2.1.1The Yku70/80 complex
The KU70/80 heterodimer is one of the most abundant DNA end-binding proteins in mammalian cells. It was originally identified as an autoantigen in patients with polymyositis-scleroderma overlap syndrome  and is now known to bind to a variety of discontinuities in double-stranded DNA (dsDNA), including single-stranded gaps and bubbles of non-complementarity. Its highest affinity, however, is to blunt, 5′ or 3′ overhanging and hairpin dsDNA ends [51,52]. At the site of DSB, KU70/80 acts as a bridging complex. In the end-to-end fusion process, KU70/80 functions as an alignment, recruitment and stimulating factor (Table 1) [53–57].
A yeast homologue of the KU70/80 autoantigen, Yku70/80 (also referred to as Hdf1/Hdf2) has been identified (Table 2) [35,37,39,58]. The amino acid sequence of Yku70 predicts a 70.647 kDa protein that shares significant homology with the human KU70 protein . Similarly, Yku80, having a molecular weight of 71.240 kDa, displays a high degree of homology with its human counterpart . As expected, the Yku70/80 heterodimer binds specifically to DNA ends in a sequence-independent manner [58,59], a function analogous to its mammalian equivalent.
Table 2. The Saccharomyces cerevisiae NHEJ factors and phenotype of the corresponding mutants
|Yeast factor||Mammalian homologue||Mutant phenotype|
|Yku70/80||KU70/80||Sensitivity to bleomycin and MMS|
| || ||No UV and HU sensitivity|
| || ||IR sensitivity observed only when coupled with a rad52 mutation|
| || ||Normal IR-induced cell cycle arrest|
| || ||Normal DSB repair after IR exposure at chromosomal level|
| || ||Mating-type switching and spontaneous mitotic recombination defect|
| || ||No growth defect at 30 °C|
| || ||No growth at 37 °C|
| || ||No gross chromosomal rearrangements|
| || ||Telomere length shortening|
| || ||No silencing at telomeres|
| || ||Defect in plasmid repair assay|
| || || |
|Mre11/Rad50/Xrs2||MRE11/RAD50/NBS1||Sensitivity to many DSB-inducing agents|
| || ||Impaired meiosis|
| || ||Reduced integration of transforming DNA|
| || ||Telomere length shortening|
| || ||No silencing at telomeres|
| || ||Poor mitotic growth|
| || ||Defective checkpoint activation|
| || ||Defect in plasmid repair assay and in assay that measures NHEJ at the chromosomal level|
| || ||Elevated rates of spontaneous mitotic recombination|
| || ||Defects in HR|
| || ||Premature senescence|
| || ||Gross chromosomal rearrangements|
| || || |
|Lig4/Lif1||DNA ligase IV/XRCC4||No sensitivity to IR and MMS of dividing cells; a slight sensitivity to these agents of non-dividing cells|
| || ||No obvious growth defects at 30 °C|
| || ||Growth at 37 °C could be affected (for details, see Section 2.3)|
| || ||No telomere length shortening|
| || ||No gross chromosomal rearrangements|
| || ||Meiosis could be affected (for details, see Section 2.3)|
| || ||Defect in plasmid repair assay|
| || || |
|Nej1||–||Defect in plasmid repair assay and in an assay that measures NHEJ at the chromosomal level|
| || ||No UV and MMS sensitivity|
| || ||No growth defect at 30 and 37 °C|
| || ||No defect in postdiauxic/stationary-phase stimulation of NHEJ|
| || || |
|Pol4||DNA polymerase μ or λ||No UV, MMS, EMS, MNNG, IR and H2O2 sensitivity, although a weak sensitivity to MMS and IR was observed in another study (see text for details)|
| || ||No mitotic growth defect|
| || ||No sporulation nor spore viability defect|
| || ||Normal levels of spontaneous mitotic recombination|
| || ||Increased levels of meiotic recombination|
| || ||Normal levels of UV-induced recombination and mutagenesis|
| || ||No chromosome loss|
| || ||Increased frequency of illegitimate mating|
| || || |
|Rad27||FEN-1||Increased sensitivity to MMS, EMS, MNU and ENU|
| || ||Moderate UV, H2O2 and IR sensitivity|
| || ||Increased spontaneous and UV-induced mutagenesis|
| || ||Increased plasmid loss|
| || ||Elevated short-sequence recombination|
| || ||No growth defects|
| || ||Temperature-sensitive phenotype|
| || ||Premature aging|
| || ||Instability of telomeric repeats|
| || ||Defects in Ty1 mobility|
| || ||Gross chromosomal rearrangements|
| || ||Viability dependent on HR proteins, but not on those involved in NHEJ|
Disruption of either the YKU70 or YKU80 gene affects mating-type switching and spontaneous mitotic recombination . In addition, it leads to sensitivity to bleomycin [38,39] and methyl methanesulfonate (MMS) [39,58], agents which lead directly or indirectly to the induction of DSB . Since yeast employs HR to repair DSB under most circumstances, loss of Yku70 and Yku80 activity significantly hypersensitizes yeast cells to ionising radiation (IR) only when the HR is debilitated [37,59]. Consistent with this are the findings showing normal DSB repair and cell cycle arrest after IR exposure in the yku70 mutant strain . Similarly to the IR sensitivity, the MMS sensitivity of the rad52 mutant strain is elevated significantly when either the YKU70 or YKU80 genes are inactivated. The hypersensitive phenotype of the yku70 rad52 double mutant was also observed after exposure to bleomycin . In contrast, inactivation of YKU70 or YKU80 does not result in any detectable increase in sensitivity towards agents such as ultraviolet (UV) light  and hydroxyurea (HU) .
The yku70 or yku80 disruption mutant also shows a temperature-sensitive phenotype for growth at 37 °C [35–37,59,61]. Interestingly, strains lacking Yku70/80 do not die immediately upon transfer to the restrictive temperature but continue dividing for several generations . After several hours, cells arrest growth and appear as enlarged single-budded cells with abnormally high DNA content, indicating a defect in the regulation of DNA replication coupled with, or causing, an arrest in G2 phase of the cell cycle [35,61]. The yku70 yku80 double mutant displays no additional growth defects. This temperature sensitivity cannot be complemented by the expression of either the single subunits or the human KU70/80 heterodimer. Yku70/80 or KU70/80 corresponding DNA binding activity is not detectable in the yku70- or yku80-deficient strains transformed with plasmids expressing human KU70 or KU80, respectively. Thus, Yku70 and KU80 or Yku80 and KU70 cannot form functional heterodimers. Furthermore, at the permissive temperature, the yku70 or yku80 mutant strain has greatly shortened telomeres, corresponding to loss of 65% of the C1–3A terminal telomeric repeat sequences [37,62–64]. Strikingly, whereas telomere length is not affected when wild type strains are incubated at 37 °C, the transfer of the yku70 or yku80 mutant strains to 37 °C leads to a further dramatic loss of telomeric repeats. Therefore, it appears that the death of Yku70/80-deficient yeast strains at 37 °C is a consequence of the loss of telomeric repeats .
Since cells expressing a functional Yku70/80 can precisely join cohesive ends of a transformed linearized plasmid, while cells deficient for one of the Yku70/80 subunits display reduced recircularization efficiency and an increased frequency of imprecisely joined products, Yku70/80 is considered to be an essential part of the S. cerevisiae NHEJ [36,37,58,64]. The recircularization efficiency (a measure of the ability of yeast cells to repair restriction enzyme generated DSB in vivo) can be established using the transformation-based plasmid repair assay. In this assay, a S. cerevisiae strain is transformed with a yeast –E. coli shuttle plasmid that has been linearized by treatment with a restriction enzyme. To normalize for differences in transformation efficiency between strains and between repeats of the same experiment, a supercoiled version of the same plasmid is transformed into the yeast strain in parallel. Since the plasmid must be recircularized in order to be propagated, the number of transformants obtained with the linear plasmid normalized to the number obtained with the supercoiled plasmid provides a quantitation of the ability of the yeast strain to mediate repair of the restriction enzyme-generated DSB. To prevent the DSB from becoming repaired by HR with the yeast genome, the sites for restriction enzyme cleavage of the plasmid are within regions that are not homologous to chromosomal sequences. DSB with cohesive ends are repaired with high efficiency in wild type yeast strains in that transformant yields are over 70% of the values obtained with supercoiled plasmid. In marked contrast, strains debilitated in YKU70 or YKU80 show a dramatic 40–100-fold decrease in transformants recovered with linearized DNA. This is not affected by the presence or absence of Rad52. Expectedly, strains mutated in both YKU70 and YKU80 are no more impaired in plasmid repair than strains mutated for either YKU70 or YKU80 alone. Additionally, in Yku70/80-deficient cells, plasmids are not repaired accurately but in an error-prone way, yielding molecules that underwent losses of up to several hundred base pairs at the joining site, with junctional overlapping sequences of 3–15 bp. Surprisingly, Yku70/80 does not play a positive role in the rejoining of plasmid molecules bearing blunt DNA termini .
As mentioned above, inactivation of YKU70[62,64] or YKU80 leads to telomeric shortening, showing that Yku70/80 plays a crucial role in telomere length maintenance. Moreover, Yku70/80 represses the transcription of RNA polymerase II genes in close proximity to telomeres, a process called telomere position effect (TPE) . In these end-regions of chromosomes, the chromatin is in a unique condensed structure that does not permit access by the transcriptional apparatus and the introduced gene is, therefore, silent [67,68]. If the telomeric chromatin structure is then disrupted, telomeric silencing is relieved, allowing the gene to be expressed, and telomeric silencing is lost [65,69,70]. TPE is severely diminished in Yku70/80-deficient cells, although the repression of the silent mating type loci, at an internal chromosome site, is maintained normally in the same cells [65,70,71], indicating that silencing in general is not affected . The association of Yku70/80 with telomeric silencing is in accord with the fact that Yku70/80 binds directly to telomeric DNA [71,72] and that Yku70 interacts with Sir4 . Yku80 also interacts with Sir4 and this interaction is mediated by the C- and N-terminal regions of Yku80 and Sir4, respectively . Sir4 is constitutively bound to Sir2, forming a Sir2/Sir4 complex, but the Sir3 association with this complex is prevented by the intramolecular interaction within Sir4 . At telomeres, Yku80 interacts with the N-terminus of Sir4, thereby inducing a conformational change within Sir4 that facilitates the recruitment of Sir3 to the Sir4 C-terminus . Therefore, it appears that Yku70/80 following binding to telomeric DNA ends and, through its interaction with Sir4, helps to recruit the Sir2/Sir3/Sir4 complex to the telomere [65,73]. Moreover, the Yku80/Sir4 interaction is likely to play a vital role in the assembly of telomeric heterochromatin, and thus in the establishment of TPE . The association of Yku70/80 with TPE is consistent with studies in vertebrate systems, which has revealed that KU70/80 is a potent inhibitor of transcription [76,77].
The dual requirement of Yku70/80 at site of DSB, where it promotes end-to-end fusion, and at telomeres, which are specifically protected from end-joining, presents a paradox. A possible resolution is that Yku70/80 performs different activities at these two different classes of DNA ends. Moreover, Yku70/80 has probably multiple separable functions at the telomere. It may be that Yku70/80 associates with subtelomeric chromatin, where it influences the formation of heterochromatin. Independently, Yku70/80 probably associates with the chromatin terminus, where it mediates telomere length regulation via interactions with telomerase and telomere end protection via inhibition of an end-processing activity . However, prior support for separable functions for Yku70/80 has come from the identification of the C-terminal truncation mutants. One mutation in Yku70, a deletion of 30 C-terminal amino acids (Yku70-c30), that abolishes DNA-binding activity, causes a Yku70 deletion phenotype; telomeres are very short, displaying G-tails, and NHEJ is not functional. On the other hand, when the C-terminal 25 amino acids of the Yku70 protein are deleted (Yku70-c25), DNA-binding capacity is indistinguishable from the wild type protein and NHEJ is fully functional. However, these same cells still display shortened telomeres and clearly detectable ssDNA overhangs of the 3′-ends. Taken together, deletion of only 5 amino acids more from the Yku70-c25 construct completely abolishes DNA-binding of the Yku70/80 complex and renders it non-functional. Thus, the extreme C-terminal domain of Yku70 is probably specifically involved in maintaining telomere integrity, but not in DNA-binding or end-joining activity. Interestingly, the terminal 25 amino acids of Yku70 contain 8 lysine residues. Deletion of 3 of these lysine residues in Yku70 with 9 amino acids deleted causes a weak decrease in telomere length when compared with wild type cells. This decrease is more pronounced in Yku70 with deletion of 20 amino acids, where 6 of the 8 lysine residues are deleted. All 8 lysine residues are removed in the Yku70-c25 mutant, and shortening of telomeres is almost as severe as observed in the yku70-deficient strain. Thus, this lysine-rich domain at the C-terminus may be important for interacting with yet unidentified protein(s) .
As stated above, Yku70/80 is associated with the ends of chromosomes in vivo [71,78,79]. Since there is enhanced degradation of broken chromosomes in the absence of Yku70/80, a role for Yku70/80 in protection of DNA ends from nucleolytic processing was proposed . Such a role is in line with crystal structure of KU70/80 (Table 1), since this heterodimer encircles the duplex DNA like the thread of a screw . Furthermore, Yku70/80 appears to be also involved in the clustering of yeast telomeres at peripheral sites in the nucleus because mutations in the YKU70 or YKU80 gene affect the subnuclear organization of yeast telomeres and thus, abolish the clustered distribution of telomeric foci. In wild type diploid yeast cells, the 64 telomeres are usually found in 6 or 7 clusters around the nuclear periphery, whereas cells mutated in either Yku70/80 subunit have around 9 clusters that seem to be located more randomly throughout the whole nucleus . Taken together, it appears that Yku70/80 is somehow involved in clustering the telomeres of several chromosomes and tethering them at sites in the nuclear periphery .
The TEL1 gene encodes a protein that shares homology with phosphatidyl-inositol-3 family of protein kinases and therefore could represent a DNA-PKcs equivalent in S. cerevisiae. Interaction between Tel1 and Yku70/80 was examined by comparing the phenotype of the tel1 yku70 or tel1 yku80 double mutant with the phenotype of the respective single mutants. Such analysis showed different roles of Tel1 and Yku70/80 in several cellular processes [39,62,65]. Moreover, the phenotype of the yku70 or yku80 mutant significantly differs from that of the tel1 mutant for TPE, repair of plasmid-based DSB, and growth and further dramatic loss of telomeric repeats at 37 °C , strengthening the notion that Tel1 and Yku70/80 operate independently from one another and, accordingly, that Tel1 is not an essential component of the S. cerevisiae NHEJ and does not represent a functional homologue of the DNA-PKcs in this organism. Indeed, Tel1 is probably the S. cerevisiae ATM homologue (for further details, see Section 5) .
2.1.2The Mre11/Rad50/Xrs2 complex
In contrast to mammalian systems, where satisfactory evidence for the involvement of the MRE11/RAD50/NBS1 complex in NHEJ is still missing, the genetic and biochemical studies carried out in S. cerevisiae unequivocally implicate the Mre11/Rad50/Xrs2 complex, a yeast homologue of MRE11/RAD50/NBS1, in NHEJ [58,65,84–86]. All individual components of the Mre11/Rad50/Xrs2 complex were cloned and shown to encode 692, 1312 and 854 amino acid proteins with a molecular weight of 77.650, 152.568 and 96.364 kDa, respectively [87–90]. By comparing their sequences with those of mammals, it has been revealed that while human MRE11 and RAD50 represent the clear structural and functional homologues of their yeast counterparts, NBS1 is a functional rather than a structural analogue of Xrs2 [91–95].
Stoichiometry of the purified Mre11/Rad50/Xrs2 complex is 2:2:1 . The complex possesses Mn2+-dependent 3′ to 5′ dsDNA and ssDNA endonuclease, ssDNA exonuclease and hairpin cleavage activities, all of which are specified by four phosphoesterase motifs residing in the N-terminal part of Mre11 [64,86,96–99]. Moreover, the complex binds and hydrolyzes ATP via Rad50 [100,101]. ATP binding and hydrolysis was shown to enhance nuclease activity of the complex in vitro  in agreement with in vivo findings that a mutant allele of the RAD50 gene that prevents ATP binding confers a null phenotype with respect to DNA repair . Finally, Mre11/Rad50/Xrs2 possesses a DNA-binding activity, which is, interestingly, exhibited by all the individual subunits separately [96,98,99,102].
It has been shown that Rad50 contains Walker A and B motifs in the N- and C-terminal parts of the protein, respectively, which are responsible for ATP binding and hydrolysis. These motifs are separated by a long coiled-coil region, in the center of which, a hinge motif possessing a sequence Cys–X–X–Cys (CXXC), is present. Crystal structure of Rad50 reveals that CXXC motif promotes Zn2+-dependent dimerization of Rad50, as the two CXXC motifs form interlocking hooks that bind one Zn2+ ion. The functional importance of the CXXC motif for the whole complex assembly and DSB repair has clearly been demonstrated in vivo using strains that were mutated individually in the two cysteine residues. These strains were IR sensitive and display impaired Mre11-binding, demonstrating that the structure of the Rad50 dimer has consequences for the structure and biological function of whole complex. Coiled-coil regions extend from the Zn2+-binding site in opposite directions and this allows them to link two distinct Mre11 dimers. The Mre11 dimer binds to the coiled-coils of two Rad50 molecules adjacent to the ATPase domain, forming a globular head, which is responsible for the DNA-binding and/or DNA end-processing activities of the complex. Hence, one Mre11/Rad50 heterodimer possesses two DNA-binding and end-processing active sites. These two active sites could bind two separate broken DNA ends at a time and, subsequently, could act as a bridging factor during NHEJ [103–108] (see text below for further details).
Participation of the Mre11/Rad50/Xrs2 complex in NHEJ in the budding yeast stemmed from the findings that showed 100-fold reduced frequency of integration of transforming DNA that shares no homology with the host genome in the rad50 mutant . Moreover, monocentric  and dicentric  plasmid systems developed for quantitative analysis of a deletion formation as a consequence of NHEJ disclosed 10- and 50-fold defects, respectively, in strains debilitated in Mre11/Rad50/Xrs2. Further in vivo experiments [58,65,84], exploiting an HO endonuclease-induced chromosomal DSB assay and a transformation-based plasmid repair assay (see Section 2.1.1), definitively established a role of the Mre11/Rad50/Xrs2 complex in the S. cerevisiae NHEJ. Plasmid repair assays showed that strains, from which MRE11, RAD50, or XRS2 was deleted, exhibited a dramatic (up to 40-fold) drop in NHEJ compared to the wild type strain [58,64,65] and that the mre11 yku70, rad50 yku70, xrs2 yku70, mre11 lig4, rad50 lig4 or xrs2 lig4 double mutants were not appreciably more debilitated in NHEJ than were the single mutants [64,65]. These data therefore provided clear in vivo evidence for the involvement of Mre11/Rad50/Xrs2 in NHEJ. Moreover, they indicated that Mre11/Rad50/Xrs2 functions epistatically with Yku70/80 and Lig4/Lif1 in NHEJ. When the nature of the residual plasmid repair events taking place in the mre11, rad50 and xrs2 mutants was analyzed, it was found that virtually all NHEJ products arising in these mutants were accurate . This is in sharp contrast with the situation for Yku70/80-deleted strains (see Section 2.1.1), indicating different roles for Yku70/80 and Mre11/Rad50/Xrs2 in NHEJ. Furthermore, the HO endonuclease-induced chromosomal DSB assay has revealed that there are distinct mechanisms of NHEJ producing either insertions or deletions and that these two pathways are differently affected by cell cycle stage when HO is expressed. This assay uses MATα strain, which lacks HML and HMR donor sequences and carries the URA3-based plasmid pGAL-HO, in which the HO gene is under the control of a galactose-inducible promoter. Upon galactose induction, a high level of HO expression and subsequent HO cleavage throughout the cell cycle can be achieved. Importantly, DSB repair in MATα locus by HR is prevented in this strain because of the HML and HMR deletions, so that virtually all repair events scored are due to either insertions or deletions arising during NHEJ. As demonstrated, the overall NHEJ efficiency of this strain was dramatically reduced (up to 70-fold) by inactivation of the RAD50, MRE11 or XRS2 genes. Moreover, inactivation of these genes markedly reduced insertions, while significantly increasing the proportion of deletions . Thus, the Mre11/Rad50/Xrs2 complex exerts an important role in the insertion-producing pathway of NHEJ and this tempts speculation that there might be a physical link between synapsis and gap-filling in NHEJ (for details on the NHEJ gap-filling factors, see text below).
In vivo NHEJ assays mentioned above [65,84,85,110] suggested that the nuclease activity of the Mre11/Rad50/Xrs2 complex may not play a role in NHEJ and that the function of this complex may rather be to bring together broken DNA ends and/or to recruit other NHEJ factors. To support this, it has been shown that two strains, each expressing different nuclease-defective Mre11 protein, were not deficient in the plasmid repair assay [64,96] and that overexpression of Exo1, 5′ to 3′ exonuclease exhibiting also flap endonuclease activity, failed to rescue NHEJ deficiency of an mre11 mutant . Furthermore, recent biochemical data  strongly supported this notion by demonstrating that Mre11/Rad50/Xrs2 serves as an end-bridging factor in NHEJ that links DNA ends and provides the scaffold upon which the NHEJ machinery is assembled. No nuclease activity of Mre11/Rad50/Xrs2 was required for this function. In vitro experiments further showed that Mre11/Rad50/Xrs2 interacts with (interaction between Mre11/Rad50/Xrs2 and Lig4/Lif1 is mediated via Xrs2 and Lif1, respectively) and specifically stimulates intermolecular ligation by Lig4/Lif1, a process that is additionally stimulated by Yku70/80.
As is the case with strains deficient in Yku70/80, those debilitated in Mre11/Rad50/Xrs2 also suffer dramatic telomeric attrition, losing approximately 65% of the terminal repeats at permissive temperature. Since no further telomere shortening was observed in the mre11 yku70, rad50 yku70 or xrs2 yku70 double mutants under these conditions, Yku70/80 and Mre11/Rad50/Xrs2 function epistatically in telomere length maintenance [64,65]. Notably, shifting of the yku70 or yku80 mutant strain from permissive to restrictive temperature led to a further dramatic loss of telomeric repeats, presumably representing the mechanism by which these mutant strains die at 37 °C (see Section 2.1.1). Surprisingly, a further striking reduction in telomere length upon transfer of strains disrupted for any of the Mre11/Rad50/Xrs2 subunit to 37 °C could be detected only in the rad50 mutant, which correlates with severely impaired growth of this mutant at the restrictive temperature . The reason why the mre11 and xrs2 mutants do not show the same growth defects and telomere shortening at the restrictive temperature is unknown, but this might indicate that there are also separate functions for each individual subunit of Mre11/Rad50/Xrs2. Similarly, the lack of observable NHEJ defects displayed by the nuclease-defective mre11 mutants (see text above) was paralleled by the lack of any telomere shortening in these mutants , although the other did have impaired telomere length maintenance . The basis of this discrepancy is unknown and calls for further examination. As already mentioned, another phenotype related to telomeric DNA is transcriptional silencing of the genes in close proximity to telomeres, a process called TPE . This process has been markedly diminished in the yku70 or yku80 mutants [65,70,71] (see Section 2.1.1). Astonishingly, TPE analysis in the mre11, rad50, or xrs2 mutants showed no essential role of Mre11/Rad50/Xrs2 in this process . This suggests partially separate functions of the Yku70/80 and Mre11/Rad50/Xrs2 complexes at the telomeres and provides evidence for a different composition of the protein complexes operating at these regions of the S. cerevisiae genome.
In addition to their abnormal phenotypes for NHEJ or telomere metabolism, cells debilitated in Mre11/Rad50/Xrs2 are defective in meiotic recombination, display premature senescence, suppression of gross chromosomal rearrangements, elevated rates of spontaneous mitotic recombination, and delayed mating-type switching ([64,65,89,96–98,112–116], reviewed in ) (Table 2). Importantly, they also display a remarkable defect in HR, the main DSB repair mechanism in the budding yeast. Part of the difficulty in identifying the role of Mre11/Rad50/Xrs2 in the particular DSB repair mechanism arises from the fact that several processes can affect the efficiency of HR and NHEJ in vivo, for example the loss of checkpoint activity in the budding yeast resembles the loss of the MRE11/RAD50/XRS2 complex with regard to its effects on NHEJ. This raises the possibility that inactivation of the Mre11/Rad50/Xrs2 complex affects HR and NHEJ, at least in part, through defects in checkpoint functions [117–119].