Yeast: An Experimental Organism for All Times
Yeast provides insights into the evolutionary origins of eukaryotes and the basis for human disease
Gerald R. Fink and Jef D. Boeke
"Nothing in biology makes sense except in the light of evolution."
-Theodosius Dobzhansky, 1973
As we stand on the threshold of the 21st century, we are confronted by the genomic sequence of Saccharomyces cerevisiae, the first eukaryote for which this information is known. This achievement has been hailed as the Rosetta Stone for understanding all other eukaryotes. What are the issues that can be addressed by knowing the sequence (and ultimately the function) of all 6,000 genes in this organism? A central question focuses on evolution: How did single-celled and multicellular nucleated organisms evolve?
At first glance, the yeast genome sequence offers few insights into these questions. Perhaps the best that can be said is that the repertoire of proteins inferred from the genome sequence establishes beyond a doubt that S. cerevisiae is a bone fide eukaryote. Not only does it have a nucleus and mitochondria, but it encodes the full complement of cytoskeletal elements such as actin, tubulin, and motor proteins that are thought to be restricted to the eukaryotes.
Yeast Permits an Experimental Test of Evolutionary Mechanisms
The ease with which S. cerevisiae can be manipulated means that questions about alternative evolutionary strategies can be addressed by direct experimental analysis as well as by traditional comparisons of DNA sequences among organisms. Sequence comparisons between organisms can suggest possible mechanisms by which evolution of one species from another might occur. But how can one decide whether such mechanisms exist? After all, conversion of one organism to another occurred over millions of years, leading many to assume that the process is refractory to experiment in laboratory time. As one biologist replied when asked about evolution in the Pleistocene, "I wasn't there, so how should I know?"
However, the strategies for genetic selections developed in S. cerevisiae coupled with the new science of genomics may actually enable a real-time test of a particular evolutionary scheme. A good example is the endosymbiont theory of the origin of mitochondria and chloroplasts. The hypothesis that these organelles arose from ancestral prokaryotic endosymbionts has received strong support by comparing the biochemistry and genomic sequences of these organelles with those of modern eukaryotes.
From such comparative analyses, perhaps the most surprising example of endosymbiosis occurs among the Apicomplexans, a group of protozoa that includes the malaria parasites. These organisms have both a mitochondrion and a plastid. The analysis of DNA sequence from these extrachromosomal genomes led to the remarkable discovery that the protozoan plastid genome is the remnant of an algal plastid genome. This organelle, surrounded by four membranes, is postulated to have been acquired through multiple endosymbiotic events-the primary symbiosis occurred when a eukaryote engulfed a photosynthetic prokaryote, and the second occurred when this alga was engulfed by the progenitor of malaria parasites. Finally, many of the genes from the endosymbiont were either lost or incorporated into the nuclear genome.
Putting the Endosymbiont Theory to Test
Can one put any of these inferences to an experimental test? The modern chloroplast and mitochondrion no longer resemble their putative ancestors-many organellar proteins are encoded by nuclear genes. Sequences homologous to mitochondrial and chloroplast genomes have been found in the nucleus, and sequences from the chloroplast genome have been found in the mitochondrial genomes of plants and protozoa. If these nuclear genes that encode mitochondrial proteins were once confined to an endosymbiont, how did they get out of the mitochondrion and enter the nucleus?
This question could be answered experimentally in S. cerevisiae because it is possible to set up a selection to reveal any movement of DNA from the mitochondrion to the nucleus. The procedure involves first transforming a nuclear gene, URA3, into the mitochondrion of a Ura- yeast strain. As the nuclear URA3 gene cannot be expressed when it is located in the mitochondrion, it fails to complement a defect caused by the cognate ura3 mutation in the nucleus. However, at a rate greater than l0-5/cell/generation, DNA escapes from the mitochondrion into the nucleus, where it complements the mutant defect, making the cells Ura+.
The extraordinary frequency with which DNA moves from the mitochondrion to the nucleus shows that this exodus continues in real time. By contrast, no reverse transfer from nucleus to the mitochondrion has been observed in S. cerevisiae, even in an assay that could have detected one event in 1010. These experiments give new meaning to the term "symbiont," as the symbiotic relationship involves cooperation between the participants in the transfer of genetic information-the endosymbiont contributes new genes, and the host contributes the enzymes that integrate those genes into its genome. This yeast experiment is important because it shows decisively that extant organisms have a mechanism that can account for the postulated evolutionary event.
S. cerevisiae Provides a Clue to the Origins of the Cytoskeleton
The ability to manipulate S. cerevisiae may also help to explain the rapid evolution of cytoskeletal proteins. In its simplest form, one of the hypotheses for eukaryote evolution posits that the nucleus originated from a protoeukaryote that was the common ancestor of the archaea and eukaryotes. A descendant of this organism endocytosed a eubacterium, which then became a mitochondrion. This model is attractive because archaea have histones as well as transcription and replication systems akin to those found in modern eukaryotes. However, the archaea do not have the complement of cytoskeletal proteins found in eukaryotes. So, how did these molecules arise?
Two proteins found in prokaryotes, both eubacteria and archaea, have functional and structural resemblances to two ubiquitous cytoskeletal proteins, actin and tubulin. The three-dimensional structure of FtsA, an ATP-binding heat shock protein, is strikingly similar to that of actin (also an ATP-binding protein); that of FtsZ, a protein required for bacterial cell division, is remarkably similar to that of tubulin (like FtsZ, a GTP-binding protein).
Despite their similarity in shape, the bacterial and eukaryotic proteins share little amino acid sequence homology: the homology between actin and FtsA is 20%; that between tubulins and FtsZ is 10-15%. Consideration of the date of divergence between eukaryotes and bacteria has led to the conclusion that there was insufficient time for the number of mutational changes required for the bacterial proteins to morph into their eukaryotic form. This puzzle has led one observer to conclude that the rate of change of these proteins altered dramatically at a key point in their history.
Accounting for Differential Evolutionary Rates
Perhaps retrotransposition was the dramatic mutational event in the history of ftsA and ftsZ evolution. Retrotransposition accelerates the mutation rate of genes because information flow from DNA->RNA->DNA is extremely error-prone, potentially peppering retrotransposed genes in one stroke with multiple changes in sequence.
Retrotransposition is often thought of as a destructive process, one that produces defective pseudogenes. However, it could also be creative-in this case, by churning out copies of ftsA and ftsZ that differ radically in DNA sequence from their bacterial progenitors. These newly minted copies could have provided the burst of novelty required to accelerate the rate of evolutionary change, with those retro- transposed copies that give an advantage being saved and those that do not being jettisoned.
Direct experiments on retrotransposition in S. cerevisiae show that the reverse transcriptase of the Ty element is capable of generating novel copies of genes, using the cDNAs of cellular messages as templates. These endogenously generated cDNAs carrying new mutations are subsequently recombined into the genome. This retrotransposition model can be extended to explain the rapid evolution that led to actin and tubulin. This model posits that the ftsA and ftsZ genes of the primitive eukaryotic nucleus were reverse transcribed, and the new versions, reinvented by retrotransposition, were returned to the nuclear genome.
There are hints that this scenario has validity. For example, ftsA and ftsZ are cotranscribed in an operon in eubacteria; meanwhile, the actin and tubulin genes are extremely closely linked in S. cerevisiae. This close linkage might be expected, if the ftsA/ftsZ message were the progenitor of the actin/tubulin cDNA in the retrotransposition event that jump-started the evolution of the latter two genes.
The possibility that retrotransposition is a sufficiently powerful mutagen to have facilitated the evolution of ftsZ to b-tubulin could be tested directly in S. cerevisiae. One could construct a yeast strain containing a conditional mutation in TUB2, the yeast b-tubulin gene. The strain would also be engineered to have a copy of the bacterial ftsZ gene inserted in the Ty element so that ftsZ is cotranscribed as part of the Ty retrotransposon mRNA (Ty::ftsZ).
Each time Ty-ftsZ transposed via its cDNA, copies of the ftsZ gene would be altered at multiple sites by mutation. This process could be repeated sequentially on each new ftsZ* gene. After each retrotransposon jump, the strain would be checked for the possibility that one of the newly created ftsZ cDNAs had been mutated to a ftsZ that could function as a b-tubulin.
How Did Single Cells Become Multicellular Organisms?
Somehow the evolution of "eukaryoteness" enabled single-celled organisms to form multicellular entities. In this sense, the eukaryotic microbes provide the central vital link between the prokaryotes and the more complex creatures visible to the naked eye. Some proto-multicellular eukaryotewas primed to make the leap(s) to multicellularity.
What was this critical adaptation, that enabled the acquisition of multicellular complexity? Was it the versatility of the eukaryotic cytoskeleton, the separation between nuclear CPU and cytoplasmic input-output devices, the explosion of intra- and eventually intercellular communication and signaling pathways, the acquisition of organellar energy factories, the enablement of sex, diploidy, and meiosis, or the ability to support a much more complex genome? The theories put forward to grapple with this heady question will shortly be overwhelmed with new data as the genomic sequences of the archetypal protoeukaryotes, including Giardia and Trichomonas species, are catalogued.
The origins of multicellularity are unknown. However, once eukaryotes arose, several distinct versions of multicellularity blossomed, including plants, animals, and fungi. It is easy to forget that yeast is a specialized unicellular form of what is, in nature, a multicellular organism. Although the multicellular pseudohyphal form of S. cerevisiae may not seem very organized, other fungi, such as mushrooms, make easily recognizable, highly developed complex structures.
Completing DNA sequence determinations for the genomes of Arabidopsis thaliana, Caenorhabditis elegans, of several fungi besides S. cerevisiae, and of higher plants and animals could help to take this key issue out of the realm of speculation. We may learn whether there are three distinct gene sets shared by fungi, plants, and animals, but not by the protoeukaryotes, that reflect three independent routes to multicellularity.
How far can one go by pushing S. cerevisiae in the opposite evolutionary direction-essentially forcing it to become a better model for much more complex pathways and creatures? How "human" can we make a lowly yeast cell, and what can we expect from forcing them into this role?
The dramatic accumulation of DNA sequences has once and for all allayed all residual skepticism about S. cerevisiae being a eukaryote. As humans, we are consumed with a desire to learn biology for practical reasons, such as the understanding of biological processes that go awry and make us sick. Can the surprisingly close evolutionary connection between humans and S. cerevisiae be better used to understand some of these disease-related processes? The use of S. cerevisiae to model these functions and dysfunctions is genomically justified, but is it possible to reconstruct the pathology of a complex, multicellular organism in a single yeast cell?
S. cerevisiae's many virtues for studying human disease have already attracted the attention of the medical community. First, the yeast genome sequence is of enormous value as a reference; it provides a complete eukaryotic sequence set, serving as both a functional and a structural model for human genes important in disease. Second, cross-referencing of S. cerevisiae and other model organism genes with their mammalian homologs has allowed more rapid and complete identification of candidate disease genes.
Third, the function of human disease genes with yeast homologs can be rapidly analyzed using transcomplementation of the corresponding yeast mutant as a surrogate assay for their human function. Fourth, S. cerevisiae itself can sometimes be "humanized" in cases where a homolog to a human component is absent or drastically different; new human functions can be added to this eukaryotic tabula rasa one by one. In short, the existing and developing technologies for functional analysis of S. cerevisiae and its genome put it at the forefront of all eukaryotes for analytical efficiency and thoroughness.
The application of these yeast technologies has already provided unexpected insights into aging, cancer, death, and mad cow disease (bovine spongiform encephalopathy)-human scourges that were thought not even to exist in S. cerevisiae. Although they do not exist as yeast illnesses, each is the reflection of some basic biological phenomenon that is probably common to all eukaryotes. The lesson seems to be: whenever a disease has a cellular basis, one can look to S. cerevisiae as a beacon for the mechanism.
The many dramatic breakthroughs in understanding the relationship of cancer to cell division are a direct legacy of genetic screens designed to identify how yeast cells proceed through their cell cycle. These include how the cell cycle is driven, regulated, and kept in check when things go wrong. The protein kinase circuitry and its logic, DNA synthesis and repair, the mitotic machinery, and the elegant checkpoint control systems that protect the integrity of the family jewels were first modeled and dissected in S. cerevisiae. Our current deep functional understanding of these processes goes along with that reached through studying sequence homologies, which reveal that the basic blueprints of cell division are conserved.
Yeast Genomics Provides Insights into Human Disease
The identification of the genes responsible for two inborn cancer susceptibility syndromes, hereditary nonpolyposis colorectal cancer (HNPCC) and ataxia telangiectasia (AT), with the corresponding yeast genes drove home the point that yeast can help us understand these processes in sickness as in health. HNPCC is associated with mutations in a set of DNA mismatch repair genes. In both yeast and human mutants, there is heightened instability in simple sequence repeats (microsatellite instability) as a consequence of a general defect in postrepli- cation mismatch repair. AT is caused by mutations in a gene encoding a large kinase that is homologous to the yeast gene MEC1, which plays central roles in DNA damage checkpoint control.
These are but two examples of a rich and growing legacy of human cancer genes with fungal homologs in which S. cerevisiae has played a key role in deciphering the etiology of the disease. Initially, yeast studies allowed the identification of the affected pathway but, once that was accomplished, the detailed knowledge of the biology of S. cerevisiae provided open avenues for investigating the human disorder. Not only were there ready platforms for testing of mechanistic models to understand the process, but the yeast also provided a high-throughput experimental system to run inexpensive in vivo tests for new chemotherapeutic agents.
Some central genes that protect us against cancer, such as the TP53 tumor suppressor gene, have no obvious yeast homolog, but nevertheless can carry out their basic biochemical function in yeast cells. Yeast cells can be "humanized" to respond to a protein the cells normally do not express. The transcriptional activation of the p53 protein activates a wide variety of yeast reporter genes. Selection for dominant-negative mutants in the human p53 gene in yeast cells led to the isolation of mutations that closely resemble those commonly isolated in a wide variety of human cancers.
Recently, human genes involved in the process of apoptosis, a process previously thought unique to multicellular organisms, were grafted into S. cerevisiae, disclosing a previously undescribed form of yeast cell death with remarkable similarities to apoptosis. The integral involvement of mitochondrial integrity in this process is conserved between yeast and mammalian systems.
A molecule implicated in a syndrome of premature human aging, the WRN1 gene, and a similar molecule involved in yet another cancer predisposition syndrome, BLM1, have a homolog in S. cerevisiae called SGS1. Studies in S. cerevisiae indicate that sgs1 mutants are hyper-rec (that is, they display elevated rates of recombination) for both homology-dependent and illegitimate recombination. Ribosomal DNA is present in the form of large tandem arrays in nearly all eukaryotes and is maintained in that form through the collaborative efforts of a number of different proteins.
Remarkably, one of the consequences of the sgs1 hyper-rec phenotype is an increased rate of formation of nonchromosomal circular forms of the ribosomal DNA; accumulation of these circles in yeast mother cells leads to their premature senescence. Perhaps it is a superficial, fortuitous coincidence that cellular senescence in S. cerevisiae should be controlled by a molecule related to one required for long life span in humans. But, given the deluge of successes, this connection between aging and ribosomal recombination in S. cerevisiae may provide a new lead into senescence mechanisms.
Even studies of the exotic and evanescent microbes responsible for spongiform neuropathies such as mad cow disease have received an important input from yeast studies. The long-controversial notion of prions, infectious proteins defined as directing a self-propagating conformational state, received a very potent boost when previously enigmatic cytoplasmically inherited agents of yeast were elegantly reinterpreted as prions. There is no sequence homology between the proteins involved in the spongiform encephalopathies and the cytoplasmically inherited yeast [URE3] and [PSI] elements, but the eccentric behavior of these yeast elements is consistent with the prion hypothesis. Moreover, several lines of experimental evidence suggest that the prion phenomenon transcends the limited universe of a previously obscure corner of animal virology and may in fact be a more general and underappreciated biological phenomenon.
Sophisticated Tools Promise Global Insights
All of these noble goals, from wonder about evolutionary origins to hard-nosed forced evolution of disease models, can now be pursued with an increasingly sophisticated yeast tool kit. The completion of the genome sequence has led to gene chips or assays bearing probes corresponding to all yeast open reading frames (ORFs) that have many interesting uses, such as transcript profiling and analysis of genome diversity between strains.
Moreover, collections of strains in which each individual gene is knocked out are being built and can be used for genome-wide functional analyses. The yeast two-hybrid system and its many variants are being exploited to identify the networks of macromolecular interactions that drive biology and can be applied on a genome-wide scale. Genome-wide transposon insertion libraries are analyzed before and after application of a genetic selection to identify the necessary genes. Other transposon insertion libraries, based on elements carrying promoterless reporter genes, are used to capture gene expression and protein localization information.
Increasingly, powerful computers and software are harnessing and trying to interpret the relentless torrent of data generated by techniques new and old. These yeast technologies are emerging at a dizzying pace; in 50 years the transcript profile of today will seem as quaint as a replica-plating experiment.
The completion of the yeast genome sequence is only the first step along the path that is revolutionizing biology. Obviously, with the addition of each new organism's genomic sequence data, the yeast system becomes more valuable. New techniques will emerge, allowing scientists to experiment in S. cerevisiae with the genomes of these organisms. As these new techniques develop, yeast will enable us to look forward to the solution of key issues in human biology and backwards into the evolutionary origins of eukaryotes.
Brachmann, R. K., M. Vidal, and J. D. Boeke. 1996. Dominant-negative p53 mutations selected in yeast hit cancer hotspots. Proc. Natl. Acad. Sci. USA 93:4091-4095.
Derr, L. K., J. N. Strathern, and D. J. Garfinkel. 1991. RNA-mediated recombination in S. cerevisiae. Cell 67:355-364.
Doolittle, R. F. 1995. The origins and evolution of eukaryotic proteins. Phil. Trans. R. Soc. Lond. B. 349:234-240.
Doolittle, W. F. 1998. You are what you eat: a gene transfer ratchet could account for bacterial genes in eukaryotic nuclear genomes. Trends Genet. 14:307-311.
Fishel, R., M. K. Lescoe, M. R. Rao, N. G. Copeland, N. A. Jenkins, I. Garber, M. Kane, and R. Kolodner. 1993. The human mutator gene homolog MSH2 and its association with hereditary nonpolyposis colon cancer. Cell 75:1027-1038. (Erratum, 77:167, 1994.)
Gilson, P. R., U.-G. Maier, and G. I. McFadden. 1997. Size isn't everything: lessons in genetics miniaturisation from nucleomorphs. Genomes Evol. 7:800-805.
Lowe, I., and L. A. Amos. 1998. Crystal structure of the bacterial cell-division protein FtsZ. Nature 391:203-206.
Morrow, D. M., D. A. Tagle, Y. Shiloh, F. S. Collins, and P. Hieter. 1995. TEL1, an S. cerevisiae homolog of the human gene mutated in ataxia telangiectasia, is functionally related to the yeast checkpoint gene MEC1. Cell 82:831-840.
Nogales, E. S., S. G. Wolf, and K. H. Downing. 1998. Structure of the alpha1 beta tubulin dimer by electron crystallography. Nature 391:199-203.
Nugent, J. M., and J. D. Palmer. 1991. RNA-mediated transfer of the gene coxII from the mitochondrion to the nucleus during flowering plant evolution. Cell 66:473-481.
Sinclair, D. A., and L. Guarente. 1997. Extrachromosomal rDNA circles-a cause of aging in yeast. Cell 91:1033-1042.
Thorsness, P. E., and E. R. Weber. 1996. Escape and migration of nucleic acid between chloroplasts, mitochondria, and the nucleus. Int. Rev. Cytol. 165:207-234.
Wickner, R. B. 1994. [URE3] as an altered URE2 protein: evidence for a prion analog in Saccharomyces cerevisiae [see comments]. Science 264:566-569.
Wilson, R. J. M., and D. H. Williamson. 1997. Extrachromosomal DNA in the apicomplexa. Microbiol. Mol. Biol. Rev. 61:1-16.