R E V I E W S
`
`Yeast evolutionary genomics
`
`Bernard Dujon
`
`Abstract | Over the past few years, genome sequences have become available from
`an increasing range of yeast species, which has led to notable advances in our
`understanding of evolutionary mechanisms in eukaryotes. Yeasts offer us a unique
`opportunity to examine how molecular and reproductive mechanisms combine to
`affect genome architectures and drive evolutionary changes over a broad range of
`species. This Review summarizes recent progress in understanding the molecular
`mechanisms — such as gene duplication, mutation and acquisition of novel genetic
`material — that underlie yeast evolutionary genomics. I also discuss how results from
`yeasts can be extended to other eukaryotes.
`
`Yeasts offer unique advantages for evolutionary genomic
`studies among eukaryotic organisms. These uni-
`cellular fungi, which are characterized by their ability
`for unlimited clonal propagation by budding or fission
`and the absence of fruiting bodies1, are easily amenable
`to microbial genetic techniques, and the limited size and
`compactness of their genomes facilitate the characteriza-
`tion of naturally or artificially evolved populations using
`sequencing. The comparison of yeast genomes, which
`is generally applied to infer evolutionary changes, can
`be complemented by experimental analyses to elucidate
`the underlying molecular mechanisms. More than a
`thousand different yeast species have now been iden-
`tified, and many more are likely to exist. Contrary to
`the commonly held view, yeasts do not represent primi-
`tive unicellular eukaryotes but instead have repeatedly
`emerged from distinct phylogenetic lineages of ‘modern’
`fungi1,2,3. The molecular evolutionary mechanisms iden-
`tified in yeasts are therefore expected to have equiva-
`lents with those in other fungi and in all animals, given
`the common evolutionary origin of both kingdoms
`among Opisthokonta4.
`The complete genome sequence of the baker’s yeast
`
`Saccharomyces cerevisiae was a milestone of early genome -
`
`ics in the 1990s that directed subsequent studies5. The
`abundance of novel genes stimulated the development
`of systematic functional genomic approaches. Today,
`S. cerevisiae offers an unparalleled reference source for
`studying basic molecular mechanisms of eukaryotic
`cells, as more than 80% of its ~5,780 protein-coding
`genes have been functionally characterized6. The second
`yeast genome to be completely sequenced was that of
`Schizosaccharomyces pombe7, but this yeast is only dis-
`tantly related to S. cerevisiae and its genetic architecture
`
`is very different, so comparing the two species did not
`offer many interpretable observations in terms of evo-
`lutionary genomics. Sequences of other yeasts at more
`evenly distributed evolutionary intervals were needed.
`About 40 different yeast species have been sequenced
`so far (FIG. 1), and genomic-level aspects of yeast evo-
`lution are gradually being unveiled. The most atten-
`tion has been focused on the Saccharomycotina (or
`Hemiascomycetes, the large subphylum of Ascomycota
`to which S. cerevisiae belongs), and the evolution of their
`genomic architecture is now reasonably well understood
`(BOX 1). Genomic data on other phylogenetic lineages
`are presently too scarce to provide an equivalent degree
`of understanding, but these lineages nevertheless offer
`interesting outgroups for comparisons.
`A major, and unexpected, lesson from yeast genomics
`is the extensive sequence divergence observed between
`different lineages. This divergence goes right down to
`the species level and reflects intense genomic changes
`that contrast with the conservation of biological prop-
`erties of yeasts for very long evolutionary times. Rather
`than offering a continuous range of gradual evolutionary
`adaptations, as proposed by classical Darwinian theory,
`genomes from distinct yeast clades, or even from species
`of the same genus, differ from one another in an abrupt
`manner. This is consistent with yeast genomes being the
`remnants of repeated bottleneck events that occurred in
`essentially clonal populations. The stochastic drift result-
`ing from such a mode of propagation is important as it
`offers the possibility for non-optimized variants to sur-
`vive and eventually colonize novel niches to which they
`may be better adapted. The recurrent emergence of this
`unicellular mode of life — with its specific reproductive
`properties — during fungal evolution is one of the main
`
`Clade
`A group of taxa that forms a
`monophyletic unit. It is
`applicable to any level of
`the taxonomical hierarchy.
`
`Institut Pasteur, Unité de
`Génétique Moléculaire des
`Levures, Centre National de
`la Recherche Scientifique/
`University Pierre and Marie
`Curie-Paris, 25 rue du
`Docteur Roux, F75724
`Paris-CEDEX 15, France.
`e-mail: bernard.dujon@
`pasteur.fr
`doi:10.1038/nrg2811
`
`512 | JULY 2010 | VOLUME 11
`
` www.nature.com/reviews/genetics
`
`
`
`© 20 Macmillan Publishers Limited. All rights reserved10
`
`LCY Biotechnology Holding, Inc.
`Ex. 1035
`Page 1 of 13
`
`

`

`R E V I E W S
`
`Status
`
`Ploidy
`
`Chr. no.
`
`Gen. size
`
`GC (%)
`
`CDS (no.)
`
`Split (%)
`
`12.1
`12.2
`12.6
`11.4
`10.2
`
`11.4
`12.3
`14.7
`9.8
`10.4
`10.7
`11.3
`10.7
`
`8.7
`
`9.5
`
`12.2
`15.4
`
`16
`16
`16
`16
`16
`
`91
`
`3
`13
`
`78886
`
`7
`
`4–9
`6
`
`781
`
`4
`
`888888
`
`n
`2n
`2n
`
`2n
`2n
`2n
`
`nnn2
`
`n
`
`2n
`n
`
`n
`
`n?
`2n
`
`nn2
`
`n
`
`nn2
`
`n
`2n
`2n
`2n
`2n
`
`C
`
`S
`
`SSSEESCSCCDCCE
`
`C
`
`E
`D
`
`Saccharomyces cerevisiae*
`Saccharomyces paradoxus*
`Saccharomyces mikatae
`Saccharomyces kudriavzevii
`Saccharomyces bayanus*
`Saccharomyces exiguus (K. exigua)
`Saccharomyces servazzii
`Saccharomyces castellii
`Candida glabrata
`Kluyveromyces polysporus (V. polyspora)
`Zygosaccharomyces rouxii
`Kluyveromyces thermotolerans (L. thermot.)
`Kluyveromyces waltii (L. waltii)
`Saccharomyces kluyveri (L. kluyveri)
`Kluyveromyces lactis
`Kluyveromyces marxianus var. marxianus
`Ashbya gossypii (E. gossypii)
`
`Strain
`S288c
`CBS432T
`IFO1815
`IFO1802
`CBS7001
`CBS379T
`CBS4311T
`CBS4309T
`CBS138T
`DSMZ70294T
`CBS732T
`CBS6340T
`NCYC2644
`CBS3082T
`CBS2359
`CBS712T
`ATCC10895
`
`Dekkera bruxellensis
`Hansenula polymorpha (O. polymorpha)
`
`CBS2499
`CBS4732T
`
`Debaryomyces hansenii
`Pichia stipitis*
`Pichia sorbitophila
`Pichia guilliermondii
`Clavispora lusitaniae
`Candida parapsilosis
`Lodderomyces elongisporus
`Candida tropicalis*
`Candida albicans*
`Candida dubliniensis
`
`Whole-genome
`duplication
`
`No complex I in
`mitochondrial DNA,
`point centromeres,
`three mating
`cassettes
`
`5S RNA
`
`CUG(cid:1)Ser
`
`
`
`SaccharomycotinaSaccharomycotina
`
`Saccharomyces
`sensu stricto
`
`Kazachstania
`
`Naumovozyma
`Nakaseomyces
`Vanderwaltozyma
`Zygosaccharomyces
`
`Lachancea
`
`Kluyveromyces
`
`Eremothecium
`
`Dekkera
`Ogataca
`
`Debaryomyces
`
`Clavispora
`
`‘Candida’ ||
`
`‘Candida’ ||
`
`Saccharomycetaceae
`
`CTG group
`
`Refs
`5
`142
`142,143
`143
`142–144
`144
`144
`143
`48
`69
`44
`44
`63
`44
`48
`144
`62
`
`38.3
`
`5,769
`
`4.4
`
`2.5
`–
`1.3
`5.6
`
`–6
`
`.0
`3.4
`
`4.5
`
`4,700
`5,204
`5,652
`5,055
`5,137
`5,230
`5,397
`5,108
`
`38.8
`32.2
`39.1
`47.3
`43.8
`41.5
`38.8
`
`52.0
`
`4,715
`
`47.9
`
`5,933
`
`–
`
`110
`145
`
`6.5
`
`– –––––6
`
`.0
`
`36.3
`41.1
`
`6,397
`5,841
`
`Dipodascaceae
`
`10.6
`12.1
`13.1
`15.5
`14.6
`14.3
`14.6
`
`43.8
`44.5
`38.7
`37.0
`33.1
`33.5
`33.2
`
`5,920
`5,941
`5,733
`5,802
`6,258
`6,107
`5,758
`
`9.4
`
`41.1
`
`5,313
`
`–
`
`–
`
`48
`146
`144
`88
`88
`88
`88
`88
`88,147
`98
`
`148,149
`
`20.5
`
`49.0
`
`6,582
`
`14.5
`
`48
`
`7
`
`###
`
`#
`150
`
`151
`151
`
`45.9
`
`–––
`
`36.0
`37.5
`43.7
`37.7
`
`4,969
`4,907
`4,814
`5,057
`
`12.5
`11.2
`11.3
`11.5
`
`– 2
`
`7.0
`
`19.0
`
`48.6
`
`6,572
`
`8.9
`
`52.0
`
`4,285
`
`8
`
`4
`
`4 6
`
`3
`
`14
`
`8
`
`n?
`
`n
`
`n
`
`n
`
`n
`
`CCEDDDDDDC
`
`P
`
`D C
`
`CDDD
`
`D
`C
`
`S
`E
`
`CBS767T
`CBS6054
`CBS7064
`ATCC6260
`ATCC42720
`CDC317
`CBS2605T
`MYA3404
`SC5314
`CD36
`
`GS115
`
`CBS7504
`
`972h–
`
`yFS275
`OY26
`
`Arxula adeninivorans
`
`Pichia pastoris* (K. pastoris)
`
`Yarrowia lipolytica
`
`Fungi with fruiting bodies
`
`Schizosaccharomyces pombe
`Schizosaccharomyces octosporus
`Schizosaccharomyces japonicus
`Schizosaccharomyces jsp.
`
`Fungi with fruiting bodies plus yeasts
`
`Cryptococcus neoformans var. grubii
`Cryptococcus neoformans var. neoformans*
`
`H99
`JEC21
`
`Malassezia globosa
`Malassezia restricta
`
`CBS7966
`CBS7877
`
`Loss of
`introns,
`budding
`
`Ascomycota
`Ascomycota
`
`Basidiomycota
`Basidiomycota
`
`Trichomonascus
`
`Komagataella
`
`Yarrowia
`
`Pezizomycotina
`
`Taphrinomycotina
`
`Pucciniomycotina
`
`Agaricomycotina
`
`Ustilagomycotina
`
`Figure 1 | Overview of the sequenced yeast genomes. The figure
`summarizes the list of sequenced yeast species with their original
`designation at the time of sequence publication (new taxonomies are
`between brackets). Coloured triangles represent clades or genera with
`their most recent designation shown on the left. The tree topology is
`adapted from REFS 137–139 and T. Boekhout and C.P. Kurtzman (personal
`communication). The branch lengths are arbitrary. Dotted lines illustrate
`uncertainty and/or incongruence between different published
`phylogenies. Deep-branch separations are very ancient140,141. Genomic
`architectures identify three major groups in Saccharomycotina (BOX 1),
`highlighted by coloured backgrounds (blue, Saccharomycetaceae
`family; orange, ‘CTG group’; purple, Dipodascaceae and related
`families). The arrows point to major evolutionary events, as deduced
`from common genomic architecture in Saccharomycotina using a
`parsimony hypothesis. For each species, the sequenced the strain is
`indicated (T, type strain) with the status of sequence completion
`(C, complete with finishing; D, draft assembly (limited number of
`
`supercontigs attributable to chromosomes); E, exploratory genome
`survey; P, work in progress; S, whole-genome shotgun (typically ~3–7(cid:114)
`coverage and/or ~500–2,000 contigs)). The ploidy is also shown along
`with, for fully analysed sequences, the number of chromosomes (haploid
`set), the genome size (in megabases, excluding ribosomal DNA), the
`GC base content (GC), the total number of predicted protein-coding
`genes (the coding sequence (CDS)) and the percentage of the CDS
`that contains introns (the ‘split’; ‘–’ indicates that data are not available
`in computed or published form). *Species for which several strains
`have been sequenced. Only the first or the more complete
`sequence is indicated. ‡Heterogeneous composition. §Hybrid, partially
`homozygotized (Génolevures Consortium, personal communication).
`||Candida is commonly used to designate clades containing human
`pathogens, although this nomenclature only designates asexual yeasts,
`irrespective of their actual phylogeny (for example, Candida glabrata).
`¶Unpublished data from the Génolevures website. #Unpublished data
`from the Broad Institute Fungal Genome Initiative website.
`
`NATURE REVIEWS | GENETICS
`
` VOLUME 11 | JULY 2010 | 513
`
`
`
`© 20 Macmillan Publishers Limited. All rights reserved10
`
`LCY Biotechnology Holding, Inc.
`Ex. 1035
`Page 2 of 13
`
`

`

`R E V I E W S
`
`Loss of heterozygosity
`Formally, the loss of one active
`allele in a heterozygous pair.
`This loss can occur by any
`mechanism (mutation, deletion
`or gene conversion using the
`other allele as template). Loss
`of heterozygosity in yeast
`genomes corresponds to
`large-scale chromosomal
`regions encompassing multiple
`neighbouring genes.
`
`Horizontal gene transfer
`A process by which an
`organism incorporates
`genetic material from another
`organism that does not belong
`to its line of ancestry. This
`process is also called lateral
`gene transfer.
`
`Allopatric
`Refers to organisms,
`populations or species
`that inhabit distinct
`geographical regions.
`
`Box 1 | Summary of major architectural features of Saccharomycotina genomes
`
`Genomes of the budding yeasts range in size from ~9 to 20 megabases (for the haploid set) and contain a limited
`number of protein-coding genes (~4,700–6,500). They have few spliceosomal introns (~2–15% of split genes) and a
`variable number of tRNA genes (~160–510). Complete sets of genes for other non-coding RNAs occur within
`budding yeast genomes, in addition to limited numbers of mobile elements belonging to various families (mostly
`class I). Traces of RNAi machinery are generally absent, except in specific cases123. The presence of autonomous
`plasmids or viral elements is highly variable. All yeast genomes contain a large number of paralogous gene
`copies, which are highly diverged in their sequences and represent various types of ancestral duplications.
`Against this common background, genomes of each of the three major ‘families’ that have been most extensively
`studied — the Saccharomycetaceae family, the ‘CTG group’ and the Dipodascaceae and related families (FIG. 1) —
`show distinctive features whose evolutionary origin is partially understood. Dipodascaceae and related yeasts
`have limited numbers of chromosomes (4–6) and dispersed 5S RNA genes. CTG yeasts and ‘protoploid’
`Saccharomycetaceae have larger numbers of chromosomes (6–8) and ‘duplicated’ Saccharomycetaceae have
`twice as many (13–16). CTG yeasts are characterized by a common deviation of the genetic code (the CUG codon
`encodes serine) with corresponding changes of numerous codons124, large centromeres and a unique mating-type
`locus (with two allelic forms). Saccharomycetaceae yeasts are characterized by point centromeres (which are
`highly conserved) and triplicate mating-type cassettes that ensure the simultaneous presence of both mating-type
`alleles in haploid cells (with some exceptions). A whole-genome duplication has occurred in this family, creating a
`subset of clades that have shorter chromosomes bearing the traces of the duplication followed by numerous gene
`deletions. Among them is the extensively studied baker’s yeast Saccharomyces cerevisiae, which serves as the
`general reference organism for yeast genomics.
`
`interests for evolutionary scientists studying yeasts, as
`is the conservation of this mode over long evolution-
`ary periods. Now that population structures of yeasts
`can be readily studied at the genomic level, this opens
`a new perspective for evolutionary genomics. The flow
`of genetic material within and between populations
`or even species can now be precisely measured and its
`consequences carefully examined.
`In this Review, I first discuss how the specificities of
`yeast reproduction influence genetic exchanges, propa-
`gation and population structures, and consequently
`the definition of species. I then examine the molecular
`mechanisms underlying evolutionary changes, covering
`aspects of gene duplications, mutations, chromosomal
`rearrangements, loss of heterozygosity (LOH), horizontal
`gene transfer and de novo gene formation, and try to
`place them in the broader perspective of other eukaryo-
`tic genomes. Other aspects of evolutionary significance,
`such as the conservation of gene families, functional
`repertoires, gene-expression variation, phenotypic
`robustness against mutational changes, the roles of epi-
`somal and mobile elements, expansions of satellite and
`other repeated sequences, and various features of RNA
`machineries, will not be comprehensively covered.
`
`Yeast populations and species
`Populations of yeasts are very different from those of
`multicellular organisms, in which sexual reproduction
`is obligatory. Under favourable conditions, all yeasts canUU
`
`propagate indefinitely by mitotic divisions (BOX 2), form-
`ing large clonal populations in either the haploid or dip-
`loid phase, or even in polyploids (although in this case
`various types of chromosomal rearrangements or loss
`can be observed). Note that, given their polyphyletic
`origin (FIG. 1), this ability for unlimited clonal growth
`in unicellular form constitutes one of the best criteria
`for differentiating yeasts from other fungi, next to the
`fact that their sexual states do not form within or on
`fruiting bodies1. The possibility of generating large
`
`subpopulations that can propagate without regular
`genetic exchange is obviously important for yeast evolu-
`tion. Many species, however, conserve the ability to per-
`form complete sexual cycles, but use it in various ways
`that affect (and often limit) the rates of genetic exchange
`in populations (BOX 3). Other yeasts are only known as
`asexual species, although some might undergo rare
`mating events. Polyploids and heterospecific hybrids
`are also encountered among yeast isolates, raising the
`difficult question of the definition of yeast species.
`
`Population structures. The analysis of sequence polymor-
`phisms at selected sets of loci suggested that, in nature,
`genetic exchanges and recombination are limited for both
`Saccharomyces paradoxus8 and S. cerevisiae9. Genomic
`data precisely quantified the level of genetic exchanges
`in these populations. The distribution of segments of
`shared genealogy among three strains of S. cerevisiae
`revealed only 314 outcrossing events during ~16 mil-
`lion successive cell divisions, indicating a frequency of
`outcrossing as low as 2 (cid:114) 10–5 per asexual generation10.
`A similar figure was reported for S. paradoxus, in which
`one sexual cycle was observed every 1,000 asexual gen-
`erations and only 1% of these cycles corresponded to
`outcrossing11. With such low levels of outcrossing, the
`recombination of advantageous or deleterious alleles
`present in populations is limited, and subpopulations
`tend to form with independent accumulation of sequence
`variations, even in sexual species. Allopatric divergence
`accentuates the phenomenon12,13. From recent genome
`analysis, it was found that natural subpopulations of the
`wild yeast S. paradoxus remain well delineated within
`geographic boundaries14. Distinct subpopulations
`were also identified for S. cerevisiae, although with a
`higher incidence of mosaics between subpopulations,
`which was attributed to human domestication14,15.
`A similarly limited gene flow between yeast subpopu-
`lations was recently observed by sequencing all 18
`known strains of Saccharomyces kudriavzevii. This led
`
`514 | JULY 2010 | VOLUME 11
`
` www.nature.com/reviews/genetics
`
`
`
`© 20 Macmillan Publishers Limited. All rights reserved10
`
`LCY Biotechnology Holding, Inc.
`Ex. 1035
`Page 3 of 13
`
`

`

`Bateson–Dobzhansky–
`Muller effect
`A negative effect of allelic
`reassortment by genetic
`recombination in hybrids
`between members of distinct
`populations. By extension, it is
`a lethal effect of reassortment
`by genetic recombination in
`crosses between parents that
`exhibit differential gene loss
`after genome duplication.
`
`Differential gene loss
`The loss of opposite copies in a
`pair of ohnologues between
`two cells that inherited the
`ohnologues from a common
`genome duplication.
`
`Sympatric
`Refers to organisms,
`populations or species
`inhabiting the same
`geographic area.
`
`Ascospores
`The four cellular products of a
`meiosis. The four ascospores
`are embedded in a sac
`called an ascus (observed
`in Ascomycota).
`
`Protoploid
`A general term
`created to designate all
`Saccharomycetaceae yeasts
`that do not originate from
`whole-genome duplication.
`
`to the discovery of a ‘balanced unlinked gene network
`polymorphism’, in which in each strain, several genes
`involved in galactose utilization are all simultaneously
`either active (Portuguese strains) or pseudogenized
`(Japanese strains), despite their unlinked locations on
`multiple chromosomes16.
`
`Hybridization. Heterospecific hybridizations are not
`rare in yeasts. Successful interspecific hybridizations
`were experimentally obtained several years ago by mass
`mating between different Saccharomyces speciess
`
`17. Strains
`of the lager beer yeast Saccharomyces pastorianus are
`known from recent genomic studies to be distinct hybrids
`between S. cerevisiae and Saccharomyces bayanus18,19.
`Other beer strains are hybrids between S. cerevisiae and
`S. kudriavzevii 20,21. These hybridization events may be
`accelerated by the stressful conditions imposed during
`alcoholic fermentations22, but the formation of heterospe-
`cific hybrids is not limited to Saccharomyces sensu stricto
`yeasts. The asexual diploid pathogenic yeasts Candida
`albicans and Candida dubliniensis form tetraploid
`hybrids23, and natural hybrids have also been reported
`between the Basidiomycota yeasts Cryptococcus neoform-
`ans and Cryptococcus gattii24. Similarly, two subgenomes
`were found in a wild isolate of Zygosaccharomyces rouxii25
`and in Pichia sorbitophila (Génolevures Consortium,
`personal communication). Despite the frequent occur-
`rence of hybrids, the role of heterospecific hybrid-
`ization in the evolution of yeasts remains unclear. In
`Saccharomyces sensu stricto hybrids, the two paren-
`tal genomes often undergo non-reciprocal exchanges
`accompanied by loss of genes, chromosomal segments
`or even complete chromosomes, producing various
`chimaeras from which novel lineages might emerge26,27.
`With its megabase-long chromosomal fragment of dis-
`tinct composition, Lachancea kluyveri might represent
`such a case28, but additional examples are needed. The
`contradictory phylogenetic relationships of some yeasts29
`may be partly due to such hybridization phenomena.
`
`Box 2 | Asymmetrical cellular division and evolutionary consequences
`
`Although clonal yeast populations seem to increase exponentially in size under
`appropriate conditions, closer examination indicates a fundamentally asymmetrical
`mode of cellular division. This mode is immediately perceptible for the large group of
`Saccharomycotina (budding yeasts) but is also true for other groups126,127. In budding
`yeasts, mitoses produce a small cell, called a ‘bud’, that separates from a larger cell,
`called the ‘mother’. Although both receive equivalent sets of genes, the bud is
`‘younger’ than the mother because newly synthesized molecules tend to migrate to
`it, whereas ancient molecules tend to remain in the mother cell128. Clonal yeast
`populations are therefore composites of young cells, which can generate an
`unlimited succession of subsequent generations, and a series of gradually older
`cells that eventually stop dividing (FIG. 2). Asymmetrical cell division has two
`consequences for yeast genome evolution. First, equivalent mutational changes or
`genomic alterations will have different probabilities of propagation in populations
`depending on which cells they occur in. If this difference is small in Saccharomyces
`cerevisiae, in which mother cells can undergo a relatively large number of successive
`mitoses, it will increase if the proliferative ability of mother cells diminishes. Second,
`in haploid clones, young and old cells differ in their ability to undergo sexual cycles.
`In S. cerevisiae, at low mating-pheromone levels, the new bud is oriented towards the
`weak signal, increasing its probability of outcrossing129,130. By contrast, old cells
`undergo mating-type switching (BOX 3), which favours intraclonal mating.
`
`R E V I E W S
`
`Speciation. Given their above properties, the definition
`of yeast species becomes a complex question, as is often
`the case in organisms in which clonal propagation dom-
`inates sexual genetic exchanges. In the Saccharomyces
`sensu stricto clade (FIG. 1), in which this question has
`been most thoroughly addressed, the reduced meiotic
`fertility of heterospecific hybrids fulfils the most classi-
`cal criterion used to define species30. However, several
`mechanisms combine to create this post-zygotic bar-
`rier, not all of which are applicable to other yeast clades.
`One mechanism is activation of the mismatch repair
`system by sequence divergence between two parental
`genomes31. Another mechanism is the consequence of
`chromosomal translocations32–34. A case of incompat-
`ibility was also reported between S. cerevisiae mito-
`chondria and a nuclear gene of S. bayanus35. However,
`lineage-specific gene losses36 are expected to have the
`largest effect in terms of lowering the meiotic fertility
`of hybrids between Saccharomyces sensu stricto species
`and other yeasts that have inherited the same genome
`duplication (see below). This corresponds to a special
`case of the Bateson–Dobzhansky–Muller effect37, an
`interspecific genome incompatibility effect that is not
`expected to occur in other phylogenetic yeast lineages.
`One differential gene loss between two lineages results in
`half of the meiotic products of their hybrids inheriting
`an incomplete gene set owing to chromosomal reassort-
`ment. Combination of several differential losses between
`the two parental genomes further lowers the probability
`of forming meiotic products with a complete gene set.
`Comparisons with other phylogenetic branches of yeasts
`are, unfortunately, not presently possible as there is a
`lack of similar experimental data on the meiotic fertility
`of hybrids, a topic that needs further investigation. Pre-
`zygotic barriers also contribute to speciation in yeasts,
`allowing the sympatric existence of distinct species that
`are otherwise able to mate38. Germinating ascospores
`of S. cerevisiae and S. paradoxus show a preference for
`own-species mating over interspecific mating39, whereas
`vegetative haploid cells of S. paradoxus do not40. The
`molecular basis for this phenomenon remains to be
`elucidated, but such a preference may have an impor-
`tant role in wild yeast populations, as mating in these
`species is suspected to occur primarily between newly
`germinated ascospores.
`
`Gene duplication mechanisms
`Building a minimal yeast genome, with one optimized
`gene per function, may sound like an attractive idea to
`synthetic biology engineers today, but nature has never
`done that. Instead, as postulated by Ohno 40 years ago41,
`all genomes show numerous traces of gene duplications,
`and the evolutionary consequences of these duplications
`in creating novel gene functions have attracted consid-
`erable attention42,43. This is also true for yeast genomes
`that contain numerous series of paralogous genes from
`ancient or recent duplications. In S. cerevisiae and related
`species (FIG. 1), some paralogous gene pairs originate
`from complete genome duplication, but many dispersed
`paralogues and tandem gene arrays also exist. In the
`‘protoploid’ Saccharomycetaceae species44, the last two
`
`NATURE REVIEWS | GENETICS
`
` VOLUME 11 | JULY 2010 | 515
`
`
`
`© 20 Macmillan Publishers Limited. All rights reserved10
`
`LCY Biotechnology Holding, Inc.
`Ex. 1035
`Page 4 of 13
`
`

`

`R E V I E W S
`
`Generation 1
`
`Generation 2
`
`Generation 3
`
`1000
`
`Ageing cell line,
`mating-type
`switching
`
`100
`
`1001
`
`etc.
`
`1
`
`10
`
`101
`
`etc.
`
`11
`
`Immortal cell line,
`no mating-type
`switching
`
`110
`
`111
`
`Homozygous
`diploids
`
`1100
`
`1101
`
`etc.
`
`etc.
`
`1110
`
`etc.
`
`1111
`
`10000
`
`10001
`
`11110
`
`11111
`
`etc.
`
`etc.
`etc.
`Figure 2 | Asymmetrical cell divisions and consequences for population structures. Mitotic divisions in
`Saccharomycotina are asymmetrical irrespective of the cellular ploidy. In both haploid and diploid cells, the parental
`cell produces a small bud, designated as the ‘daughter cell’ (circles with narrow outlines), and a larger sibling,
`designated as the ‘mother’ cell (circles with bold outlines), which keeps the original size and shape of the parental cell.
`After increasing in size, the buds reproduce the same phenomenon, producing a bud and becoming mother cells (for
`example, the cell marked 110). The bud lineages (right arrows) undergo an unlimited number of successive mitoses,
`whereas mother cells can either undergo mitosis (for example, cell 10) — producing both a new bud that will begin an
`immortal lineage (for example, cell 101) and a one-generation-older ‘mother’ cell (for example, cell 100) — or stop
`dividing (for example, cell 10000). In Saccharomyces cerevisiae, mother cells generally undergo 20–30 successive
`mitoses, producing the impression of a nearly exponential growth of the population in which, in any generation,
`one-half of the cells are newly formed buds (labelled 1xx1 in generation 3, in which x means 0 or 1), one-quarter are
`first-time ‘mothers’ (1x10), one-eighth are second-time ‘mothers’ (1100), and so on. This asymmetry points to similarities
`between yeasts and filamentous fungi in the sense that the immortal bud lineages are equivalent to the tips of hyphae,
`but the residual number of cellular divisions of mother lineages is specific to yeasts. Note that in Saccharomycetaceae
`species undergoing mating-type switching (symbolized by green and purple colours) by action of the HO endonuclease
`(BOX 3), immortal bud lineages never switch. Mating between members of the same descent (double-headed arrow)
`produces homozygous diploids, except for the mating locus itself.
`
`phenomena occur so frequently that one-third of their
`protein-coding genes are members of paralogous fami-
`lies despite the absence of an ancestral whole-genome
`duplication. The numbers of paralogues are even higher
`in Yarrowia lipolytica and in species of the ‘CTG group’.
`Four distinct mechanisms contribute to this universal
`genome redundancy in yeast genomes, and it would be
`interesting to better quantify their relative rates of action
`in future studies.
`
`Expansions of tandem gene arrays. Clusters of identical
`or similar protein-coding genes exist in various eukaryo-
`tic genomes, but their role in evolution remains largely
`speculative. In yeasts, tandem gene arrays are generally
`not conserved45 except in specific cases, such as B-type
`cyclin genes. Several examples indicate the role of tan-
`dem gene array expansions in rapid adaptive evolution.
`In S. cerevisiae, tandem expansion of the CUP1 locus
`occurs when cells are grown under selective pressure
`for copper resistance46. Because this cluster is made of
`identical gene copies47, its expansion (or shrinkage) can
`be accounted for by unequal homologous recombination
`
`events. However, in other cases tandem arrays are
`generally made of sequence-diverged gene copies, which is
`more consistent with functional diversification than with
`the simple need for copy-number increases, although the
`degree of possible concerted evolution in such tandems
`remains to be analysed. In the pathogenic yeast Candida
`glabrata, two large gene clusters are observed that are
`unique to this species48. One of them corresponds to the
`expansion of six additional YPS genes (compared with
`two YPS genes in S. cerevisiae), which encode extracellular
`glycosylphosphatidylinositol (GPI)-linked aspartyl pro-
`teases that are required for virulence49. Expansion of this
`cluster corresponds with the roles of these enzymes in
`processing a GPI-linked cell-wall adhesin that is nec-
`essary for the adherence of C. glabrata to mammalian
`cells. The other cluster is an expansion of the unique
`S. cerevisiae MNT3 gene into eight copies in C. glabrata.
`This gene encodes an α-1,3-mannosyltransferase
`involved in cell wall biogenesis. Again, the sequence
`divergence between the tandem copies is consistent
`with functional diversification of the encoded proteins,
`and the cluster varies in size among clinical isolates of
`
`CTG group
`This term is used here to
`designate a monophyletic
`group of yeast species that
`share a common genomic
`architecture and a common
`deviation from the universal
`genetic code (the CUG codon
`specifies serine) but are
`taxonomically classified in
`diverse families, some of which
`contain yeasts that do not share
`these genomic properties.
`
`516 | JULY 2010 | VOLUME 11
`
` www.nature.com/reviews/genetics
`
`
`
`© 20 Macmillan Publishers Limited. All rights reserved10
`
`LCY Biotechnology Holding, Inc.
`Ex. 1035
`Page 5 of 13
`
`

`

`Box 3 | Variation of sexual cycles and evolutionary consequences
`
`Typically, mating in yeasts starts by the fusion of haploid cells that are similar in size
`and shape (isogamy), which produces zygotes131. Fusion involving diploid cells is also
`possible, and this forms polyploids132. Mating between haploid cells is usually
`followed by karyogamy, which produces a cell with a single diploid nucleus that
`either gives rise to a diploid clone or enters meiosis. Karyogamy is sometimes
`delayed (a process that can be enhanced by mutational alteration in Saccharomyces
`cerevisiae) such that dikaryotic cells may undergo mitosis before nuclear fusion. In
`such a case, the progeny may receive only one of the two nuclei that cohabited for
`some time in the same cell; this situation offers an opportunity for the two parents
`to exchange mitochondrial DNA (cytoduction) and various other autonomous
`genetic elements, such as plasmids, viruses or retrotransposons, without
`exchanging the bulk of their chromosomes. Depending on the species, mating can
`occur between genetically identical cells (homothallism, which is rare in yeasts) or
`requires two cells of distinct mating-types (heterothallism). In Ascomycota yeasts,
`
`mating-type is controlled by a unique chromosomal locus (designated MAT or T MTL)
`that exists in two idiomorphic forms, usually designated a and alpha. This locus
`contains a limited set of genes that control the expression of numerous other genes
`by various molecular mechanisms that have been extensively elucidated in
`S. cerevisiae but that are beyond t