A variety of proteases have been implicated in yeast PCD (programmed cell death) including the metacaspase Mca1 and the separase Esp1, the HtrA-like serine protease Nma111, the cathepsin-like serine carboxypeptideases and a range of vacuolar proteases. Proteasomal activity is also shown to have an important role in determining cell fate, with both pro- and anti-apoptotic roles. Caspase 3-, 6- and 8-like activities are detected upon stimulation of yeast PCD, but not all of this activity is associated with Mca1, implicating other proteases with caspase-like activity in the yeast cell death response. Global proteolytic events that accompany PCD are discussed alongside a consideration of the conservation of the death-related degradome (both at the level of substrate choice and cleavage site). The importance of both gain-of-function changes in the degradome as well as loss-of-function changes are highlighted. Better understanding of both death-related proteases and their substrates may facilitate the design of future antifungal drugs or the manipulation of industrial yeasts for commercial exploitation.
- programmed cell death (PCD)
- Saccharomyces cerevisiae
Understanding the molecular mechanisms regulating and executing PCD (programmed cell death) in fungi is important from a cell biology perspective as some mechanisms may be common to all eukaryotes ). It may also facilitate the design of future antifungal drugs  or the manipulation of industrial yeasts. The model yeast Saccharomyces cerevisiae has the capacity to undergo PCD in response to a range of environmental triggers or endogenous signals . According to the nature and strength of the death-inducing stimulus, PCD in yeasts can be primarily apoptotic, although secondary necrosis invariably follows . Programmed necrosis has also been reported in response to ROS (reactive oxygen species), ER (endoplasmic reticulum) stress and depletion of polyamines . PCD is a complex process, often involving the controlled de-construction of the cellular machinery in a manner regulated and implemented by proteases. A variety of proteases have been implicated in yeast PCD and these are the subject of this paper.
The yeast degradome and the proteolytic landscape
Currently, 112 yeast proteases and 60 non-protease homologues of proteases are listed in the MEROPS database (available at http://merops.sanger.ac.uk). Lineage-specific expansion is apparent for some families of protease in yeast, including calpains, ubiquitin-specific peptidases, amidophosphoribosyltransferases, pitrilysin-like metalloproteases, aminopeptidase Ys, subtilisins and S59 family nucleopore-associated endopeptidases, making them attractive targets for antifungal research. Numerous studies have focused on yeast homologues of known mammalian or plant PCD regulators. In most cases, these studies have confirmed a role for the proteases in yeast PCD, highlighting the functional conservation of cell-death mechanisms across the tree-of-life.
Proteases associated with PCD in yeast
Many forms of metazoan apoptosis are dependent on caspases . True caspases appear to be restricted to the metazoa, but distant homologues, paracaspases and metacaspases, exist in fungi, plants and protists . Caspases, legumains, separases, clostripains and gingipains are MEROPS C14-family (clan CD) proteases. Yeast possesses a single metacaspase-encoding gene, MCA1 , widely reported to participate in PCD . Metacaspases are believed to lack true caspase activity and specificity [10–12], preferring arginine and lysine to aspartate at the P1 position. Most studies have focused on substrates of type-II metacaspases, whereas yeast Mca1 is a type-I metacaspase. Heterologous Mca1 expression and in vitro cleavage of synthetic peptides may also have affected specificity. Lee et al.  reported that some native substrates of yeast Mca1 are probably cleaved at aspartate residues, a characteristic of true caspases. Additionally, metacaspases and true caspases can share common death-related substrates, with both caspase 3 and mcII-Pa from Picea abies cleaving Tudor staphylococcal nuclease . Although H2O2, osmotic stress and numerous other stimuli induce metacaspase-dependent cell death in yeast [9,15], there are many reports of Mca1-independent PCD as well .
H2O2-induced apoptosis triggers cleavage of Mcd1, a yeast homologue of the Rad21 cohesin required for dsDNA (double-stranded DNA) repair, sister chromatid cohesin and chromosome condensation . Cleavage is dependent on Esp1, a caspase-like separase normally associated with an inhibitory securin, Pds1. During normal growth Pds1 is degraded by the anaphase-promoting complex, releasing Esp1 that cleaves Mcd1 (and substrates such as Slk19 and Scc1); the fragments remain in the cytoplasm. However, in apoptotic cells, a 40-amino-acid C-terminal peptide fragment translocates from the nucleus to the mitochondria. This is a gain-of-function cleavage, since the mitochondrial import of the Mcd1 fragment leads to a loss of mitochondrial membrane potential, contributing to the onset of apoptosis. Overexpression of Esp1 causes accumulation of Mcd1 in mitochondria, even without external stimuli. Many cells produce endogenous ROS that is required for mitochondrial import. Esp1 has caspase 1-like activity  and caspase 1 inhibitors reduce Mcd1 degradation.
Nma111 is an HtrA-like serine protease with a PDZ domain, which in response to cellular stresses forms intranuclear aggregates . Nma111 contains HtrA-like sequence repeats, but only the N-terminal repeat has a functional catalytic domain. As with many trypsin-like serine proteases, the active site contains a GGSGS consensus motif and also contains a nuclear localization signal, with nuclear import mediated by Kap95 . Cells lacking Nma111 survive heat-shock better than wild-type cells and fail to undergo H2O2-induced apoptosis. Overexpression enhances apoptosis under stress conditions. Nma11 can cleave the only yeast IAP (inhibitor of apoptosis protein) Bir1 , although Bir1 cleavage is not sufficient to induce PCD. Bir1 itself does not, however, appear to interact with Mca1 as might be expected of a bona fide IAP.
Disruption of N-glycosylation of proteins within the ER leads to apoptosis in yeast . MCA1 deletion does not abrogate this response ; however, loss of Kex1 decreases caspase-like activity within dying cells and improves the survival of cells suffering ER-stresses, acetic acid-induced PCD or chronological aging . Kex1 is a cathepsin A-like serine carboxypeptidase, preferring basic amino acid residues and C-terminal arginine and lysine residues. It is unclear if Kex1 has caspase-like activity or if downstream proteolytic components include such activity. Purified Kex1 cannot cleave synthetic caspase 6 substrates  but other caspase substrates have not yet been tested.
Vacuoles replace lysosomes in fungi. Yeast vacuoles contain acid proteases including proteinase A (Pep4), proteinase B (Prb1), carboxypeptidase Y (Prc1), carboxypeptidase S (Cps1) and aminopeptidase I (Lap4). Vacuolar and lysosomal proteases are mainly non-specific and are synthesized as inactive zymogens, passing via the ER and Golgi to the vacuoles . Lysosomal endopeptidases are responsible for the majority of bulk protein breakdown in yeast, dispose of protein aggregates and degrade membrane proteins. Aminopeptidases and carboxypeptidases finally reduce the products to free amino acids. Complete digestion is so swift that the degradation intermediates are virtually undetectable . Pep4, the yeast equivalent of cathepsin D, is principally located in the vacuole. During H2O2-and acetic acid-induced apoptosis, Pep4 is released from the vacuole and relocates to the nucleus and cytoplasm, initiating mitophagy-independent degradation of mitochondria and nucleoporins .
The ubiquitin–protease system has long been implicated as a key player in yeast PCD [26,27]. Inactivation of UBP10 (ubiquitin-processing protease 10), a deubiquitinating enzyme, yields a complex phenotype characterized by a subpopulation of cells exhibiting typical markers of apoptosis. Cells lacking Ubp10 possess endogenous caspase-like activity, which is only partially lost upon deletion of MCA1. Loss of Mca1 suppresses the ubp10-null mutant phenotype, whereas Mca1 overexpression, even without external stimuli, impairs growth and viability of ubp10 cells. Later stages of acetic acid-induced PCD in yeast require a transient elevation of proteasome-related chymotrypsin, trypsin and PGPH (peptidylglutamyl peptide hydrolase)-like activities . Levels of proteasomal catalytic components (Pre2, Pre3) appear to be unchanged, suggesting a role for a proteasome regulatory component, or altered proteasome accessibility/localization. Inhibition of the proteasomal chymotrypsin-like activity significantly reduces acetic acid-induced PCD . In many plant and mammalian cell-types, blocking of proteasome function promotes PCD . In yeast, overexpression of Stm1, a known target of the proteasome, promotes apoptosis, but shows a deficiency in apoptosis when deleted . Therefore, under some cell-death scenarios, the activation of the proteasome may be important for the prevention of yeast PCD as well as its activation.
Caspase-like activities associated with yeast PCD
Positional scanning of synthetic peptides  and peptide-based synthetic substrate libraries  have been used to ascertain caspase substrate specificities and to assess caspase-like activities during yeast and plant PCD [8,11,15,33–39]. Activities against caspase 3, 6 and 8 substrates, but not caspase 1 substrates, have been reported in yeast; see Table 1. This contrasts with PCD-associated caspase-like activities in plants which commonly display caspase 1-like activity . Plants contain type-I and type-II metacaspases, whereas fungi only possess type-I, which may explain some differences. Overexpression of yeast Mca1 leads to elevated caspase 6-and 8-like activities , but deletion of Mca1 does not remove all activities. Since pro-apoptotic stimuli such as ageing and osmotic stress elicit strong caspase-like activities, even in the absence of Mca1, other proteases with caspase-like activity must be partly responsible.
Several metacaspase-independent death-associated caspase-like activities have been reported in plants. Subtilisin-like serine proteases have caspase-like activity  and DEVDase and TATDase activities have been identified . A screen of 119 non-essential yeast proteases reveals 38 that show some defects in either caspase 1, 6 or 8 substrate cleavage and seven (Yps7, Ybl091ca, Map1, Pim1, Lap2, Aap1 and Rbd2) show reduced death upon stimulation with pro-apoptotic doses of acetic acid (Figure 1A). These proteases are up-regulated during PCD, but show limited changes in stressed cells (Figures 1B and 1C). The future challenge is to unravel the primary and secondary roles of these caspase-like proteases and to identify their downstream targets. The challenge is complex since some proteases target multiple proteins and many substrates are degraded by several proteases.
Inhibitors of caspases often block or reduce the level of PCD in yeast [21,33]. Again plants and fungi differ. Plant PCD is commonly affected by caspase 1 and 3 inhibitors, whereas caspase 3 and 6 inhibitors appear more potent in fungi. The importance of these activities remains to be seen, as general cysteine and serine protease inhibitors also block plant PCD, without affecting overall caspase-like activity.
Protein abundance is a balance between translation rates and degradation rates. Using TAP (tandem affinity purification)-tagged protein expression constructs and a translational block with cycloheximide, Belle et al.  measured the half-lives of 3751 yeast proteins. The median half-life was observed to be 43 min, but many proteins were much more unstable. Sequencing-based MS/MS (tandem MS) of the yeast degradome gives more detail about yeast protein turnover . During mid-exponential-phase growth, 1111 peptide fragments originating from 205 parent protein substrates were observed. Cytoplasmic and mitochondrial proteins were the largest components. No proteins uniquely associated with the cell wall or nucleus were identified, nor were membrane proteins (perhaps reflecting technical bias in the extraction procedures used). Although the authors asserted that ‘lysosomal’ degradation played little role in the breakdown of the substrates, roles for vacuolar proteases cannot be excluded since vacuolar persistence may be transient and yeast PCD is sometimes associated with translocation of peptides between subcellular compartments, for example, Pep4 . Strong preferences for P1-site cleavages at lysine>leucine>arginine>alanine>methionine>serine>threonine>phenylalanine>tyrosine>asparagine are observed, whereas cleavages at aspartate, glutamate, glycine, isoleucine, proline and valine are detected at much lower levels than expected. P1′-sites have a preference for amino acids with short side chains e.g. glycine, alanine, serinine and threonine. Degradation of mitochondrial proteins is slightly different with P1-site cleavages at lysine>leucine>serine>phenylalanine>arginine=alanine>valine>glutamine and P1′-site cleavages at glycine>arginine>serine>lysine>glutamate>threonine=alanine>aspartate, implying distinct processing routes.
The detection of arginine–arginine, lysine–arginine and lysine–lysine cut sites suggests the activity of pro-protein convertases such as Kex2. Trypsin-like and chymotrypsin-like proteolysis, favouring lysine–arginine and phenylalanine–tyrosine cuts was also highlighted, perhaps attributable to the proteasomal peptidases Pup1 and Pre2. Leucine and alanine cleavage at P1-sites might equate to the activity of proteasomal components as well . Surprisingly, over half of the degraded cytosolic regulatory proteins and enzymes were not ubiquitinylated, highlighting the importance of non-proteasomal pathways or ubiquitin-independent proteasomal activities in the degradome.
Studies of global proteolysis have been undertaken in human cells undergoing apoptosis [43,44], facilitating both substrate and cleavage site identification. Dix et al.  utilized PROTOMAP, a technology combining SDS/PAGE with shotgun LC-MS/MS, to compare control and staurosporine-treated cells, whereas Mahrus et al.  captured novel N-terminal peptides from etoposide-induced apoptotic cells. In these studies alone, a total of 261  and 292  caspase substrates were reported. The two datasets overlap (64 proteins) and both datasets overlap with substrates listed in the CASBAH database . Substrates specifically cleaved during the cell death response include ROCK1 (Rho-associated kinase 1), linked to membrane blebbing and the CAD/DFF40 endonuclease. Often multiple components of biosynthetic and transcriptional complexes are targeted . Future studies of the degradome in wild-type yeast and protease deficient/overexpression strains could reveal much about fungal PCD.
Evolutionary conservation of caspase substrates and cleavage sites
Some caspase substrates are cleaved in a species-specific manner (e.g. cyclin A is targeted during apoptosis of Xenopus oocytes, although there is no homologous cleavage site in mammals ). Nevertheless, it seems unlikely that similar death phenotypes arise from proteolysis of completely different complements of target substrates in different species . Indeed, Shao et al.  revealed that the P4–P3′ caspase 1 cleavage site in GAPDH (glyceraldehyde-3-phosphate dehydrogenase) (Asp189) is actually conserved in organisms as diverse as man, mouse, cow and yeast. The CASBAH database (http://bioinf.gen.tcd.ie/casbah/) describes 793 caspase substrates (mostly from humans, rats and mice), with cleavage site details for 586 proteins . A total of 241 yeast homologues of the caspase substrates listed in the database can be found in the yeast proteome. Since only 16% of human proteins are conserved in S. cerevisiae , the conservation of 41% of caspase substrates is much higher than expected by chance. Moreover, alignment of the caspase cleavage sites in human substrates with their yeast homologues reveals considerable conservation (Figure 2A). Cluster analysis based on the overall similarity of cleavage sites reveals that the yeast substrates share highest similarity with human caspase 6 and 8 cleavage sites (Figure 2B), supporting earlier observations that caspase-like activities in yeast are caspase 6- and 8-associated.
Conclusions about death-related proteases and proteolysis
Proteolysis is not only a simple form of post-translational modification but an important regulatory process. By altering substrate abundance, proteases can elicit loss-of-function effects that directly influence the activation state of metabolic and developmental pathways. Furthermore, proteolytic cleavage of substrates can elicit gain-of-function changes, with neopeptide fragments showing novel properties. Examples linked to PCD (reviewed in ) include the cleavage of executioner procaspases 3 and 7 zymogens by initiator caspases, degradation of Bid by caspase 8 and granzyme B, and cleavage, destabilization and degradation of the retinoblastoma-associated protein relieving its block on S-phase progression. Cleavage of caspases 7 and 9 generates antagonistic neoepitopes that recruit IAPs, removing their inhibition of apoptosis, whereas cleavage of DFF (DNA fragmentation factor) 45/ICAD (inhibitor of caspase-activated deoxyribonuclease) allows the formation of a catalytically active DFF40/CAD (caspase-activated deoxyribonuclease) dimer that can degrade DNA during apoptosis. Loss-of-function effects may require degradation of a substantial proportion of a protein. Gain-of-function requires cleavage of only a small proportion to mediate a biological effect.
The identification of yeast protease substrates should promote further understanding of the role of individual proteases in basic biological processes, and may provide novel targets for drug development. Traditional post-genomic screens for antifungal drugs have often focused on essential genes; however, an alternative strategy might be to identify death-related substrates, which at reduced levels, promote the intrinsic fungal PCD response. The roles that substrate level loss-of-function and gain-of-function changes have on fungal PCD remain largely unexplored, and so should be a significant focus for the future.
Original research was supported by the Biotechnology and Biological Sciences Research Council [grant numbers BB/C501176/1 and CUF_Y3_11] and a Biotechnology and Biological Sciences Research Council studentship to D.W.
We acknowledge all the past, present and visiting members of the Ramsdale laboratory, thanking in particular Yahui Huang, Paul Airs, Emma Sheils and Virginia Cabezon for their contributions.
8th International Meeting on Yeast Apoptosis: An Independent Meeting held at Keynes College, University of Kent, Canterbury, U.K., 2–6 May 2011. Organized and Edited by Paula Ludovico (University of Minho, Braga, Portugal).
Abbreviations: DFF, DNA fragmentation factor; ER, endoplasmic reticulum; IAP, inhibitor of apoptosis protein; MS/MS, tandem MS; PCD, programmed cell death; ROS, reactive oxygen species; UBP10, ubiquitin-processing protease 10
- © The Authors Journal compilation © 2011 Biochemical Society