Abstract

Among genes conserved from bacteria to mammals are those involved in replicating and repairing DNA. Following the complete sequencing of four hemiascomycetous yeast species during the course of the Génolevures 2 project, we have studied the conservation of 106 genes involved in replication, repair, and recombination in Candida glabrata, Kluyveromyces lactis, Debaryomyces hansenii, and Yarrowia lipolytica and compared them with their Saccharomyces cerevisiae orthologues. We found that proteins belonging to the replication fork and to the nucleotide excision repair pathway were—on the average—more conserved than proteins involved in the checkpoint response to DNA damage or in meiotic recombination. The meiotic recombination proteins Spo11p and Mre11p-Rad50p, involved in making meiotic double-strand breaks (DSBs), are conserved as is Mus81p, involved in resolving meiotic recombination intermediates. Interestingly, genes found in organisms in which DSB-repair is required for proper synapsis during meiosis are also found in C. glabrata, K. lactis, and D. hansenii but not in Y. lipolytica, suggesting that two modes of meiotic recombination have been selected during evolution of the hemiascomycetous yeasts. In addition, we found that SGS1 and TOP1, respectively, a DEAD/DEAH helicase and a type I topoisomerase, are duplicated in C. glabrata and that SRS2, a helicase involved in homologous recombination, is tandemly duplicated in K. lactis. Phylogenetic analyses show that the duplicated SGS1 gene evolved faster than the original gene, probably leading to a specialization of function of the duplicated copy.

Introduction

A common concern of all living organisms is how to replicate, maintain, and transfer to the next generation an intact pool of chromosomes. For that reason, they have developed a number of partly redundant machineries in order to ensure proper duplication and repair of their genome content. Proteins involved in these machineries are most of the time conserved during evolution from bacteria to mammals (Cann and Ishino 1999; Lindahl and Wood 1999; Pâques and Haber 1999; Zhou and Elledge 2000; Burgers et al. 2001). However, constraints are clearly different between organisms with small and compact genomes and those with large genomes containing numerous repetitive elements. Constraints are also different between unicellular organisms with short generation times and metazoans. Finally, constraints are different between organisms going through a mode of sexual reproduction followed by meiosis and those whose reproduction mode is asexual. During the course of the Génolevures 2 project (Dujon et al. 2004), four hemiascomycetous yeast species were fully sequenced. Candida glabrata is a pathogenic yeast, the second causative agent of human candidiasis, phylogenetically related to Saccharomyces cerevisiae (Bennett, Izumikawa, and Marr 2004). Kluyveromyces lactis is also related to S. cerevisiae and has been used for genetic studies or industrial applications like the production of β-galactosidase (Bolotin-Fukuhara et al. 2000). Debaryomyces hansenii is a marine yeast that can tolerate high salinity levels, phylogenetically close to the pathogenic Candida albicans (Lépingle et al. 2000). Yarrowia lipolytica is a more distantly related yeast, able to grow as individual yeast cells or as a mycelium (Casarégola et al. 2000). The evolutive distance between S. cerevisiae and Y. lipolytica, measured as the amino acid divergence between proteins, is comparable to the entire phylum of Chordates (Dujon et al. 2004). However, genome sizes and general organization are comparable among the five hemiascomycetes sequenced. Hence, constraints on DNA replication and repair should be similar, and any difference detected should reflect a mechanistic difference between such machineries. In the present work, we have analyzed the genomes of the four newly sequenced hemiascomycetous yeasts to look for the presence of 106 genes known to be involved in replication, repair, and recombination in S. cerevisiae. We found that some machineries are very well conserved whereas others have diverged more rapidly. In addition, two genes (SGS1 and TOP1) are duplicated in C. glabrata and one (SRS2) is duplicated in K. lactis.

Materials and Methods

Analysis of Gene Families

We started from protein families built from sequence similarities during Génolevures 2 (Dujon et al. 2004) (http://cbi.labri.fr/Genolevures). When a family contained only five members and one gene per sequenced species, we considered that this gene was the correct orthologue of the S. cerevisiae gene. This happened in 41 cases out of 106. In the three cases of larger gene families (the RFC, RPA, and MCM families, containing altogether 14 genes), orthologues could not be chosen among paralogues based on sequence similarity but were determined based on synteny conservation. No homologue to a S. cerevisiae gene was found in any species in only three cases. Finally, in the remaining 48 cases, more than one gene matched with the S. cerevisiae gene in at least one sequenced species. In these cases, we used three different criteria to select the correct putative orthologue. First, we performed global alignments using the Needleman-Wunsch algorithm and rejected all alignments with less than 20% identity, unless a portion of the protein showed a very strong similarity to the S. cerevisiae protein. Second, we looked for synteny conservation between the S. cerevisiae gene and the corresponding gene in a region covering 10 genes upstream and 10 genes downstream. Synteny was considered conserved if at least three genes (including the query) were conserved in the correct order. Evidence of synteny was found in 171 cases out of 240 (63 cases out of 64 for C. glabrata, 54 out of 60 for K. lactis, 38 out of 65 for D. hansenii, and 16 out of 51 for Y. lipolytica). In six cases, synteny conservation was not found with S. cerevisiae but with at least another species. Third, a possible conserved motif was searched using the Conserved Domain Database (CDD) (http://www.ncbi.nlm.nih.gov:80/structure/cdd/cdd.shtml). If none of these three approaches gave a significant result, the corresponding gene was discarded. In the three cases of species-specific gene duplication (SGS1, TOP1, and SRS2), multiple alignments of the homologues in the five species were performed using ClustalW. We also performed tBlastn searches using the orthologue sequence in the closest species as the query sequence. This approach detected five novel orthologues not found by previous methods. In addition, we used PSI-Blast (on the National Center for Biotechnology Information [NCBI] server, with E value threshold = 0.1, restricted to fungal genomes) to look for S. cerevisiae genes absent in C. glabrata and K. lactis, but we did not find any sequence similar to them. Finally, in 10 specific cases (RFA3, RAD28, LIF1, LIF2/NEJ1, MEI4, MER1, REC104, SAE3, SPO13, and TAM1/NDJ1), when no orthologue was found by the above methods in one or more species, we tried to detect degenerate homologues using the possible synteny conservation with S. cerevisiae. For example, RFA3 (YJL173c) is located between YJL172w and YJL174w in S. cerevisiae. We examined the DNA sequence between these two genes in D. hansenii in order to find a possible degenerate orthologue. When a gene was found at the expected position, we tried to align its sequence with the S. cerevisiae gene. Using this method, we found a possible orthologue in only one case, TAM1/NDJ1 in K. lactis. All the results are summarized in table 1 and the supplementary table.

Table 1

Conservation of Replication, Repair, and Recombination Genes of Saccharomyces cerevisiae in the Four Completely Sequenced Hemiascomycetous Yeasts


Genea
 

ORF
 

Human Homologue
 

Functionb
 

Cg
 

Kl
 

Dh
 

Yl
 
Replication and DNA topology        
CDC9 YDL164c Lig I Ligase 
DNA2 YHR164c Dna2 Nuclease 1c 
MCM2 YBL023c Mcm2 Helicase 
MCM3 YEL032w Mcm3 Helicase 
MCM4 YPR019w Mcm4 Helicase 
MCM5 YLR274w Mcm5 Helicase 
MCM6 YGL201c Mcm6 Helicase 
MCM7 YBR202w Mcm7 Helicase 
POL1 YNL102w Pol α Polymerase 
POL12 YBL035c Pol α (subunit) Polymerase 
POL2 YNL262w Pol ε Polymerase 
POL3 YDL102w Pol δ Polymerase 
POL30 YBR088c PCNA Clamp 
POL5 YEL055c — Polymerase 
PRI1 YIR008c p48 Polymerase 
PRI2 YKL045w p58 Polymerase 
RAD27 YKL113c Fen1 Nuclease 
RFA1 YAR007c RPA1 SSB 
RFA2 (i) YNL312w RPA2 SSB 
RFA3 YJL173c RPA3 SSB d  
RFC1 YOR217w Rfc1 Clamp loader 
RFC2 YJR068w Rfc2 Clamp loader 
RFC3 YNL290w Rfc3 Clamp loader 
RFC4 YOL094c Rfc4 Clamp loader 
RFC5 YBR087w Rfc5 Clamp loader 
TOP1 YOL006c Topo I Topoisomerase 
TOP2 YNL088w Topo II Topoisomerase 
TOP3 YLR234w Topo III Topoisomerase 
SGS1 YMR190c e Helicase 
NER        
RAD1 YPL022w ERCC4 (XPF) Nuclease 
RAD10 YML095c ERCC1 Nuclease 
RAD14 (i) YMR201c XPA DNA binding 
RAD16 YBR114w — DNA binding 1f 
RAD7 YJR052w — DNA binding 
RAD2 YGR258c ERCC5 (XPG) Helicase 
RAD23 YEL037c HR23A, HR23B DNA binding 
RAD4 YER162c XPC DNA binding 
RAD25 YIL143c ERCC3 (XPB) Helicase 
RAD3 YER171w ERCC2 (XPD) Helicase 
Repair        
RAD26 YJR035w ERCC6 (CSB) TCR 
RAD28 YDR030c CSA TCR d  
REV3 YPL167c Pol ζ Polymerase 
RAD6 YGL058w HR6B Ubiq.-conjug. 1(2i) 
RAD18 YCR066w Rad18 DNA binding 
RAD5 YLR032w —  
POL4 YCR014c Pol λ Polymerase  
MMR        
MLH1 YMR167w Mlh1 DNA binding 
PMS1 YNL082w Pms2 DNA binding 1f 
MLH2 YLR035c Pms1    
MLH3 YPL164c   
MSH1 YHR120w  
MSH2 YOL090w Msh2 DNA binding 1f 
MSH3 YCR092c Msh3 DNA binding 
MSH6 YDR097c GTBP DNA binding 
NHEJ        
DNL4 YOR005c Lig IV Ligase 
LIF1 YGL090w Xrcc4  d   
LIF2/NEJ1 YLR265c —  d   
HDF1 YMR284w Ku70 DNA binding 
HDF2 YMR106c Ku80 DNA binding 
MRE11 YMR224c Mre11 Nuclease 
RAD50 YNL250w Rad50 DNA binding 1f 
XRS2 YDR369c Nbs1    
HIR        
EXO1 YOR033c — Nuclease 
RAD51 YER095w g Strand exch. 
RAD52 YML032c h DNA binding 
RAD54 YGL163c Rad54  1f 
RAD55 YDR076w g    
RAD57 YDR004w g  
RAD59 YDL059c —    
RDH54/TID1 YBR073w Rad54B  
SRS2 YJL092w — Helicase 
Meiotic recombination        
DMC1 (i) YER179w Dmc1 Strand exch.  
HOP1 YIL072w — Synap. comp.  
HOP2 (i) YGL033w GT198   
MEI4 YER044c-a —  d d  
MEI5 YPL121c —   
MEK1 YOR351c Chk2 Kinase 1(i)  
MER1 YNL210w — Splicing d   
MER2 (i) YJR021c —  1(i)   
MER3 (i) YGL251c — Helicase  
MRE2/NAM8 YHR086w PABP-1 Splicing 
MSH4 YFL003c — DNA binding   
MSH5 YDL154w — DNA binding   
MUS81 YDR386w Mus81 Nuclease 
MMS4 YBR098w — Nuclease  
REC102 YLR329w —    
REC104 YHR157w —  d d   
REC114 (i) YMR133w —  1(i)   
RED1 YLR263w —    
SAE2 YGL175c —   
SAE3 YHR079c-b —  d d d  
SPO11 YHL022c Spo11 Topoisomerase 
SPO13 YHR014w —  d d d  
TAM1/NDJ1 YOL104c — DNA binding   
ZIP1 YDR285w — Synap. comp. 1c  
ZIP2 YGL249w —    
Checkpoints        
CHK1 YBR274w Chk1 Kinase  
RAD24 YER173w Rad17 Clamp loader 
DDC1 YPL194w Rad9 Clamp  
MEC3 YLR288c Hus1 Clamp  
RAD17 YOR368w Rad1 Clamp 
DDC2 YDR499w ATRIP DNA binding  
MEC1 YBR136w ATR Kinase 
RAD53 YPL153c Chk2 Kinase 
RAD9 YDR217c —    
TEL1
 
YBL088c
 
ATM
 
Kinase
 
1
 
1
 
1
 
1
 

Genea
 

ORF
 

Human Homologue
 

Functionb
 

Cg
 

Kl
 

Dh
 

Yl
 
Replication and DNA topology        
CDC9 YDL164c Lig I Ligase 
DNA2 YHR164c Dna2 Nuclease 1c 
MCM2 YBL023c Mcm2 Helicase 
MCM3 YEL032w Mcm3 Helicase 
MCM4 YPR019w Mcm4 Helicase 
MCM5 YLR274w Mcm5 Helicase 
MCM6 YGL201c Mcm6 Helicase 
MCM7 YBR202w Mcm7 Helicase 
POL1 YNL102w Pol α Polymerase 
POL12 YBL035c Pol α (subunit) Polymerase 
POL2 YNL262w Pol ε Polymerase 
POL3 YDL102w Pol δ Polymerase 
POL30 YBR088c PCNA Clamp 
POL5 YEL055c — Polymerase 
PRI1 YIR008c p48 Polymerase 
PRI2 YKL045w p58 Polymerase 
RAD27 YKL113c Fen1 Nuclease 
RFA1 YAR007c RPA1 SSB 
RFA2 (i) YNL312w RPA2 SSB 
RFA3 YJL173c RPA3 SSB d  
RFC1 YOR217w Rfc1 Clamp loader 
RFC2 YJR068w Rfc2 Clamp loader 
RFC3 YNL290w Rfc3 Clamp loader 
RFC4 YOL094c Rfc4 Clamp loader 
RFC5 YBR087w Rfc5 Clamp loader 
TOP1 YOL006c Topo I Topoisomerase 
TOP2 YNL088w Topo II Topoisomerase 
TOP3 YLR234w Topo III Topoisomerase 
SGS1 YMR190c e Helicase 
NER        
RAD1 YPL022w ERCC4 (XPF) Nuclease 
RAD10 YML095c ERCC1 Nuclease 
RAD14 (i) YMR201c XPA DNA binding 
RAD16 YBR114w — DNA binding 1f 
RAD7 YJR052w — DNA binding 
RAD2 YGR258c ERCC5 (XPG) Helicase 
RAD23 YEL037c HR23A, HR23B DNA binding 
RAD4 YER162c XPC DNA binding 
RAD25 YIL143c ERCC3 (XPB) Helicase 
RAD3 YER171w ERCC2 (XPD) Helicase 
Repair        
RAD26 YJR035w ERCC6 (CSB) TCR 
RAD28 YDR030c CSA TCR d  
REV3 YPL167c Pol ζ Polymerase 
RAD6 YGL058w HR6B Ubiq.-conjug. 1(2i) 
RAD18 YCR066w Rad18 DNA binding 
RAD5 YLR032w —  
POL4 YCR014c Pol λ Polymerase  
MMR        
MLH1 YMR167w Mlh1 DNA binding 
PMS1 YNL082w Pms2 DNA binding 1f 
MLH2 YLR035c Pms1    
MLH3 YPL164c   
MSH1 YHR120w  
MSH2 YOL090w Msh2 DNA binding 1f 
MSH3 YCR092c Msh3 DNA binding 
MSH6 YDR097c GTBP DNA binding 
NHEJ        
DNL4 YOR005c Lig IV Ligase 
LIF1 YGL090w Xrcc4  d   
LIF2/NEJ1 YLR265c —  d   
HDF1 YMR284w Ku70 DNA binding 
HDF2 YMR106c Ku80 DNA binding 
MRE11 YMR224c Mre11 Nuclease 
RAD50 YNL250w Rad50 DNA binding 1f 
XRS2 YDR369c Nbs1    
HIR        
EXO1 YOR033c — Nuclease 
RAD51 YER095w g Strand exch. 
RAD52 YML032c h DNA binding 
RAD54 YGL163c Rad54  1f 
RAD55 YDR076w g    
RAD57 YDR004w g  
RAD59 YDL059c —    
RDH54/TID1 YBR073w Rad54B  
SRS2 YJL092w — Helicase 
Meiotic recombination        
DMC1 (i) YER179w Dmc1 Strand exch.  
HOP1 YIL072w — Synap. comp.  
HOP2 (i) YGL033w GT198   
MEI4 YER044c-a —  d d  
MEI5 YPL121c —   
MEK1 YOR351c Chk2 Kinase 1(i)  
MER1 YNL210w — Splicing d   
MER2 (i) YJR021c —  1(i)   
MER3 (i) YGL251c — Helicase  
MRE2/NAM8 YHR086w PABP-1 Splicing 
MSH4 YFL003c — DNA binding   
MSH5 YDL154w — DNA binding   
MUS81 YDR386w Mus81 Nuclease 
MMS4 YBR098w — Nuclease  
REC102 YLR329w —    
REC104 YHR157w —  d d   
REC114 (i) YMR133w —  1(i)   
RED1 YLR263w —    
SAE2 YGL175c —   
SAE3 YHR079c-b —  d d d  
SPO11 YHL022c Spo11 Topoisomerase 
SPO13 YHR014w —  d d d  
TAM1/NDJ1 YOL104c — DNA binding   
ZIP1 YDR285w — Synap. comp. 1c  
ZIP2 YGL249w —    
Checkpoints        
CHK1 YBR274w Chk1 Kinase  
RAD24 YER173w Rad17 Clamp loader 
DDC1 YPL194w Rad9 Clamp  
MEC3 YLR288c Hus1 Clamp  
RAD17 YOR368w Rad1 Clamp 
DDC2 YDR499w ATRIP DNA binding  
MEC1 YBR136w ATR Kinase 
RAD53 YPL153c Chk2 Kinase 
RAD9 YDR217c —    
TEL1
 
YBL088c
 
ATM
 
Kinase
 
1
 
1
 
1
 
1
 

NOTE.—(i): intron(s). RAD6 contains two predicted introns.

a

Number of orthologues to Saccharomyces cerevisiae genes (left column) for each yeast species are listed in the last four columns. Genes are classified according to the pathway they belong to (see supplementary table for gene names in the four new hemiascomycetes). Essential genes are underlined.

b

Functions are indicated when known. Clamp and Clamp loader refer to the processivity factor (PCNA) and its associated loading complex (RFC) during S-phase replication or to the alternate complexes associated to checkpoint activity (Ddc1p-Mec3p-Rad17p and RFC-Rad24p). Strand exch.: Rad51p and Dmc1p exhibit RecA-like strand exchange activity during homologous recombination. Synap. comp.: protein is part of the meiotic synaptonemal complex. SSB: single-strand DNA-binding complex. TCR: protein involved in transcription-coupled repair. Ubiq.-conjug.: ubiquitin-conjugating enzyme (E2). Precise catalytic activities of nucleases, helicases, topoisomerases, polymerases, kinases, and DNA-binding proteins were not detailed to save space.

c

Probable pseudogene or intron (high-quality sequence containing in-frame stop codons or frameshifts).

d

Cases in which the synteny with S. cerevisiae was used to try to find a putative orthologue (see Materials and Methods).

e

Blm, Wrn, RecQ4, RecQL, RecQ5.

f

Possible pseudogene or intron (low-quality sequence containing in-frame stop codons or frameshifts).

g

Rad51B, Rad51C, Rad51D, Xrcc2, Xrcc3.

h

Rad52, Rad52B.

Calculation of the Mean Conservation of S. cerevisiae Gene Products and Their Orthologues in the Four Other Genomes

A predicted gene from a given genome (Cagl, Klla, Deha, or Yali) was considered orthologous to a S. cerevisiae gene if it was found in a region of conserved synteny between the two genomes. Using this criterion, 3,935 orthologues were found with C. glabrata, 3,440 with K. lactis, 279 with D. hansenii, and 107 with Y. lipolytica. The percentage of amino acid identity between a given S. cerevisiae gene and its corresponding orthologue was obtained from the Smith-Waterman alignment between the two sequences. Means were calculated using these values. For construction of synteny maps and precise parameters used for alignments see Dujon et al. (2004).

Phylogenetic Analyses

Multiple alignments of amino acid sequences were performed using T-coffee (Notredame, Higgins, and Heringa 2000). Gaps and poorly aligned sequences were excluded from alignments using Gblocks (Castresana 2000). Tree reconstruction was performed by the maximum likelihood algorithm as implemented in PHYML (Guindon and Gascuel 2003). The substitution process was modeled by the Jones, Taylor, and Thornton (JTT) model, the heterogeneity of substitution rates among sites was modeled by a gamma distribution, with four categories and a parameter estimated from the data set. Tree topology and support of internal branches were inferred by 500 bootstrap calculations. Calculations of the nonsynonymous/synonymous substitution rate ratio (ω = Dn/Ds) were performed with the maximum likelihood method (Goldman and Yang 1994) implemented in the PAML package version 3.14 (Yang 1997).

Results

In order to study the evolution of DNA replication, repair and recombination pathways among five completely sequenced yeast genomes, we studied the conservation of 106 different S. cerevisiae genes, selected on their known function as deduced from genetics or biochemistry. Out of 106 genes, 101 have orthologues in C. glabrata, 100 in K. lactis, 85 in D. hansenii, and only 70 in Y. lipolytica (table 1 and supplementary table). The only five genes not detected in C. glabrata are involved in meiotic recombination and nonhomologous end joining (NHEJ). In Y. lipolytica, one or more gene is not detected in each pathway, except in the nucleotide excision repair (NER) pathway. When a gene is not found, it means either that the gene is not present in the considered organism or that the sequence is too diverged to be recognized using our criteria (see Materials and Methods). Out of 29 S. cerevisiae essential genes, only three (DNA2, RFA3, and DDC2) are not perfectly conserved in all four hemiascomycetes studied (table 1). This suggests that most of the essential genes in S. cerevisiae also encode products that are essential (or at least important enough to be selected for) in the four other species. Seven genes in our list of 106 S. cerevisiae genes are split by an intron. Only two of these contain an intron in at least one of the four other hemiascomycetes. Similarly, five introns are predicted by the sequence in the four newly sequenced hemiascomycetes, but none of them is found in the same gene in at least two species. This suggests that introns are differentially lost and acquired during evolution, in accordance with a former study on 13 partially sequenced hemiascomycetous yeasts (Bon et al. 2003).

High Conservation of Genes Involved in S-Phase Replication

It is not surprising that almost all the proteins playing a role in chromosome replication (Burgers et al. 2001) are conserved throughout the hemiascomycete evolution. One notable exception however, is Rfa3p, one of the tripartite components of the yeast single-strand binding protein complex, which is not detected in D. hansenii and Y. lipolytica, the two other members (Rfa1p and Rfa2p) of the same heterotrimeric complex being found. No RFA3 gene relic was found in these two species. In addition, the gene encoding Dna2p, involved in processing Okazaki fragments, contains three in-frame stop codons in Y. lipolytica and thus is most probably not properly translated. Dna2p function is at least partly redundant with Rad27p, suggesting that Rad27p is necessary and sufficient to process Okazaki fragments in Y. lipolytica, or that the translated N-terminal part of the Dna2 protein is sufficient to carry out its essential function.

TOP1 and SGS1 are specifically duplicated in C. glabrata (table 1). The two copies of C. glabrata Top1p (CAGL0E02431g and CAGL0J11660g, supplementary table) are almost perfectly aligned with ScTop1p, except in the N-terminal part of the protein. Synteny shows that CAGL0E02431g is the correct orthologue, the other copy being present in a duplicated chromosomal block present in both S. cerevisiae and C. glabrata (G. Fischer and B. Dujon, unpublished data). The duplicated copy was conserved in C. glabrata but not in S. cerevisiae in which no trace of a pseudogene or a relic could be found in the duplicated block (I. Lafontaine and B. Dujon, unpublished data). Consistent with that, the phylogenetic tree shows that CAGL0E02431g is the closest homologue of ScTop1 (fig. 1A). Calculation of synonymous (Ds) and nonsynonymous (Dn) substitutions show that Ds values are very high (Ds > 5). Because synonymous sites are saturated, Dn/Ds ratios are not a reliable measure of evolutionary rates. Hence, we took in consideration only Dn values. They are low and similar for both paralogues (DnCAGL0E02431g = 0.22; DnCAGL0J11660g = 0.18, as compared to S. cerevisiae). This suggests that both genes have evolved at a similar rate and have both probably retained their catalytic activity. Two copies of Sgs1p were found in C. glabrata (CAGL0L00407g and CAGL0H00759g). According to synteny results, CAGL0L00407g is the correct orthologue, and CAGL0H00759g is found in a duplicated block in S. cerevisiae and C. glabrata only (G. Fischer and B. Dujon, unpublished data). Like previously, the copy in the duplicated block has been erased, and no trace of a pseudogene or relic can be detected in S. cerevisiae. (I. Lafontaine and B. Dujon, unpublished data) The phylogenetic tree shows that the closest homologue of ScSgs1p is CAGL0L00407g, the other copy being more diverged (fig. 1B). Interestingly, the duplicated copy is shorter than the orthologue. It is deleted for the N-terminal part containing the Top3-binding domain and the C-terminal part containing the DNA-binding domain of the Sgs1 protein (fig. 2). Again, synonymous sites are saturated, but Dn values are low. However, the Dn value of the duplicated copy is higher than that of the orthologue (DnCAGL0L00407g = 0.2; DnCAGL0H00759g = 0.46, as compared to S. cerevisiae), meaning that not only the copy lost two important parts of the protein (still retaining the helicase motif) but also the remaining part diverged more rapidly.

FIG. 1.—

Phylogenetic tree of TOP1, SGS1, and SRS2 families. Trees were obtained by the maximum likelihood method. Branch lengths are estimated under the model of amino acids substitution JTT (Jones, Taylor, and Thronton 1992). Percentages of bootstrap values for internal nodes are indicated on each branch (see Materials and Methods for details). When the number of homologous sequences among the Génolevures species was less than 10, additional homologous sequences were retrieved among the other available fungal genomes. (A) phylogeny of TOP1. (B) Phylogeny of SGS1. (C) Phylogeny of SRS2, which is a subset of a larger family of 10 members, also containing HMI1, a mitochondrial helicase and its three orthologues. (D) phylogenetic tree of the five yeast species studied here (Dujon et al. 2004). Numbers refer to the branch in which gene duplications occurred. 1: HMI1/SRS2 duplication. 2: SRS2 tandem duplication. 3: TOP1 and SGS1 duplications. Sace: S. cerevisiae, Sapa: S. paradoxus, Saba: S. bayanus, Saca: S. castellii, Sakl: S. kluyveri, Saku: S. kudriavzevii, Cagl: C. glabrata, Klwa: K. waltii, Klla: K. lactis, Caal: C. albicans, Deha: D. hansenii, Yali: Y. lypolytica, Scpo: Schizosaccharomyces pombe.

FIG. 1.—

Phylogenetic tree of TOP1, SGS1, and SRS2 families. Trees were obtained by the maximum likelihood method. Branch lengths are estimated under the model of amino acids substitution JTT (Jones, Taylor, and Thronton 1992). Percentages of bootstrap values for internal nodes are indicated on each branch (see Materials and Methods for details). When the number of homologous sequences among the Génolevures species was less than 10, additional homologous sequences were retrieved among the other available fungal genomes. (A) phylogeny of TOP1. (B) Phylogeny of SGS1. (C) Phylogeny of SRS2, which is a subset of a larger family of 10 members, also containing HMI1, a mitochondrial helicase and its three orthologues. (D) phylogenetic tree of the five yeast species studied here (Dujon et al. 2004). Numbers refer to the branch in which gene duplications occurred. 1: HMI1/SRS2 duplication. 2: SRS2 tandem duplication. 3: TOP1 and SGS1 duplications. Sace: S. cerevisiae, Sapa: S. paradoxus, Saba: S. bayanus, Saca: S. castellii, Sakl: S. kluyveri, Saku: S. kudriavzevii, Cagl: C. glabrata, Klwa: K. waltii, Klla: K. lactis, Caal: C. albicans, Deha: D. hansenii, Yali: Y. lypolytica, Scpo: Schizosaccharomyces pombe.

FIG. 2.—

Functional domains of yeast Sgs1 proteins and human RecQ homologues. Domains were defined according to the CDD, except the TopIII-binding domain defined as in Mullen, Kaliraman, and Brill (2000). For each protein the number of amino acids (according to the NCBI genome annotation, in the case of the five human orthologues) is indicated above the C-terminus. Right: amino acids surrounding the DEAH motif (bold) are shown. In RecQ5, two DEAH motifs are present in the protein, separated by 340 amino acids.

FIG. 2.—

Functional domains of yeast Sgs1 proteins and human RecQ homologues. Domains were defined according to the CDD, except the TopIII-binding domain defined as in Mullen, Kaliraman, and Brill (2000). For each protein the number of amino acids (according to the NCBI genome annotation, in the case of the five human orthologues) is indicated above the C-terminus. Right: amino acids surrounding the DEAH motif (bold) are shown. In RecQ5, two DEAH motifs are present in the protein, separated by 340 amino acids.

Genes Involved in DNA Repair

NER is the main mechanism used to remove pyrimidine dimers induced by UV cross-linking or chemical damage such as those caused by benzopyrene, aflatoxin, and cisplatin (Lindahl and Wood 1999). In humans, NER-defective individuals are affected by xeroderma pigmentosum (XP) a disorder associated with hypersensitivity to sunlight and a 1,000-fold increase in the occurrence of skin cancer as compared to normal individuals. The S. cerevisiae homologues of XP genes are conserved in other yeast species (table 1). The transcription-coupled repair pathway is not completely conserved in K. lactis and Y. lipolytica because they both lack RAD28, the yeast homologue of human CSA involved in Cockayne syndrome. Genes involved in the mutational repair pathway (RAD18, RAD6) and its dedicated error-prone DNA polymerase REV3 (Pol ζ) are well conserved in all species, as well as the postreplicational repair helicase RAD5. Interestingly, the RAD6 gene, which is the most highly conserved of all 106 genes among the five species (fig. 3A), has two introns in D. hansenii and none in the four other yeast species. POL4, encoding the orthologue of Pol λ, a β-like DNA polymerase involved in base excision repair is not detected in D. hansenii. Finally, genes involved in the mismatch repair (MMR) pathway are completely conserved in C. glabrata and K. lactis. For the two more distant species, only the core MutS and MutL homologues, MLH1, MSH2, MSH3, MSH6, and PMS1 (Kolodner 1996), are found (table 1). Interestingly, MSH1 whose role is essential for maintenance of mitochondrial DNA is conserved in all four species.

FIG. 3.—

(A) Multiple alignment of the Saccharomyces cerevisiae RAD6 gene product with its four orthologues using the ClustalW software (Thompson, Higgins, and Gibson 1994). Identical amino acids are shown by a star, amino acids belonging to a ‘strong’ group are indicated by a column, and amino acids belonging to a ‘weaker’ group are shown by a single dot. The brackets show the location of the two introns in the Debaryomyces hansenii orthologue. Note the missing acidic tail in Yarrowia lipolytica. (B) Multiple alignment of the S. cerevisiae SPO11 gene product with its four orthologues using ClustalW. The conserved catalytic Tyr135 residue is boxed. Only the central well-conserved part of the protein is shown.

FIG. 3.—

(A) Multiple alignment of the Saccharomyces cerevisiae RAD6 gene product with its four orthologues using the ClustalW software (Thompson, Higgins, and Gibson 1994). Identical amino acids are shown by a star, amino acids belonging to a ‘strong’ group are indicated by a column, and amino acids belonging to a ‘weaker’ group are shown by a single dot. The brackets show the location of the two introns in the Debaryomyces hansenii orthologue. Note the missing acidic tail in Yarrowia lipolytica. (B) Multiple alignment of the S. cerevisiae SPO11 gene product with its four orthologues using ClustalW. The conserved catalytic Tyr135 residue is boxed. Only the central well-conserved part of the protein is shown.

Double-Strand Break Repair Genes

Genes in this category have been subdivided in two subclasses (table 1), those involved in NHEJ and those involved in homologous recombination (HR). DNA end-joining is a conserved process through evolution, from yeast to man. In human B and T lymphocytes, V(D)J recombination of immunoglobulin chains and of T-cell receptors is achieved by NHEJ, involving a number of genes including RAG1 and RAG2 (Grawunder, West, and Lieber 1998). In S. cerevisiae, there is no Rag protein but the end-joining machinery is very well conserved, and NHEJ has been mostly studied in this organism using HO and I-SceI–induced DSBs (for review see Haber 1995). It was recently shown that a V(D)J recombination substrate was correctly and precisely processed in yeast when the human RAG1 and RAG2 genes are coexpressed, showing that the whole yeast end-joining machinery is proficient to form signal joints (Clatworthy et al. 2003). In hemiascomycetes, the Ku complex is conserved, along with the Ligase IV orthologue (table 1). In the MRX complex, MRE11 and RAD50 are found in all species, whereas XRS2 is only detected in C. glabrata and K. lactis. XRS2 is the less well-conserved gene of the MRX complex, having no structural but a functional homologue in humans, called NBS1 (Carney et al. 1998). TBlastn search on D. hansenii and Y. lipolytica, using NBS1 as the query sequence did not reveal any homologue either. Surprisingly, the Ligase IV–interacting factor, LIF1, is only found in C. glabrata, whereas LIF2/NEJ1 is only found in K. lactis. LIF2/NEJ1 is a haploid-specific gene that regulates the efficiency of NHEJ in yeast cells, depending on whether they express only one set of mating-type proteins (a or α proteins) or both sets (a and α proteins) (Frank-Vaillant and Marcand 2001; Valencia et al. 2001). Because Lif2p was found to specifically interact with Lif1p in a two-hybrid screen (Frank-Vaillant and Marcand 2001), it is therefore surprising that K. lactis does not have a recognizable LIF1 gene. The Sir1 protein, involved in silencing HMLα and HMRa loci in S. cerevisiae, was not detected in C. glabrata (Fabre et al. 2005). This species has lost its ability to mate, supposedly because both a and silent mating-type cassettes are now expressed. Therefore, it is possible that selection pressure to keep the haploid regulator gene LIF2 in this species was abolished, eventually leading to the loss of this gene. Among genes involved in HR, the RAD52 gene product essential to this process is found in all four species along with the RecA homologue, RAD51, whose product catalyzes strand invasion and strand exchange (for review see Pâques and Haber 1999). Rad54p, Rdh54p, and Exo1p are also well conserved, whereas other accessory proteins that facilitate HR, like Rad55p and Rad59p, are not found in the most distant species. Finally, two copies of Srs2p were found to be duplicated in tandem in K. lactis (KLLA0F14256g and KLLA0F14234g, supplementary table). Multiple alignments show that both copies are very well conserved in their N-terminal part, in which the UvrD helicase domain is found (data not shown). Phylogeny demonstrates that both copies are at the same distance from ScSrs2p (fig. 1C) and that Dn values are almost identical (DnKLLA0F14256g = 0.5; DnKLLA0F14234g = 0.42, as compared to S. cerevisiae). This suggests that either both copies have evolved at the same rate, that the duplication is fairly recent in the evolution of this yeast, or that there is a high level of gene conversion between tandemly duplicated genes. Tandem paralogues are often more conserved than dispersed paralogues; this is a general trend of tandem duplications (Dujon et al. 2004).

Weak Conservation of Genes Involved in Meiotic Recombination

HR during meiosis is a highly regulated process by which genetic information is reshuffled between homologous chromosomes (for review see Zickler and Kleckner 1998). During this process, DSBs are generated by the Spo11p topoisomerase and then processed by the meiotic recombination machinery involving the Mre11p-Rad50p-Xrs2p complex. Spo11p is homologous to the A subunit of type VI topoisomerases, such as those found in archaebacteria (Bergerat et al. 1997). Homologues to SPO11 are found in all four yeast species, despite extensive sequence divergence. The Tyr135 residue essential for its catalytic activity is conserved, strongly suggesting that the four orthologues are functional in vivo (fig. 3B). The occurrence of crossovers is also regulated during meiosis, although little is known about the precise mechanism by which a recombination intermediate is resolved as a crossover or as a noncrossover, in vivo. It involves—at least—two different pathways: the Msh4-Msh5 pathway and the Mus81-Mms4 pathway. Msh4 and Msh5 proteins function as heterodimers in S. cerevisiae, and the corresponding mutants show a reduced frequency of meiotic crossovers as compared to wild-type strains (Pochart, Woltering, and Hollingsworth 1997). The msh5 mutant is profoundly affected at an early stage during meiotic recombination, showing a decreased level of early recombination intermediates leading to crossovers (Börner, Kleckner, and Hunter 2004). Their simultaneous absence in D. hansenii and Y. lipolytica might reflect a different mechanism to control crossovers in these species. The S. cerevisiae Mus81-Mms4 complex is able to process branched structures arising during mitotic or meiotic replication/recombination that are not canonical Holliday junctions (Fricke, Bastin-Shanower, and Brill 2005). Mus81p is conserved in all species, whereas Mms4p was not found in Y. lipolytica. However, although Mus81p is known to be conserved throughout evolution, its partner is poorly conserved (Ögrünç and Sancar 2003). It is therefore possible that a functional homologue of Mms4p is also present in Y. lipolytica but not detected.

The other genes involved in the meiotic recombination pathway are most of the time poorly conserved in D. hansenii and absent in Y. lipolytica, with the exception of MRE2, whose product is involved in the splicing of MER2 and MER3 messenger RNAs in S. cerevisiae. MER2 is predicted to contain an intron only in C. glabrata. Mre2p belongs to the U1 snRNP in S. cerevisiae and therefore splices many transcripts other than those involved in meiosis. It is thus probable that the MRE2 gene does not play a role anymore in meiotic recombination in K. lactis or D. hansenii. In conclusion, the only genes that are found in all five yeast species are genes involved in initiating recombination by making and processing DSBs (SPO11, MRE11, RAD50) in resolving recombination intermediates (MUS81) or the general splicing factor MRE2.

Checkpoint Proteins

Signaling DNA damage during the cell cycle is regulated by a series of proteins that activate the so-called “checkpoints” (for review see Zhou and Elledge 2000). Most of them are conserved except in Y. lipolytica. RAD9 is the only gene that is missing in D. hansenii in addition to Y. lipolytica. Most probably, DDC2, which has a human functional homologue (ATRIP, table 1), is also conserved in Y. lipolytica but is too diverged to be recognized.

Conservation of DNA Maintenance Pathways During Evolution

Given that some of the pathways are very well conserved in the five hemiascomycetous yeasts (e.g., replication or NER proteins) and others are missing several components, we wanted to know if amino acid conservation was the same among the different pathways. We performed pairwise Smith-Waterman alignments between each S. cerevisiae protein and its putative orthologues. Percentages of identity are shown in figure 4 for each species in each pathway. The average identity for each pathway was also calculated. Note that proteins have been classified in a pathway according to one of their functions, although some of them act in several distinct pathways. The best example is the MRE11-RAD50-XRS2 complex, classified in the NHEJ pathway, but which is known to be involved in formation and processing of meiotic DSBs, S-phase checkpoint activation, and HR (for review see Haber 1998). Nevertheless, when amino acid conservations of each pathway are compared, they generally follow the phylogenetic tree, i.e., C. glabrata is the closest to S. cerevisiae and Y. lipolytica is the farthest (fig. 4). The only exception is the meiotic recombination pathway, in which the only three genes to be conserved in Y. lipolytica (MRE2, MUS81, and SPO11) show a higher identity to S. cerevisiae orthologues than the corresponding D. hansenii, K. lactis, and C. glabrata genes. In order to determine if evolutionary rates were similar in the five species for these three genes, we calculated the Dn and Ds rates of nonsynonymous and synonymous substitutions. Because informative sites are saturated (Ds > 5), we took into consideration only Dn values. Using this criterion, we confirmed that SPO11 and MRE2 (but not MUS81) evolved slower in Y. lipolytica (DnSPO11 = 0.76; DnMRE2 = 0.69, as compared to S. cerevisiae) than in D. hansenii (DnSPO11 = 0.98; DnMRE2 = 0.73, as compared to S. cerevisiae). As a control, we also determined the average level of amino acid conservation between all S. cerevisiae proteins and their orthologues in each of the four species and used it as a baseline (see Materials and Methods). As expected, conservation follows the phylogenetic tree, i.e., C. glabrata proteins share a higher percentage of identity with S. cerevisiae proteins (60%) than Y. lipolytica proteins (50%). Hence, the only pathway in which proteins reach the amino acid identity baseline in each species is the replication pathway; most of the others (and all of them in Y. lipolytica) are below the baseline. This suggests that proteins involved in pathways whose average amino acid conservation is under the baseline diverge more rapidly than the average orthologous proteome, perhaps reflecting more flexibility in proteins involved in repair and recombination than in proteins involved in replication.

FIG. 4.—

Values of Smith-Waterman identity scores for conserved proteins in each pathway for each species. Each dot corresponds to a pairwise alignment between a protein and its Saccharomyces cerevisiae orthologue. Rad51p and Dmc1p are indicated just below or above the corresponding dot (see text). Average identity of each pathway is indicated above the 100% line. Average identity of all orthologous proteins for each species is indicated in parentheses following the species name.

FIG. 4.—

Values of Smith-Waterman identity scores for conserved proteins in each pathway for each species. Each dot corresponds to a pairwise alignment between a protein and its Saccharomyces cerevisiae orthologue. Rad51p and Dmc1p are indicated just below or above the corresponding dot (see text). Average identity of each pathway is indicated above the 100% line. Average identity of all orthologous proteins for each species is indicated in parentheses following the species name.

Protein Complexes

Among the 106 genes we have studied, many were known to encode products belonging to multiprotein complexes. One might expect that selection pressure would be the same for all members of a protein complex because if one of the members accumulates mutations faster than the other members, interaction between the different members could be rapidly lost and complex functionality disrupted. Therefore, one expects that in some cases all the members of a given complex are absent (they all evolved faster and are hence not recognizable anymore, Snel and Huynen 2004). Out of 21 known complexes, 12 are found in all organisms and 1 (Msh4p-Msh5p) is found only in C. glabrata and K. lactis, suggesting that either MSH4 and MSH5 genes evolved faster and are not detected anymore in the two more distant yeast species or that they appeared in the common ancestor of S. cerevisiae and K. lactis. The last eight complexes contain one or two members that are not conserved in each species (fig. 5).

FIG. 5.—

Conservation of known protein complexes in hemiascomycetous yeasts. Left: protein complexes for which each species contains at least one member. Right: complexes in which at least one member is present in at least a species and absent in at least another.

FIG. 5.—

Conservation of known protein complexes in hemiascomycetous yeasts. Left: protein complexes for which each species contains at least one member. Right: complexes in which at least one member is present in at least a species and absent in at least another.

Discussion

In the present study, we have analyzed the content of four fully sequenced hemiascomycetous yeast genomes to find orthologues of 106 S. cerevisiae genes involved in replication, repair, and recombination pathways. The aim of this work was (1) to identify orthologous pathways in other yeast species and to investigate the conservation of these pathways; (2) get insights into the evolution of genes involved in such pathways, particularly the frequency with which gene duplication/loss occurred; and (3) try to draw conclusions about the biological properties of these hemiascomycetous yeast species based on their gene content.

Conservation of Pathways

Pathways have been defined arbitrarily because many proteins belong to several pathways and therefore all pathways are interconnected with each other. However, despite such interconnection, some pathways such as meiotic recombination and checkpoints are less conserved than others, such as replication and NER. There are two independent criteria that may be used to estimate the conservation of pathways. The first—the presence/absence criterion—is used to determine the ratio of genes that are found in each species over the total number of genes in this pathway (table 1). The second—the conservation criterion—is used to calculate the average conservation in amino acid of proteins belonging to a given pathway for each species (fig. 4). Not surprisingly, the replication machinery comes first using both criteria, and almost all genes are present in each species and exhibit a high level of similarity with S. cerevisiae genes (table 1 and fig. 4). The NER pathway is very well conserved (all the genes are found in each species), but amino acid conservation is lower in K. lactis, D. hansenii, and Y. lipolytica than for proteins belonging to the HR pathway, in which many accessory proteins are not found in the more distant species (table 1 and fig. 4). In terms of presence/absence, the meiotic recombination machinery is missing several members, even in species related to S. cerevisiae. Most of the genes that are not found in C. glabrata and K. lactis belong to this pathway (table 1). We found that, in general, proteins interacting with DNA are more conserved than structural proteins, proteins that are part of a scaffold and other cofactors. It is striking that Rad51p and Dmc1p catalyzing strand exchange reactions are the most conserved of their respective pathways. Similarly, the Mre11p-Rad50p complex and Spo11p, necessary to make and process meiotic DSBs, are conserved along with Mus81p, involved in resolving recombination intermediates. All these proteins interact directly with DNA and are much more conserved than proteins involved in making the synaptonemal complex or other structural proteins and cofactors.

Gene Duplications During Evolution

Paralogous sets of genes play a key role in defining functional biological systems. For example, the MutS family of proteins contains six members in S. cerevisiae (MSH1-6, table 1), with distinct functions and specializations. Another example is the replicative helicase, formed by assembly of six distinct subunits, encoded by six different genes (MCM2-7), arising from successive gene duplications during evolution. In the present work, we found that both SGS1 and TOP1 were duplicated in C. glabrata and that SRS2 is tandemly duplicated in K. lactis. SGS1 encodes a DEAD/DEAH helicase of the RecQ/BLM/WRN family and has been shown to interact genetically with Top3p (Gangloff et al. 1994) and Top1p (Tong et al. 2001) and physically with Top2p (Watt et al. 1995). The duplicated Sgs1p and the duplicated Top1p both arose from duplication events prior to the S. cerevisiaeC. glabrata speciation, and both duplicated genes have been conserved in C. glabrata and lost in S. cerevisiae (fig. 1D). Given that Dn values are rather low for both genes, it is probable that both duplicated proteins are under selection pressure in C. glabrata. This could imply being part of an alternative complex involved in replication and/or repair or being part of a Sgs1-containing complex that would be specific to the life cycle of this pathogenic yeast. Interestingly, the duplicated copy of Sgs1p lacks its N-terminal and C-terminal parts (fig. 2) but retains the central helicase domain. It is therefore possible that it lost its DNA-binding activity but is still active as a helicase, maybe as part of a multicomponent complex. In humans, there are five homologues of Sgs1p, and two of them (RecQ5 and RecQL, fig. 2) are shorter versions, lacking either the N-terminal part (RecQL) or both the N- and C-terminal parts (RecQ5) but retaining their helicase domain. It is interesting that in C. glabrata, a short copy of Sgs1p was also found. We performed local and global alignments between the Sgs1p copy and the five human orthologues and concluded that although being a shortened version of Sgs1p, the C. glabrata copy is closer to WRN, BLM, and RecQ4 (RTS) than to the human RecQ5 and RecQL helicases. We therefore concluded that evolution of this protein family in C. glabrata and man was different. The Top1p duplication is interesting because this gene is duplicated in vertebrates but not in S. cerevisiae, Schizosaccharomyces pombe, or plants (Zhang et al. 2004). In vertebrates, one gene product is addressed to the nucleus and the other to mitochondria (Zhang et al. 2001). In C. glabrata, the duplicated copy (CAGL0J11660g) is predicted to encode a nuclear product, but no obvious nuclear nor mitochondrial addressing signal could be found in the original gene (CAGL0E02431g; Y. Pommier, personal communication). However, we know that the S. cerevisiae orthologue functions in the nucleus. Hence, this suggests that both gene products in C. glabrata are nuclear, and therefore that the evolution of this protein family in C. glabrata and in vertebrates was also different. It was previously shown that the very conserved lysine residue (K41) in Srs2p was essential for the adenosine triphosphatase (ATPase) activity (Krejci et al. 2004). This residue is present among the five species in the center of the completely conserved motif 35G36P37G38T39G40K41T42K43. In addition, in the duplicated copy of SRS2 in K. lactis, this motif is also completely conserved, suggesting that all the Srs2p orthologues are functional in the four other hemiascomycetes. The Hmi1p helicase is a paralogue of Srs2p in S. cerevisiae. Hmi1p is a mitochondrial protein and is essential for maintenance of mitochondrial DNA (Sedman et al. 2000). It was found in all species except Y. lipolytica, and the phylogenetic tree shows that the duplication of the SRS2/HMI1 gene ancestor occurred in the common ancestor to S. cerevisiae and D. hansenii (fig. 1C and D). Consistent with this observation, the conserved ATPase motif in Hmi1p only differs by one amino acid (Thr39 → Ser39) from Srs2p. This conservative mutation is found in all four species in which a HMI1 orthologue is detected, strengthening the idea that the formation of paralogues occurred before speciation of our yeasts. Finally, it was shown that the C-terminal part of the Hmi1 protein contains the mitochondrial targeting signal (Lee et al. 1999). Alignments of Srs2p and Hmi1p orthologues show that both proteins are very well conserved in the N-terminal part, containing the ATPase motif, but conservation of the C-terminal part is weak. Therefore, Srs2p/Hmi1p is probably a case of gene duplication before speciation, leading after alteration of the C-terminal part of one of the duplicated copies to a specialization of function, with both proteins being DNA helicases but one addressed to the mitochondria and the other to the nucleus. Our results strongly suggest that in each case of gene duplication, both copies are probably functional and have retained their catalytic activity, although they might be active in different cell compartments and/or on different substrates (subfunctionalization) (Lynch and Conery 2000).

Pathway Conservation, Evolution, and Yeast Biological Properties

We showed that genes belonging to the meiotic recombination machinery are poorly conserved in hemiascomycete species (in terms of presence/absence). However, K. lactis, D. hansenii, and Y. lipolytica undergo meiosis (Herman and Roman 1966; Kreger-van Rij and Veenhuis 1975; Casarégola et al. 2000). This means that although most of the genes necessary to go through meiotic recombination in S. cerevisiae are not detected in other yeasts, they must have functional orthologues able to carry out similar functions. Interestingly, the most highly conserved protein of the HR pathway is Rad51p in each species, and the most highly conserved protein of the meiotic recombination pathway is Dmc1p in the three species in which it is detected (fig. 4). Because DMC1 and RAD51 presumably come from the duplication of a common ancestor, our results suggest that this duplication occurred after the divergence between Y. lipolytica and the four other yeast species. It has also been proposed that organisms undergoing meiosis can be classified in two different groups (Stahl et al. 2004). In group I, organisms do not depend on meiotic DSB-repair functions to achieve synapsis (Drosophila melanogaster, Caenorhabditis elegans, Neurospora crassa), whereas in group II organisms, synapsis may only occur if DSB-repair is functional (e.g., S. cerevisiae). In group II organisms, the DMC1, HOP2, and MND1 genes are found, whereas they are apparently absent in group I organisms. This would suggest that group I organisms have lost these three genes or that they have been independently acquired during evolution. Therefore, Y. lipolytica would be classified as a group I organism because none of these three genes is found, whereas the four other species all contain these three genes (table 1 and data not shown for MND1). In addition, the Msh4-Msh5 protein complex involved in crossover control is also missing in this yeast. Taken together, these data suggest that although Y. lipolytica undergoes meiotic recombination (Wickerham, Kurtzman, and Herman 1970; Gaillardin, Charoy, and Heslot 1973), its properties are most probably very different from the four other hemiascomycetous yeasts.

In order to determine if the differences observed among the different yeast species for the NHEJ pathway could reflect a difference in the efficiency of DSB-repair mechanisms, we irradiated haploid cells with a source of γ-radiation. γ-Rays are known to induce single- and double-strand breaks in chromosomes, and resistance to ionizing radiations is a measure of how efficient the DSB-repair systems are in a given organism (Esposito and Wagstaff 1981). At low energy (50 Gys), the four hemiascomycetes are slightly more resistant than a haploid S. cerevisiae strain to ionizing radiations (supplementary figure). At higher doses (300 Gys), all five yeast species show the same sensitivity to γ-rays. We concluded that, most probably, no gene dramatically affecting the efficiency of DSB-repair was missing in the four species. This suggests again that the NHEJ pathway is functional, despite the apparent absence of some of its members. The higher resistance at low doses may be hypothesized by the existence of a more efficient pathway, for example HR with the sister chromatid, that would occur more often in those species as compared to S. cerevisiae, perhaps because of a longer S-G2 phase of the cell cycle. Further experimentation will be required to determine if cell cycles are the same in these five hemiascomycetes.

Supplementary Materials

One supplementary table: list of homologues in each yeast species.

One supplementary figure: comparison of survival to γ-irradiation between the five hemiascomycetes.

SUPPLEMENTARY FIGURE. Comparison of survival to γ-irradiation between the five hemiascomycetes. For each haploid strain, approximately 300 cells were plated on YPGlu plates and irradiated at different doses (0, 50, 100, and 300 Gy) using a 137Cs source, at a dose rate of 4 Gy/min. After 3 days of incubation at 30°C, survival was determined as the number of colony forming units (CFU) at each dose divided by the number of CFU at 0 Gy. The average of two independent experiments is shown for each species.

Franz Lang, Associate Editor

We thank our colleagues for fruitful discussions, particularly G. Fischer for careful reading of the manuscript, J. Haber for many suggestions, and A. Thierry for her expertise with in silico intron designing. This work was supported by the Consortium National de Recherche en Génomique (to Génoscope and to Institut Pasteur Génopole), the CNRS (GDR2354, Génolevures sequencing consortium), the Ministère de la Jeunesse, de l‘Education et de la Recherche (ACI IMPBio no. IMPB114 “Génolevures en ligne”), and the “Conseil Régional d’Aquitaine” (“Génotypage et Génomique Comparée”). A.K. is the recipient of a doctoral fellowship from the “Ministère de l‘Education Nationale, de l’Enseignement Supérieur et de la Recherche.” B.D. is a member of the Institut Universitaire de France.

References

Bennett, J. E., K. Izumikawa, and K. A. Marr.
2004
. Mechanism of increased fluconazole resistance in Candida glabrata during prophylaxis.
Antimicrob. Agents Chemother.
 
48
:
1773
–1777.
Bergerat, A., B. de Massy, D. Gadelle, P. C. Varoutas, A. Nicolas, and P. Forterre.
1997
. An atypical topoisomerase II from Archaea with implications for meiotic recombination.
Nature
 
386
:
414
–417.
Bolotin-Fukuhara, M., C. Toffano-Nioche, F. Artiguenave et al. (11 co-authors).
2000
. Genomic exploration of the hemiascomycetous yeasts: 11. Kluyveromyces lactis.
FEBS Lett.
 
487
:
66
–70.
Bon, E., S. Casaregola, G. Blandin et al. (11 co-authors).
2003
. Molecular evolution of eukaryotic genomes: hemiascomycetous yeast spliceosomal introns.
Nucleic Acids Res.
 
31
:
1121
–1135.
Börner, G. V., N. Kleckner, and N. Hunter.
2004
. Crossover/noncrossover differentiation, synaptonemal complex formation, and regulatory surveillance at the leptotene/zygotene transition of meiosis.
Cell
 
117
:
29
–45.
Burgers, P. M., E. V. Koonin, E. Bruford et al. (21 co-authors).
2001
. Eukaryotic DNA polymerases: proposal for a revised nomenclature.
J. Biol. Chem.
 
276
:
43487
–43490.
Cann, I. K. O., and Y. Ishino.
1999
. Archaeal DNA replication: identifying the pieces to solve a puzzle.
Genetics
 
152
:
1249
–1267.
Carney, J. P., R. S. Maser, H. Olivares, E. M. Davis, M. Le Beau, J. R. Yates III, L. Hays, W. F. Morgan, and J. H. J. Petrini.
1998
. The hMre11/hRad50 protein complex and Nijmegen breakage syndrome: linkage of double-strand break repair to the cellular DNA damage response.
Cell
 
93
:
477
–486.
Casarégola, S., C. Neuveglise, A. Lepingle, E. Bon, C. Feynerol, F. Artiguenave, P. Wincker, and C. Gaillardin.
2000
. Genomic exploration of the hemiascomycetous yeasts: 17. Yarrowia lipolytica.
FEBS Lett.
 
487
:
95
–100.
Castresana, J.
2000
. Selection of conserved blocks from multiple alignments for their use in phylogenetic analysis.
Mol. Biol. Evol.
 
17
:
540
–552.
Clatworthy, A. E., M. A. Valencia, H. J. H., and M. A. Oettinger.
2003
. V(D)J recombination and RAG-mediated transposition in yeast.
Cell
 
12
:
489
–499.
Dujon, B., D. Sherman, G. Fischer et al. (67 co-authors).
2004
. Genome evolution in yeasts.
Nature
 
430
:
35
–44.
Esposito, M. S., and J. E. Wagstaff.
1981
. Mechanisms of mitotic recombination. Pp. 341–370 in J. N. Strathern, E. W. Jones, and J. R. Broach, eds. The molecular biology of the yeast Saccharomyces—life cycle and inheritance. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.
Fabre, E., H. Muller, P. Therizols, I. Lafontaine, B. Dujon, and C. Fairhead.
2005
. Comparative genomics in hemiascomycete yeasts: evolution of sex, silencing and subtelomeres. Mol. Biol. Evol. (in press).
Frank-Vaillant, M., and S. Marcand.
2001
. NHEJ regulation by mating type is exercised through a novel protein, Lif2p, essential to the Ligase IV pathway.
Genes Dev.
 
15
:
3005
–3012.
Fricke, W. M., S. A. Bastin-Shanower, and S. J. Brill.
2005
. Substrate specificity of the Saccharomyces cerevisiae Mus81-Mms4 endonuclease.
DNA Repair
 
4
:
243
–251.
Gaillardin, C., V. Charoy, and H. Heslot.
1973
. A study of copulation, sporulation and meiotic segregation in Candida lipolytica.
Arch. Mikrobiol.
 
92
:
69
–83.
Gangloff, S., J. P. McDonald, C. Bendixen, L. Arthur, and R. Rothstein.
1994
. The yeast type I topoisomerase Top3 interacts with Sgs1, a DNA helicase homolog: a potential eukaryotic reverse gyrase.
Mol. Cell. Biol.
 
14
:
8391
–8398.
Goldman, N., and Z. Yang.
1994
. A codon-based model of nucleotide substitution for protein-coding DNA sequences.
Mol. Biol. Evol.
 
11
:
725
–736.
Grawunder, U., R. B. West, and M. R. Lieber.
1998
. Antigen receptor gene rearrangement.
Curr. Opin. Immunol.
 
10
:
172
–180.
Guindon, S., and O. Gascuel.
2003
. A simple, fast and accurate algorithm to estimate large phylogenies by maximum likelihood.
Syst. Biol.
 
52
:
696
–704.
Haber, J. E.
1995
. In vivo biochemistry: physical monitoring of recombination induced by site-specific endonucleases.
BioEssays
 
17
:
609
–620.
———.
1998
. The many interfaces of Mre11.
Cell
 
95
:
583
–586.
Herman, A., and H. Roman.
1966
. Allele specific determinants of homothallism in Saccharomyces lactis.
Genetics
 
53
:
727
–740.
Jones, D. T., W. R. Taylor, and Thornton J.M.
1992
. The rapid generation of mutation data matrices from protein sequences.
Comput. Appl. Biosci.
 
8
:
275
–282.
Kolodner, R.
1996
. Biochemistry and genetics of eukaryotic mismatch repair.
Genes Dev.
 
10
:
1433
–1442.
Kreger-van Rij, N. J. W., and M. Veenhuis.
1975
. Electron microscopy of ascus formation in the yeast Debaryomyces hansenii.
J. Gen. Microbiol.
 
89
:
256
–264.
Krejci, L., M. Macris, Y. Li, S. Van Komen, J. Villemain, T. Ellenberger, H. Klein, and P. Sung.
2004
. Role of ATP hydrolysis in the anti-recombinase function of Saccharomyces cerevisiae Srs2 protein.
J. Biol. Chem.
 
279
:
23193
–23199.
Lee, C. M., J. Sedman, W. Neupert, and R. A. Stuart.
1999
. The DNA helicase, Hmi1p, is transported into mitochondria by a C-terminal cleavable targeting signal.
J. Biol. Chem.
 
274
:
20937
–20942.
Lépingle, A., S. Casaregola, C. Neuveglise, E. Bon, H. Nguyen, F. Artiguenave, P. Wincker, and C. Gaillardin.
2000
. Genomic exploration of the hemiascomycetous yeasts: 14. Debaryomyces hansenii var. hansenii.
FEBS Lett.
 
487
:
82
–86.
Lindahl, T., and R. D. Wood.
1999
. Quality control by DNA repair.
Science
 
286
:
1897
–1905.
Lynch, M., and J. S. Conery.
2000
. The evolutionary fate and consequences of duplicate genes.
Science
 
290
:
1151
–1155.
Mullen, J. R., V. Kaliraman, and S. J. Brill.
2000
. Bipartite structure of the SGS1 DNA helicase in Saccharomyces cerevisiae.
Genetics
 
154
:
1101
–1114.
Notredame, C., D. Higgins, and J. Heringa.
2000
. A novel method for multiple sequence alignments.
J. Mol. Biol.
 
302
:
205
–217.
Ögrünç, M., and A. Sancar.
2003
. Identification and characterization of human MUS81-MMS4 structure-specific endonuclease.
J. Biol. Chem.
 
278
:
21715
–21721.
Pâques, F., and J. E. Haber.
1999
. Multiple pathways of recombination induced by double-strand breaks in Saccharomyces cerevisiae.
Microbiol. Mol. Biol. Rev.
 
63
:
349
–404.
Pochart, P., D. Woltering, and N. M. Hollingsworth.
1997
. Conserved properties between functionally distinct MutS homologs in yeast.
J. Biol. Chem.
 
272
:
30345
–30349.
Sedman, T., S. Kuusk, S. Kivi, and J. Sedman.
2000
. A DNA helicase required for maintenance of the functional mitochondrial genome in Saccharomyces cerevisiae.
Mol. Cell. Biol.
 
20
:
1816
–1824.
Snel, B., and M. A. Huynen.
2004
. Quantifying modularity in the evolution of biomolecular systems.
Genome Res.
 
3
:
391
–397.
Stahl, F. W., H. M. Foss, L. S. Young, R. H. Borts, M. F. F. Abdullah, and G. P. Copenhaver.
2004
. Does crossover interference count in Saccharomyces cerevisiae?
Genetics
 
168
:
35
–48.
Thompson, J. D., D. G. Higgins, and T. J. Gibson.
1994
. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice.
Nucleic Acids Res.
 
22
:
4673
–4680.
Tong, A. H., M. Evangelista, A. B. Parsons et al. (13 co-authors).
2001
. Systematic genetic analysis with ordered arrays of yeast deletion mutants.
Science
 
294
:
2364
–2368.
Valencia, M., M. Bentele, M. B. Vaze, G. Herrmann, E. Kraus, S.-E. Lee, P. Schär, and J. E. Haber.
2001
. NEJ1 controls non-homologous end-joining in Saccharomyces cerevisiae.
Nature
 
414
:
666
–669.
Watt, P. M., E. J. Louis, R. H. Borts, and I. D. Hickson.
1995
. Sgs1: a eukaryotic homolog of E. coli RecQ that interacts with topoisomerase II in vivo and is required for faithful chromosome segregation.
Cell
 
81
:
253
–260.
Wickerham, L. J., C. P. Kurtzman, and A. I. Herman.
1970
. Sexual reproduction in Candida lipolytica.
Science
 
167
:
1141
.
Yang, Z.
1997
. PAML: a program package for phylogenetic analysis by maximum likelihood.
Comput. Appl. Biosci.
 
13
:
555
–556.
Zhang, H., J. M. Barcelo, B. Lee, G. Kohlhagen, D. B. Zimonjic, N. C. Popescu, and Y. Pommier.
2001
. Human mitochondrial topoisomerase I.
Proc. Natl. Acad. Sci. USA
 
98
:
10608
–10613.
Zhang, H., L. H. Meng, D. B. Zimonjic, N. C. Popescu, and Y. Pommier.
2004
. Thirteen-exon-motif signature for vertebrate nuclear and mitochondrial type IB topoisomerases.
Nucleic Acids Res.
 
32
:
2087
–2092.
Zhou, B.-B. S., and S. J. Elledge.
2000
. The DNA damage response: putting checkpoints in perspective.
Nature
 
408
:
433
–439.
Zickler, D., and N. Kleckner.
1998
. The leptotene-zygotene transition of meiosis. Pp. 619–697 in A. Campbell, W. W. Anderson, and E. W. Jones, eds. Palo Alto, CA: Annual Review of Genetics.