|Home | About | Journals | Submit | Contact Us | Français|
Dna2 is a dual polarity exo/endonuclease, and 5′ to 3′ DNA helicase involved in Okazaki Fragment Processing (OFP) and Double-Strand Break (DSB) Repair. In yeast, DNA2 is an essential gene, as expected for a DNA replication protein. Suppression of the lethality of dna2Δ mutants has been found to occur by two mechanisms: overexpression of RAD27scFEN1, encoding a 5′ to 3′ exo/endo nuclease that processes Okazaki fragments (OFs) for ligation, or deletion of PIF1, a 5′ to 3′ helicase involved in mitochondrial recombination, telomerase inhibition and OFP. Mapping of a novel, spontaneously arising suppressor of dna2Δ now reveals that mutation of rad9 and double mutation of rad9 mrc1 can also suppress the lethality of dna2Δ mutants. Interaction of dna2Δ and DNA damage checkpoint mutations provides insight as to why dna2Δ is lethal but rad27Δ is not, even though evidence shows that Rad27ScFEN1 processes most of the Okazaki fragments, while Dna2 processes only a subset.
Chromosomal DNA is subjected to constant damage not only from exogenous agents, such as ionizing radiation, ultraviolet radiation and base-damaging drugs, but also from damage arising from faulty DNA replication and spontaneous base loss. Unrepaired or misrepaired damage can lead to cell lethality or genome instability, such as translocations or aneuploidy. The DNA damage checkpoint involves a cascade of protein kinases that arrests or slows down the cell cycle, allowing time to repair the damage. Additional functions of the checkpoint include stabilizing the DNA replication fork and stimulating DNA repair. Yeast (Saccharomyces cerevisiae) Mec1 and Tel1 are two phosphoinositol kinase-related protein kinases that initiate the DNA damage response in response to single-strand and double-strand DNA breaks.1,2 Two downstream kinases, Rad53 and Chk1, are activated by Mec1/Tel1 kinases after DNA damage, although Mec1 is the primary upstream kinase.3 Activated Rad53 then positively regulates an additional kinase, Dun1.4 Dun1 regulates levels of ribonucleotide reductase and arrests cells having DNA damage in late anaphase.4 Additional activities of Rad53 after DNA damage include stimulation of DNA repair by phosphorylation of Rad55, inhibition of firing of late origins, induction of transcription of DNA repair genes and phosphorylation of the nuclease Exo1 and its apparent inhibition.5–7 Exo1 is a 5′ to 3′ nuclease involved in mismatch base repair, double-strand break (DSB) repair and 5′ to 3′ degradation of uncapped telomeres.8–12
Rad9 and Mrc1 are “mediators” that transduce the signal between Mec1/Tel1 and the Rad53, Chk1 and Dun1 kinases.13–15 RAD9 is, in fact, the original defining gene of the DNA damage checkpoint response.16 Rad9 is a scaffold protein that binds the forkhead domain of Rad53 after phosphorylation by Mec1.17 The Rad9/Rad53 interaction results in Rad53 kinase activation through Rad53 dimerization and subsequent transphosphorylation.14,18 Activated Rad53 delays the G2/M transition by inhibiting the degradation of cohesin and inhibiting spindle elongation.19,20 Activated Rad53 stabilizes stalled replication forks.21–23 rad9Δ mutants are sensitive to IR, UV and prematurely enter mitosis with damaged DNA.24 Mrc1 is a non-essential DNA replication protein that binds to Mcm2, the replicative helicase, Cdc45, a subunit of the replicative helicase and DNA polymerase ε (pol ε), the leading strand polymerase.25,26 Mrc1 couples DNA unwinding with replicative synthesis, and in mrc1Δ mutants, replication fork rate is reduced.13,27,28 After DNA damage Mrc1 is phosphorylated by Mec1, primarily in the N terminus, allowing Mrc1 to activate Rad53 by an undefined mechanism.29 The mechanism of activation of Rad53 by Mrc1 presumably differs from Rad9 activation of Rad53, since Mrc1 travels with the replication fork.30 However, to test whether Mrc1 phosphorylation by Mec1 is required for Rad53 activation, Osborn and Elledge mutated all the potential Mec1/Tel1 phosphorylation sites in Mrc1, creating the mutant mrc1-AQ. Activation of Rad53 by HU treatment is abolished in the rad9Δ mrc1-AQ mutant, and the rad9Δ mrc1-AQ mutant is sensitive to low doses of HU. Unlike the mrc1Δ mutant, the mrc1-AQ mutant does not exhibit a slow DNA replication phenotype and is specifically checkpoint-defective.30
Dna2 is a 5′ to 3′ helicase and 5′ to 3′ and 3′ to 5′ exo/endonuclease with single-strand annealing and strand exchange activity.31–36 dna2-1 mutants are defective in DNA replication and accumulate short DNA fragments at the restrictive temperature,37,38 and dna2Δ mutants are inviable. Rad27scFEN1 (Flap Endonuclease) is a 5′ to 3′ exo/endonuclease that processes the majority of the Okazaki fragments.39–41 Dna2 assists Rad27scFEN1 in RNA primer removal through a coordinated Rad27scFEN1, Dna2, RPA interaction at a subset of Okazaki fragments.42–46 rad27Δ mutants are viable but exhibit a very high frequency of mutation at repeated DNA sequences resulting from delayed OFP.47 The dna2-1 mutants do not show an increased frequency of these mutations, and therefore, the high mutational rate of rad27Δ mutants suggests Rad27scFEN1 processes the majority of the Okazaki fragments. This raises an obvious question: since OFP is essential, why are dna2Δ mutants inviable and rad27Δ mutants viable?
Comparison of the genetic interaction network of DNA2 and RAD27 provided a clue to the answer.48–51 dna2 mutants show synthetic lethality with approximately 46 other genes, and two-dimensional clustering of these revealed that DNA2 and RAD27 share the greatest number of interactions, consistent with function in the same pathway. This overlapping set of mutants affect DNA replication, DNA repair, histone modification and the cellular stress response. However, we also noticed that there are differences. In particular, dna2 mutants are not synthetically lethal with DNA damage checkpoint mutants mec1, rad53, rad9, rad24, rad17, ddc1, mec2, while rad27 mutants are synthetically lethal with these checkpoint genes. The only DNA damage checkpoint mutants that are synthetically lethal with dna2 are mrc1Δ, tof1Δ and csm3Δ; however, these are also part of the replication progression complex and important for DNA synthesis.25 Therefore, the synthetic lethality presumably results from the role of MRC1, TOF1 and CSM3 in DNA replication.
Previously, all suppressors of dna2Δ affected components of the OFP system. In this work, we report the surprising observation of a spontaneously arising suppressor of dna2Δ mutants that, instead, disrupts the DNA damage checkpoint through inactivation of the Rad9 “mediator.” We propose that the DNA repair functions of Dna2 contribute to its essential function, and in the absence of this repair function, the cell cycle inhibition by the checkpoint leads to permanent arrest rather than to survival, as it does in rad27Δ cells, which are repair proficient.
dna2Δ::kanMX strains are inviable; however, the inviability is suppressed by a deletion of PIF1. When a DNA2/dna2Δ::kanMX PIF1/pif1Δ::HIS3 diploid is sporulated and dissected, the dna2Δ::kanMX spores always segregate with pif1Δ::HIS3. During one dissection of a DNA2/dna2Δ::kanMX PIF1/pif1Δ::HIS3 diploid, however, we noticed a kanMX (kanomycin resistant) his3 (did not grow on histidine minus plates) colony, suggesting that it was dna2Δ PIF1. This indicated that the dna2Δ mutant could be suppressed by a mutation other than pif1Δ::HIS3. To check if the suppressor (supX) represented a genetic suppressor or a nonspecific suppressor (e.g., suppression resulting from overproduction of heat shock proteins), the viable dna2Δ PIF1 supX strain was crossed to a BY4741-derived strain and sporulated. Thirteen tetrads were dissected (Fig. 1). If the dna2Δ strain is suppressed by one suppressor, the expected number of dna2Δ strains recovered would be 13. In keeping with this, 28 DNA2 spores and 10 dna2Δ::KanMX spores were recovered. The recovery of dna2Δ strains at about 20 to 25% of dissected tetrads suggests that one mutation was segregating in the cross that is suppressing the lethality of dna2Δ. The dna2Δ spores in this cross were temperature-sensitive and ionizing radiation (IR)-sensitive (not shown). The IR sensitivity of the dna2Δ supX strain is significantly different from the IR resistance of a dna2Δ pif1Δ strain.52 Therefore, supx was not likely pif1Δ.
To identify the supX gene, DNA was prepared from a segregant (MB220) of the cross of BY4741(WT) × dna2Δ, and the genome was sequenced and compared to strain BY4741 (see Materials and Methods). The complete list of mutations in MB220 can be found in Supplemental Table 1. Three genes were found to have C-terminal deletions. Among these, the only logical candidate for a suppressor of dna2Δ is RAD9. No mutations were found in the PIF1 gene.
RAD9 encodes a 1,309 amino acid protein, and the suppressor mutation consisted of a stop codon at amino acid 320. Rad9 possesses at least four domains required for full activity. Amino acids 1–231 are required to activate Chk1; residues 390–458 and residues 593–620 are phosphorylated by Mec1 and contain the Rad53 interaction domain; residues 754–947 contain a Tudor domain, which binds methylated histones, and residues 1,027–1,310 contain two BRCT repeats that bind phosphorylated histones.14,53–55 The BRCT repeats are required for Rad9 dimerization, a necessary step in Rad53 activation. The 1–320 amino acid fragment in the rad9 suppressor thus possesses the region required for activation of Chk1 but is missing the Rad53 interaction domain, the Tudor domain and the BRCT domain.14 Since the chromatin binding region, dimerization region and Mec1 phosphorylation sites are missing, the mutant is therefore presumably completely defective in activation of Rad53 and Chk1. The rad9-320 mutation partially accounts for the IR sensitivity of the strain, although the dna2Δ rad9-320 strain is significantly more sensitive to irradiation than a rad9Δ strain (Budd M and Campbell JL, unpublished data). We designate the suppressive allele (supX above) as rad9-320, indicative of the site of the mutation.
To confirm that the rad9-320 mutation was suppressing dna2Δ, we asked whether expression of RAD9 could restore lethality to dna2Δ rad9-320. RAD9 cloned into a yeast plasmid under control of the GAL1 promoter was introduced into dna2Δ rad9-320. Since overexpression of RAD9 can in itself cause slow growth, we first determined induction conditions that were not lethal in a rad9Δ DNA2 strain (Fig. 2A). On 2% glucose plates, the plasmid had no effect on the growth of the strain. 0.05% galactose is the lowest concentration of galactose that results in detectable induction of the GAL1 promoter, with a level induction of 1/100 that of full induction of 2%.56 Concentrations of 0.05% and 0.2% galactose resulted in minor inhibition of growth of rad9Δ DNA2. This level of expression was sufficient to completely suppress the IR sensitivity of the strain, however (not shown). We then used these induction conditions in the dna2Δ rad9-320 strain carrying the GAL1-RAD9 plasmid. On glucose containing plates, the RAD9 plasmid resulted in no detectable growth inhibition. On 0.05% and 0.2% galactose, the RAD9 plasmid, but not control plasmids, completely blocked growth of the dna2Δ rad9-320 strain. These results show that expression of RAD9 causes inviability of the strain dna2Δ rad9-320. We conclude that the DNA damage checkpoint is deleterious rather than beneficial in the dna2Δ mutant.
The dna2Δ rad9-320 strain was viable but grew more slowly than wild type. We reasoned that this might be due to continued induction of the S-phase checkpoint by MRC1, and that mutation of the checkpoint function of MRC1 might further enhance the growth of dna2Δ rad9-320. We first tested the combined effect of rad9Δ mrc1-AQ double mutation on the temperature sensitivity of dna2-1 strains. The suppressive effect of rad9Δ, mrc1-AQ and rad9 mrc1-AQ mutants on dna2-1 growth at 23°C, 28°C and 34°C is shown in Figure 3. dna2-1 strains cease growth at 28°C; dna2-1 rad9Δ and dna2-1 mrc1-AQ cease growth at 34°C. Thus, although rad9Δ suppresses dna2Δ lethality, it does not suppress dna2-1 temperature sensitivity. However, dna2-1 rad9Δ mrc1-AQ is capable of growth at 34°C. Thus, the combined rad9Δ mrc1-AQ mutation is a strong suppressor of the temperature-sensitive phenotype of dna2-1 strains, similar to the pif1Δ suppression of dna2-1 strains.52 We then tested the ability of mrc1-AQ combined with rad9Δ as an additional suppressor of the dna2Δ lethality. Five independent WT/dna2Δ::natR rad9Δ::kanMX mrc1-AQ::HIS3 diploids were sporulated and dissected with an average of 12 to 13 tetrads per diploid. The data from one of the crosses is illustrated in Figure 4. The results of the 62 total diploids scored were as follows: (the number following the genotype indicates the number of spores identified): WT, 22; rad9Δ, 32; mrc1-AQ, 29; rad9Δ mrc1-AQ, 27; dna2Δ rad9Δ. 5; dna2Δ mrc1-AQ, 1; dna2Δ rad9Δ mrc1-AQ, 17; dna2Δ, 0. Assuming 100% viability, about 31 spores from each genotype should be recovered (62 tetrads × 4 spores + 8 genotypes = 31). From this data, we first point out that the five viable dna2Δ rad9Δ spores serve as an independent control, showing that complete deletion of RAD9 suppresses dna2Δ inviability, consistent with rad9-320 likely functioning as a Rad9-null in suppression. Second, 23 dna2Δ strains were recovered in combination with either or both of the checkpoint mutants, rad9 and/or mrc1-AQ, and none were recovered when dna2Δ was not combined with a checkpoint mutant. All the dna2Δ rad9Δ, dna2Δ mrc1-AQ and dna2Δ rad9Δ mrc1-AQ strains could form colonies upon restreaking at 23°C but not at 36°C. mrc1-AQ alone barely suppressed dna2Δ, and only one suppressor was recovered. rad9Δ is a more efficient suppressor than mrc1-AQ, and rad9Δ mrc1-AQ provided more efficient suppression, since they were recovered at higher frequency, although the dna2Δ rad9Δ mrc1-AQ colonies were not necessarily larger than the dna2Δ rad9Δ colonies (Fig. 4, compare squares and circles).
We have identified rad9Δ as a novel suppressor of the lethality of dna2Δ using full genome sequencing of a spontaneous suppressor. Suppression of the complete absence of Dna2 is surprising, because rad9Δ suppresses the growth defect of only a subset of dna2 point mutants and does not suppress the inability of dna2-1 mutants to grow at 34°C. The rad9-320 mutation identified in this work gives rise to a stop codon at amino acid 320, which is expected to create a functionally null allele. Even if the fragment survives nonsense-mediated decay, the only functions associated, to date, with the N-terminal domain remaining are Chk1 activation and binding to the replication initiator protein Dbp11.53 Indeed, in independent experiments, we verified that rad9Δ suppressed dna2Δ (Fig. 4), and that expression of RAD9 reversed the suppression by rad9-320 (Fig. 2). None of the additional mutations identified by DNA sequencing of the original suppressor strain appears to be a candidate suppressor mutation. DNA sequencing did not detect any mutations in the PIF1 gene, which we had previously identified as a dna2Δ suppressor. Interestingly, the rad9Δ mrc1AQ double mutation resulted in more efficient suppression than either mutation alone. We conclude that inefficiently processed DNA intermediates arising due to the absence of Dna2 do not activate either the Rad53 or Chk1 cell cycle checkpoint functions in the dna2Δ rad9Δ mrc1AQ mutants. In this way, terminal arrest is avoided, allowing further division in the complete absence of Dna2.
Previously, specific endogenous suppressors of dna2 lethality were attributed to increasing the efficiency of OFP. This may be mediated by overexpression of Rad27ScFEN1, which is directly involved in OFP, and of BLM, WRN, Exo1 and Mph1, which provide auxiliary functions and stimulate the nuclease activities of Dna2 and Rad27scFEN1.57–61 In Schizosaccharomyces pombe, overexpression of FEN1, pol δ or DNA ligase suppress a dna2 helicase mutant.62 A related phenomenon involves suppressing the production of irreparable toxic structures such as long 5′ flaps by Pif1 helicase. pif1 mutation allows dna2-1 strains to grow at 37°C.52,57 Pif1 is a 5′ to 3′ helicase that unwinds forked DNA substrates, RNA/DNA duplexes and removes telomerase bound to telomeres.63,64 A dna2Δ pif1Δ strain is viable but temperature sensitive; however, a dna2Δ pif1Δ pol32Δ strain does grow at 37°C.52 Thus, Pol32 is an additional suppressor of dna2 mutants. Pol32 is a stimulatory subunit of pol δ;.65 The suppression of dna2Δ by pif1Δ and further suppression by pol32Δ suggests that long flaps created during OPF by pol δ, Pol32 and Pif1 are potentially lethal substrates unless they are processed by Dna2 (Fig. 5). Additional suppressors of dna2 mutants are not involved in OFP. E1A overexpression was one of the first identified suppressors of Dna2.66 E1A is encoded by the human adenovirus, activates viral transcription and can promote cell cycle progression. TOR overexpression also suppresses some dna2 helicase mutants.67 The mechanisms involved are not understood.
Our current work identifies a new compensatory pathway that we propose involves desensitizing dna2 mutants to replication stress encountered in its absence. In addition to our newly documented suppression of the lethality of deletion of DNA2 by deletion of RAD9, rad9Δ and mec1 mutants have been observed to suppress the growth defects of a subset of dna2 point mutants and the synthetic lethality of a dna2-2 ctf4Δ mutant.67,68 rad9 and mrc1 mutants desensitize other DNA replication mutants to replication damage as well. cdc9 rad9 mrc1 mutants progress further into DNA synthesis than cdc9 rad9, which progress further than cdc9 mutants at 36°C.69 The lethality of orc2-1 and orc1-4 mutants at the restrictive temperature is suppressed by rad9Δ mutation.70 This is of general interest, because desensitizing cells to DNA damage during DNA replication is a hallmark of cancer cells and is the cause of genome instability. The most widespread desensitizing mutation is p53, a common mutation in human cancer.
We propose that cell cycle checkpoint activation in the absence of Dna2 probably results from persistant 5′ flaps created by combined polδ/Pif1 DNA strand displacement synthesis during Okazaki fragment processing (see Fig. 5 for details). Supporting this mode of checkpoint activation, Rad53 is highly phosphorylated in dna2-1 strains, and phosphorylation is Mec1-dependent, while Rad53 remains largely unphosphorylated in dna2-1 pif1Δ strains.52 Furthermore, we calculate that this mechanism could easily generate sufficient single-stranded DNA to activate the DNA damage response. It has been estimated that at least 104 bp of single-stranded DNA is required to activate Rad53 after an HO-induced DSB, based on the rate of 5′ to 3′ resection and the time of appearance of phosphorylated Rad53.71,72 Yeast contains about 2.3 × 107 nt, and therefore, about 2 × 105 Okazaki fragments, which are about 125 nt in length, are synthesized during S phase.73 If polδ/Pif1 strand displacement causes about 1% of the flaps to become 30 nt or longer, then about 2 × 103 flaps greater than 30 nt long or 1.2 × 105 nt of single-stranded DNA, is expected to be generated during S phase in a dna2Δ mutant, more than is necessary to activate the checkpoint. After checkpoint activation, cells normally repair the DNA damage, deactivate the checkpoint and resume cell division. dna2-1 strains repair the DNA damage (long RPA bound 5′ flaps) inefficiently since Rad53 is constitutively activated and dna2-1 strains divide very slowly, a 4 hour generation time at 23°C. dna2Δ strains would be unable to repair the unprocessed Okazaki fragments during the G2 division delay; thus the Rad9/Mrc1-dependent G2 division delay contributes to cell death rather than recovery.
In addition to failing to induce cell cycle arrest, another mechanism is also probably involved in the suppression of the dna2Δ lethality by rad9 and rad9 mrc1AQ. Dna2 has recently been shown to be involved in resection of DSBs in yeast.9,74–76 Although the DSB break repair function of Dna2 is missing, the viability of dna2Δ rad9-320 suggests that in the absence, but not in the presence, of Rad9, lethal endogenous DSBs are channeled into an alternative repair pathway. The Exo1-dependent DSB repair pathway is a good candidate, given previously documented genetic interactions and overlaps of function between Dna2 and Exo1.52 Exo1 is a 5′ to 3′ nuclease with homology to Rad27scFEN1 and is involved in mismatch base repair, resection at the ends of DSBs and decapped telomeres and can function in OFP when either Rad27scFEN1 or Dna2 is defective.9,10,77–79 Overproduction of Exo1 partially suppresses the growth defect of dna2-1 cells at 30°C and suppresses the growth defect of rad27Δ mutants at 37°C.50,78 dna2-1, dna2-2, dna2Δ pif1Δ and rad27Δ mutants are synthetically lethal with exo1Δ mutants.50,52,78 In processing of DSBs, either Exo1 or Dna2 can provide the resection function.9,74–76 Several recent lines of evidence suggest that activation of RAD9 and RAD53 inhibit processes dependent on Exo1 nucleolytic functions. Cdc13 is an essential telomere capping protein. After Cdc13 inactivation, extensive 5′ to 3′ degradation occurs at the telomere end, continuing as far as 15 kb from the end.80,81 The degradation is mainly Exo1-dependent.10 In the absence of Rad9, 5′ to 3′ degradation proceeds significantly further and faster, suggesting Rad9/Rad53 inhibits Exo1.82,83 Other evidence that Exo1 is inhibited by a Rad53-dependent step is the observation that in rad53 mutants, replication forks stalled by MMS or HU exhibit long regions of single-stranded DNA on the nascent lagging strand, and the appearance of single-stranded DNA is dependent on Exo1. In exo1Δ rad53 mutants, there is no single-stranded DNA, and exo1Δ suppresses the MMS sensitivity of a rad53 mutant.84 In RAD53 cells, stalled forks remain fully duplex and do not collapse, which could be explained if Exo1 is inhibited.85 In the rad9Δ and rad9 mrc1 mutants, Rad53 is deactivated, and the putative Rad53-dependent inhibition of Exo1 is relieved. In addition, Rad9 and the H3K79 histone methyltransferase Dot1 inhibit 5′ to 3′ degradation at DSBs and telomeres.86 A rad9Δ mutant would relieve much of the inhibition of Exo1, and rad9Δ mrc1AQ may relieve the remainder of the inhibition, allowing Exo1 to function in OFP/DSB repair as a backup repair pathway in the absence of Dna2 (Fig. 5).
The question remains why rad27Δ mutants are viable and dna2Δ mutants are inviable. We have previously proposed that the synthetic lethality of rad27Δ with recombinational repair mutants and the viability of dna2 mutants in combination with the same recombinational repair mutants suggests that DSBs occur more frequently in rad27Δ, and that, therefore, Rad27ScFEN1 processes most of the Okazaki fragments. The remaining OFs are inefficiently processed by Dna2, Exo1 and RnaseH. The synthetic lethalities also suggest that Rad27ScFEN1 is not required for DSB repair but is dedicated to DNA replication, thus a rad27Δ mutant requires the use of the G2 division delay to repair damage, allowing its survival (albeit with high frequency of mutation). Hence, rad27 rad9 is lethal.48,51,87 With dna2Δ, instead, even though there are fewer OFP failures and fewer replication-induced DSBs, G2 division delay is lethal because Dna2 is required for efficient repair. Potential backup repair pathways are suppressed in the presence of RAD9, and its deletion results in dna2 survival.
The strains are listed in Table 1. The plasmid BG1805-GAL1-Rad9 (2 µ ori, URA3) was obtained in the strain Y258 from Open Biosystems. The plasmid BG1805-Rad9 was purified from Y258, transformed into Eschericia coli and then purified. Standard genetic techniques were used for tetrad analysis and genetic analysis. The chromosomal DNA from MB220 was isolated using the Qiagen Genomic-tip P/20. The kanMX disruption cassette confers resistance to G418/geneticin, and natR cassette confers resistance to Nourseothricin.
Libraries were prepared from genomic DNA isolated from BJ4741 and its derivative strain MB220 carrying the dna2Δ suppressor, following the standard Illumina protocol. The libraries were sequenced on the Illumina Genome Analyzer II using a single read sequencing protocol producing 15.4 and 20.2 million 64 nt reads passing quality filtering, respectively. The sequence output corresponds to 80.2× coverage of the S. cerevisiae genome for BJ4741 and 105.2× coverage for MB220. To identify polymorphisms in MB220, each strain was compared to the reference S. cerevisiae genome (version S288C, downloaded from the Saccharomyces Genome Database, www.yeastgenome.org) and to each other using the VAAL software package, which can detect both single nucleotide polymorphisms (SNPs) and deletion/insertion events (indels).88 To identify genes affected by the mutations, genome annotations in GFF format (downloaded from SGD ftp://ftp.yeastgenome.org/yeast/data_download/chromosomal_feature/saccharomyces_cerevisiae.gff) were loaded into a MySQL database, and the output of the TruePoly module of VAAL containing the differences between MB220 and BJ4741 was compared against it. For each polymorphism, we recorded MB220, BJ4741 and reference genome sequences flanking the mutation, overlapping gene(s), gene description, associated GO terms and other available functional annotations, nature of the mutation (i.e., stop codon, non-synonymous substitution, synonymous substitution) and amino acid change produced by it. Our analysis88 identified 476 SNPs and 21 small indels. 347 of those either did not fall within genes (115) or resulted in nucleotide substitutions that did not affect amino acid sequence (232). One hundred and forty-five mutations produced a single amino acid change, two deleted one and six amino acids (without frame change), and three produced stop codons. The complete list of identified polymorphisms can be found in Supplemental Table 1 (MB220_vs._BJ4741_polymorphisms_all). Mutations producing stop codons, their genomic coordinates and affected genes are: polymorphism 19, chr4 902499, RAD9 (DNA damage-dependent checkpoint protein); polymorphism 432, chr13 74415, YML099W-A (dubious open reading frame unlikely to encode a protein); polymorphism 471, chr15 286230, TAT2 (high-affinity tryptophan and tyrosine permease).
This work was supported by Public Health Service, NIGMS 078666 and Army Research Office (ARO) 09-1-0041.