Indeed, this isoform stimulated acetic acid-induced apoptosis, abolished the Bcl-xL anti-apoptotic effect and stimulated the pro-apoptotic activity of Bax. Despite this, p53 does not appear to be an apoptotic substrate of PKCα (Table 1). Concerning PKCδ, the well-known pro-apoptotic function of this isoform in mammalian cells [61,87] was widely confirmed in yeast. PKCδ stimulated acetic acid- and coleon U-induced apoptosis and activated p53 (Table 1). Although several studies have reported the phosphorylation of p53 and the activation of its transcriptional mechanism by PKCδ in mammals [89–92], work in yeast has revealed for the first time the involvement of this isoform in the regulation of a p53 transcription-independent mechanism. Interestingly, stimulation of H2O2-induced apoptosis by PKCδ in yeast was only achieved in the presence of p53 [85]. This corroborated similar studies performed in mammalian cells, in which the stimulation of H2O2-induced apoptosis was almost abolished in the presence of rottlerin (a selective inhibitor of PKCδ) and in null-p53 cells [89]. This may suggest that, for this stimulus, the pro-apoptotic activity of PKCδ is highly dependent on a specific substrate that appears to be p53. It also suggests that Bcl-xL is not a PKCδ apoptotic substrate (Table 1). Concerning PKCε, this isoform is frequently regarded as having anti-apoptotic properties in mammalian cells [61,87]. Although the mechanisms responsible for its anti-apoptotic function are still not well clarified, it appears to involve the regulation of Bcl-2 family proteins [87], such as the inhibition of the pro-apoptotic proteins Bax and Bad [93,94]. Indeed, the yeast cell model revealed that PKCε also regulates Bcl-xL anti-apoptotic activity [18].

Table 1.   Effects of mammalian PKC isoforms on distinct apoptotic stimuli and human apoptotic proteins expressed in yeast. (↑), Increase of effect; (↓), decrease of effect; (NE), no effect; (), not determined.
 cPKCαnPKCδnPKCεaPKCζReferences
Apoptotic stimuli
 Acetic acid[18]
 H2O2NENENE[85]
 Coleon UNENE[79]
Apoptotic proteins
 Bcl-xLNE[18]
 Bax[84]
 p53NENE[85]

Although several studies suggest that PKCε favours life over death, other studies report the involvement of this isoform in apoptosis promotion [95]. When expressed in yeast, PKCε stimulated acetic acid-, coleon U- and H2O2-induced yeast apoptosis and p53 activity (Table 1). Taken together, the results obtained in yeast suggest that the apoptotic function of PKCε is highly dependent on its accessibility to key apoptotic proteins, such as Bcl-xL and p53. The translocation of PKCε to distinct subcellular compartments (e.g. as a result of the distinct stimuli applied) may expose the isoform to distinct substrates and this may be responsible for its distinct activities. Concerning PKCζ, its anti-apoptotic role in mammalian cell death signalling pathways is also well-accepted [61]. However, it was observed that PKCζ stimulated acetic acid-induced yeast apoptosis. Despite this, when this isoform was co-expressed with Bcl-xL, it markedly enhanced the Bcl-xL anti-apoptotic activity, with a complete abolishment of acetic acid-induced apoptosis [18]. Interestingly, the yeast assay also suggests that p53 is not an apoptotic substrate of PKCζ (Table 1). Taken together, although PKCα, δ, ε and ζ stimulated acetic acid-induced apoptosis, only PKCε stimulated the H2O2 apoptotic effect (Table 1). This indicates that, for H2O2, the apoptotic substrates of PKCα, δ and ζ are probably not conserved in yeast. These studies emphasize that the apoptotic function of a specific PKC isoform is highly dependent on the stimulus applied.

Overall, valuable insights into the genetic, molecular and functional profile of individual PKC isoforms have been provided by the yeast PKC assays. Additionally, potent and selective small molecule modulators of these kinases have been identified. The studies discussed above certainly represent an important contribution to the design of new therapeutic strategies against cancer through the identification of key apoptotic targets amongst the PKC family isoforms.

p53 tumour suppressor protein

The p53 tumour suppressor protein is a sequence-specific DNA-binding transcription factor that regulates the expression of an assortment of genes involved in cell cycle, apoptosis and numerous other processes. Mutations in the TP53 gene are a feature of 50% of all reported cancer cases. In the other cases where the TP53 gene itself is not mutated, the p53 pathway is often partially inactivated. This exemplifies the critical role of p53 in human cancers. The regulation of the p53 activity is therefore considered as a hallmark in the fight against cancer. Hence, over the last 30 years, p53 has become the focus of intensive basic and clinical research [96].

The versatility of the yeast cell model for studying several cellular processes, such as cell cycle and apoptosis, associated with the fact that no orthologues of human p53 have been discovered so far in this eukaryotic cell [97,98], has justified the attractiveness of this cell system for studying p53. Indeed, despite the complexity of p53 regulation in higher eukaryotes, at the beginning of the 1990s, it was shown that human wt p53 was also a sequence-dependent transcription factor in yeast [99,100]. Additionally, it was shown that p53 activating kinase (PAK1) was an essential gene for the activation of p53 transcriptional activity in yeast [101]. Subsquently, additional evidence was provided corroborating the remarkable similarities between the p53 transcriptional activity in yeast and mammalian cells [98]. With this interesting discovery, yeast became a powerful tool for uncovering major aspects of the function of p53. For example, using yeast, regulators of p53 transcriptional activity were identified. It was shown that the p53 transcriptional activity in yeast was subjected to redox regulation and required thioredoxin reductase [102], as subsequently confirmed in mammalian cells [103]. Moreover, Ishioka et al. [104] developed the technique of functional analysis of separated alleles in yeast (FASAY) with the aim of understanding the status of p53 in cancer cells. Several modified versions of this FASAY assay were subsequently used to identify tumour-derived p53 gene mutations and to understand how these mutations could interfere with the p53 function in human tumour cells (Fig. 5). Indeed, the identification and classification of tumour-derived p53 mutants based upon their degree of loss of function has a high clinical value. Yeast-based p53 functional assays were also revealed to be a powerful tool for molecular epidemiology. The assumption that carcinogens leave fingerprints suggested that an analysis of the frequency, type and site of mutations in genes frequently altered in carcinogenesis may provide clues with respect to the identification of factors contributing to carcinogenesis [97,105]. Using these assays, it was demonstrated that p53 mutants had a partial loss of transcriptional activity [106]. The ability and efficacy of p53 mutants to bind and activate crucial reporters of cell cycle and apoptosis were also studied in yeast [107,108]. Moreover, 2314 p53 mutants, representing all the possible amino acid substitutions caused by a point mutation, were constructed, expressed and evaluated in yeast aiming to understand the effect of all these mutations on p53 function [109]. In another study, yeast was used to identify intragenic suppressor mutations that were able to restore the activity to nonfunctional p53 mutants [110].

Figure 5.

 The p53 FASAY assay. It consists of the amplification of p53 cDNA by RT-PCR (using p53 mRNA from tumour samples or other tissues) and the co-transformation of a yeast reporter strain with the PCR product and a gapped expression plasmid. (A) p53 status is evaluated using plates without histidine. Yeast clones expressing a functional p53 (p53wt) are His+ as a result of a capacity to express the HIS3 reporter gene (REP) and will therefore grow; yeast clones expressing a mutated p53 cDNA (p53mt) are His- and will not grow. (B) A modified version of the FASAY assay, with ADE2 instead of HIS3 as a REP [122]. Using plates with low concentration of adenine, p53 mutations are identified via the colour of the colonies. White colonies indicate a wt p53 that allowed ADE2 expression and red colonies indicate a mutant p53. (Adapted from [123]).

Recently, an elegant yeast p53 transactivation assay for the high-throughput screening of factors that can influence p53 function, including mutations, cofactor proteins and small molecules, was developed [111]. Using this assay, the interaction of p53 with its endogenous negative regulators, MDM2 and 53BP1, was analyzed. The results obtained showed that MDM2 and 53BP1 led to a reduction in the p53 transactivation activity. However, unlike MDM2, which also reduced the transactivation activity of p53 mutants with a partial transactivation function, the impact of 53BP1 was lost or greatly reduced with specific partial function p53 mutants. Concerning the ability of human MDM2 to reduce p53 transactivation activity in yeast, as initially reported in 1993 [112], it was also shown that MDM2 can interact with endogenous yeast pathways to ubiquitylate and sumoylate p53. Ubiquitylation led to p53 degradation, whereas sumoylation was essential for the localization of p53–MDM2 complexes to yeast nuclear bodies [113].

The induction of growth arrest by human wt p53 expressed either in S. cerevisiae [114] or S. pombe [115] has also been highly explored in the study of p53. In S. cerevisiae, it was reported that wt p53 caused a mild growth inhibition [114,116], which was markedly increased by the co-expression of human cell cycle-regulated protein kinase [114]. Similarly, it was also observed that the expression of human wt p53 in S. cerevisiae induced growth inhibition associated with S-phase cell cycle arrest, which was markedly increased by PKCδ and ε. Additionally, using the selective p53 transcriptional inhibitor pifithrin (PFT)-α, a complete abrogation of p53-induced growth arrest was obtained. This supported the transcription of genes involved in the yeast cell cycle by human p53 [79] (Fig. 4). Interestingly, with the exploitation of this p53 phenotype in S. pombe, it was shown that p53-induced cell cycle arrest was abrogated by the cell cycle regulator protein phosphatase cdc25 [115]. Subsequently, it was shown that the regulation of cell cycle progression by p53 in mammalian cells involved its interaction with cdc25 [117]. More recently, Hadj Amor et al. [118] reported that the expression of wt p53 in S. cerevisiae resulted in apoptotic cell death. It was also revealed that p53 interfered with the expression of the gene encoding the anti-apoptotic and anti-oxidative protein, thioredoxin. Because this protein has a crucial function in the protection of yeast cells against ROS, these results suggested that p53 may induce apoptosis in part by down-regulation of anti-apoptotic proteins. These different phenotypes obtained with the expression of human p53 in yeast may be attributed to different p53 cellular levels as a result of the use of different expression vectors.

Apoptosis induced by p53 is firmly established as a central mechanism of tumour suppression [96]. In 1995, it was noted that the overexpression of a mutant p53, lacking most of the DNA binding domain and completely deficient in transactivation function, could nonetheless trigger apoptosis [119]. That study was the first reference concerning the capacity of p53 to induce apoptosis independent of its transcriptional function. At present, despite the prominence of p53 nuclear transcriptional activity in the induction of apoptosis, a transcription-independent mechanism involving mitochondrial p53 translocation is receiving increased attention. Indeed, recent findings have provided encouragement regarding further exploration of the potential of mitochondrial p53-based cancer therapeutics [96,120]. The first evidence for the conservation in yeast of this transcription-independent p53-mediated apoptosis was recently provided [85] (Fig. 4). It was also shown that, besides the activation of p53 transcriptional activity, nPKCδ and ε triggered the p53 translocation to mitochondria. This provided new insights about the regulation of p53 translocation to mitochondria. Additionally, the results obtained supported the possibility that mitochondrial p53 localization is a subtle deciding factor that dictates whether cells die or arrest growth. Indeed, mitochondrial p53 localization was only detected in yeast under an apoptotic cell death scenario.

The direct participation of p53 in the intrinsic mitochondria-mediated apoptotic pathway may involve interaction with the multidomain members of the Bcl-2 family, and particularly the activation of one of its pro-apoptotic members Bax or Bak, to induce mitochondrial outer membrane permeabilization. However, the underlying mechanisms remain to be completely clarified. The recent discovery of Ybh3p in yeast [121] may contribute to an understanding of how interactions between p53 and Bcl-2 family members promote mitochondrial outer membrane permeabilization.

Taken together, the studies referred to above show that, although no p53 orthologues have been found in yeast, the p53 pathway is highly conserved in this eukaryotic organism. Additionally, several lines of evidence are provided that highlight yeast as a powerful cell model for use in genetic, molecular and functional studies of p53.

Concluding remarks

Besides the recognized contribution of yeast to our present understanding of many fundamental aspects of cellular biology in higher eukaryotes, in recent years, yeast has proved to be a powerful model organism for unravelling the molecular basis of complex human diseases such as cancer. Indeed, the ease of genetic manipulation of yeast cells, in association with the conservation of many features of higher eukaryotic physiology, a tractable genome, a short generation time and a large network of researchers who have generated a vast arsenal of experimental tools, has justified the extensive use of this model organism in biomedicine.

In this review, the impact of using ‘humanized yeast systems’ in the dissection of cancer-related molecular processes and in the identification of genetic and pharmacological regulators of this disease is discussed. Studies involving the heterologous expression of proteins in yeast are sometimes viewed with skepticism, with toxicity often considered as a mere response to unphysiological expression levels. Although it is unlikely that all of the protein functions revealed by yeast will be relevant for human diseases, the numerous insights into protein functions that are obtained from yeast justifies the exploitation of this cell system.

More recently, sequencing of the human genome promised the identification of new disease-causing genes. This brought an enormous expectation for the therapy of severe human disorders such as cancer. However, this gave rise to the discovery of genes whose normal function in these pathologies is commonly unknown. The genetic manipulations required to uncover gene function are often extremely difficult to perform in human cells. This lead to the ‘rebirth’ of yeast as a valuable tool for uncovering the function of human genes. We therefore anticipate that promising new discoveries will continue in this area using yeast as a model organism.

Acknowledgements

This work was supported by FCT (Fundação para a Ciência e a Tecnologia) and FEDER funds through the COMPETE program under the project FCOMP-01-0124-FEDER-015752 (reference FCT PTDC/SAU-FAR/110848/2009) and by the University of Porto/Santander Totta. This work was also supported by FCT through REQUIMTE (grant number PEst-C/EQB/LA0006/2011) and through the fellowships of I. Coutinho (SFRH/BD/36066/2007), M. Leão (SFRH/BD/64184/2009), J. Soares (SFRH/BD/78971/2011) and C. Pereira (SFRH/BPD/44209/2008).

Ancillary