Genetic Analysis of the Hsm3 Protein Function in Yeast Saccharomyces cerevisiae NuB4 Complex

In the nuclear compartment of yeast, NuB4 core complex consists of three proteins, Hat1, Hat2, and Hif1, and interacts with a number of other factors. In particular, it was shown that NuB4 complex physically interacts with Hsm3p. Early we demonstrated that the gene HSM3 participates in the control of replicative and reparative spontaneous mutagenesis, and that hsm3Δ mutants increase the frequency of mutations induced by different mutagens. It was previously believed that the HSM3 gene controlled only some minor repair processes in the cell, but later it was suggested that it had a chaperone function with its participation in proteasome assembly. In this work, we analyzed the properties of three hsm3Δ, hif1Δ, and hat1Δ mutants. The results obtained showed that the Hsm3 protein may be a functional subunit of NuB4 complex. It has been shown that hsm3- and hif1-dependent UV-induced mutagenesis is completely suppressed by inactivation of the Polη polymerase. We showed a significant role of Polη for hsm3-dependent mutagenesis at non-bipyrimidine sites (NBP sites). The efficiency of expression of RNR (RiboNucleotid Reducase) genes after UV irradiation in hsm3Δ and hif1Δ mutants was several times lower than in wild-type cells. Thus, we have presented evidence that significant increase in the dNTP levels suppress hsm3- and hif1-dependent mutagenesis and Polη is responsible for hsm3- and hif1-dependent mutagenesis.


Introduction
The chromatin of eukaryotic cells consists of nucleosome core particles containing a histone octamer core wrapped approximately two times with 147 bp of DNA. The histone core comprises a heterotetramer of two copies each of the histones H3 and H4 and two heterodimers of H2A and H2B [1]. Histones are highly charged basic proteins, which bind with chaperones that prevent them from interacting nonspecifically with other proteins and DNA and help regulate their proper deposition into nucleosomes [2].
The chromatin assembly of a genome imposes limitations on many cellular processes that require accessibility to chromosomal DNA. The posttranslational acetylation of the histone N-terminal tails has been shown to be an important mechanism by which cells regulate accessibility to chromatin [3]. The tail acetylation of newly synthesized H3 and H4 molecules is a transient modification. This modification changes both the charge and structure of lysine residues and is catalyzed by histone acetyltransferases.
Hat1 is most often represented in the nucleus of a yeast cell as a NuB4 complex that contains Hat1p, Hat2p, and Hif1p [16,17]. Hat2p possess histone chaperone activity and, therefore, are thought to mediate the interactions of these varied NuB4 complexes with histones [17][18][19][20][21]. Hif1p is a member of the N1 family of histone chaperones and specifically interacts with histones H3 and H4. Hif1p can participate in the deposition of histones onto DNA, suggesting that Hat1p may be directly involved in the chromatin assembly process [16]. It was shown that NuB4 complex physically interacts with Hsm3p [22][23][24]. It was shown earlier that the gene HSM3 participates in the control of replicative and reparative spontaneous mutagenesis, and that hsm3∆ mutants increase the frequency of mutations induced by different mutagens [25][26][27][28][29][30][31]. It was previously believed that the HSM3 gene controlled only some minor repair processes in the cell, but later it was suggested that it had a chaperone function with its participation in proteasome assembly [32][33][34]. This was confirmed quite recently: Takagi et al., 2012 [35] succeeded in establishing the spatial structure of the Hsm3 protein, as well as showing the interaction of this protein with the subunits of the proteasome complex. It means that that the product of this gene may have more than one functional domain. Analysis of various mutant alleles of the HSM3 gene revealed that the C-terminal domain of the Hsm3 protein is responsible only for controlling induced and spontaneous mutagenesis and does not affect proteasome assembly [36].
To determine possible role of the HSM3 and HIF1 genes in HAT1 complex function, we conducted genetic analysis of properties of three hsm3∆, hif1∆, and hat1∆ mutants. Besides, our findings show that Polη is responsible for hsm3and hif1-dependent mutagenesis.
Media: Standard yeast media of complete and minimal composition were used in the work [39]. In some experiments, a liquid YPD was used without the addition of agar. When working with auxotrophic mutants, metabolites required for growth were added to the minimal medium at a rate of 20 mg/L for amino acids and 3 mg/L for nitrogen bases. As a selective medium for accounting for the frequency of canavanine resistance mutations, a minimal medium was used with the addition of a liquid YPD in an amount of 10 mL/L and required for the growth of amino acids and nitrogenous bases. Depending on the strains used, canavanine concentrations were up to 80 mg/L. Taking into account the frequency of induced mutations at five loci, YPD with an alcohol instead of glucose was used, the composition of which was described earlier [37].
Sensitivity against UV irradiation: Cell killing tests were performed on plates by growing overnight a culture of the respective strain in liquid YPD at 30 • C. Cells were washed and resuspended in water at a density of 1 × 10 7 cell/mL. Cells were irradiated with a UV lamp BUV-30 (UV-C range). Aliquots were withdrawn at different times, diluted, and plated onto YPD plates to determine the number of survivors.
Mutation frequency: Mutation tests were performed on plates by growing overnight a culture of the respective strain in liquid YPD at 30 • C. Cells were washed and resuspended in water at a density of 1 × 10 7 cell/mL. Cells were irradiated with a UV lamp BUF-30. Aliquots were withdrawn at different times, diluted, and plated onto YPD plates to determine the number of survivors. To determine the mutation frequency, undiluted aliquots were plated onto a medium YPD with an alcohol instead of glucose, the composition of which was described earlier [37].
Mutation rates: Mutation rates were determined according to the methods: fluctuation test [40] and ordered seeding [41]. The first method allows the determination of the rate of spontaneous yeast cell mutations in the process of fast growth on complete medium. After incubation for three days, 12 separate colonies were scored, and each colony was suspended in 1 mL of water and plated on selective medium with canavanine, at a concentration that rules out the possibility of growth of canavanine-sensitive cells. When estimating the number of plated cells, we diluted suspensions and plated them on complete medium. After incubation for three or four days, the number of canavanine-resistant colonies and the total number of cells on the plate were counted. The occurrence frequency of spontaneous mutations was estimated using the special formula [32].
Using the method of perfect order plating, one can register the frequency of spontaneous mutations arising in the process of slow growth on selective medium containing lower concentrations of canavanine in which cells are grown over 8 to 10 divisions. Cells were incubated in 2 mL of complete liquid medium for 2 days; next, 1 mL of grown culture was diluted in 5 mL of water. A special replicator having 150 appliances was embedded into the suspension, and drops were placed onto plates with selective medium. After a14-day incubation, the number of canavanine-resistant colonies and the total number of cells grown in 150 spots was counted. The number of grown non-mutant cells was determined after washing away cells from individual reprints on which no canavanine-resistant colony was visualized. The rate of mutation was determined by dividing the number of canavanine-resistant colonies by the number of cells in all reprints [33].
In total, five replicates of the experiment are shown on the graphs and in the tables, and the mean values with 95% confidence intervals are given.
PCR cycling conditions were as follows: 1 cycle of 5 min at 95 • C, followed by 39 cycles of 15 s at 95 • C and 20 s at 52 • C. Melting curve analysis was 5 s incremental increases of 1 • C from 55 • C to 95 • C.
Control reactions with primer and template free reaction mixtures were included. Two biological and three technical replicates were performed for each sample. The results were processed using the CFX Manager program (Bio-Rad, Watford, UK).
Statistical analysis: Experimental data are shown as the means standard deviations from at least three biological replicates, and statistical differences were determined by the Student's t-test. Significance was determined at the level of p < 0.05.

Results
The physical interaction between the products of the genes HAT1, HIF1, and HSM3 has been demonstrated in previous studies [22][23][24]. We suppose that this interaction plays a functional role and Hsm3 protein is a subunit of NuB4 complex. In this work, we conducted a comparative study of the genetic properties of hif1∆ and hsm3∆ mutations and their interaction with a deletion mutation in the HAT1 gene, which codes for the catalytic subunit of the complex. HSM3 and HIF1: It has been shown that the mutants for the HAT1 gene affect the repair processes in yeast [42]. Therefore, we first tested spontaneous mortality of the single hif1∆ and hsm3∆ mutants. The percentage of lethal clones was assessed after a day of incubation in solid complete medium. Clones with fewer than 16 cells were considered to be lethal. In the test for spontaneous mortality, both mutants showed the same pattern. These mutants did not change significantly the rate of spontaneous death of cells compared to the wild-type strain ( Table 2). According to our research and the data presented in work [38], in the double hif1∆ rad52∆ and hsm3∆ rad52∆ mutants, hif1∆ and hsm3∆ mutants were equally epistatized to recombination-deficient mutants. These data support the hypothesis advanced earlier that hsm3∆ mutation (maybe hif1∆) leads to the destabilization of the D-loop during the post-replicative repair [38,43] and thus reduces the load on the path of the recombination repair, which is blocked in rad52∆ mutant. Table 2. Rate of spontaneous lethal clones after 1 day of incubation on a dense nutrient medium.

Strains
Rate of Spontaneous Lethal Clones (%) To further test the survival of hsm3∆ and hif1∆ mutants, we performed experiments with UV light. We found that both single mutants showed the same sensitivity to UV as a wild-type strain ( Figure 1A). The double hsm3∆ hif1∆ mutant practically did not differ in this parameter from single mutants ( Figure 2A). Thus, mutations in the HIF1 and HSM3 genes have no significant effect on the survival of yeast cells after UV irradiation.
In order to study the role hsm3∆ and hif1∆ mutants in the DNA damage response, we were constructed double mutant strains with genes from the major DNA repair pathways. Nucleotide excision repair-deficient strain, rad2∆, was used for RAD3 epistasis group, while rad52∆ was chosen for the recombination repair epistasis group. The hif1∆ mutation, like the hsm3∆ mutation, in combination with rad2∆ did not lead to a change in the UV resistance of the double mutant hif1∆ rad2∆ cells compared to the cells of a single rad2∆ mutant ( Figure 1C,D). Based on the data obtained, it can be concluded that hsm3∆ and hif1∆ mutations do not affect NER. rad54∆ and rad52∆ mutants are weakly sensitive to UV, which corresponds to the previously obtained data [38]. Earlier, we have shown that hsm3∆ mutation does not change the UV sensitivity of rad52∆ mutant [38]. Double hif1∆ rad52∆ mutant had the same sensitivity as single radiosensitive mutant ( Figure 1B). Taken together, hsm3∆ and hif1∆ mutations did not affect the radiosensitivity of yeast cells.
It has been shown that the HSM3 gene participates in the control of replicative and reparative spontaneous mutagenesis [29]. Therefore, we set out to compare the spontaneous mutagenesis of the hif1∆ mutant to that of the hsm3∆ mutant. When studying the rate of occurrence of spontaneous mutations, we used two different methods: the fluctuation test (the "Coulson-Lee median" method), which that allows to determine level of spontaneous replication mutagenesis, and the method of ordered seeding, which can be judged on the level of reparative mutagenesis. In both cases, the frequency of mutations in the CAN1 gene was calculated. From the data in Table 3 it is seen that that the presence of hsm3∆ and hif1∆ mutations in cells leads to a slight change in the rate of replicative spontaneous mutations compared to wild-type cells. In the ordered seeding test, both mutations led to a sharp increase in the rate of spontaneous mutagenesis. hsm3∆ and hif1∆ mutations increased the rate of spontaneous reparative mutagenesis in this test by approximately 18 times. These results support the notion that hsm3∆ and hif1∆ have the same phenotype.  In order to study the role hsm3Δ and hif1Δ mutants in the DNA damage response, we were constructed double mutant strains with genes from the major DNA repair pathways. Nucleotide excision repair-deficient strain, rad2Δ, was used for RAD3 epistasis group, while rad52Δ was chosen for the recombination repair epistasis group. The hif1Δ mutation, like the hsm3Δ mutation, in combination with rad2Δ did not lead to a change in the UV resistance of the double mutant hif1Δ rad2Δ cells compared to the cells of a single  Earlier, we studied in detail the effect of hsm3∆ mutation on UV-induced mutagenesis [29]. As shown previously, hsm3∆ mutation significantly increased the induced mutagenesis when exposed to various mutagenic factors. As shown in Figure 2A, the hif1∆ mutation, like hsm3∆ mutation, significantly increased the frequency of UV-induced mutagenesis. Unexpectedly, the double hsm3∆ hif1∆ mutant showed the level of UV-induced mutagenesis peculiar to wild-type strain, as well as a double mutant hsm3-1 hif1∆ mutant with a point mutation of the C-terminal domain of the Hsm3p. (Figure 2A). Hence the loss of both the Hif1p and Hsm3p results in increased UV-induced mutagenesis; inactivation of both proteins leads to a phenotype corresponding to the wild-type cells.
In all previous experiments, we used an unsynchronized cell culture. As it is known, in such a culture there is a mixture of cells in the G1, S, and G2 phases. The repair of UV-induced damage occurs by the mechanism of the nucleotide excision repair (NER) at all three stages of the cell cycle. In NER-defective strains, damage repair will be performed in the S and G2 phases by the process of postreplicative and recombination repairs. To assess the effect of mutations in the genes encoding the subunits of the HAT1 complex, we have disrupted the key gene that provides NER in the single mutants mentioned above. hsm3∆ mutation significantly increases the mutagenesis of rad2∆ mutant with broken nucleotide excision repair [37]. We tested how the mutation in the HIF1 gene, which codes for another subunit of the NuB4 complex, affects UV mutagenesis. As shown in Figure 1B, the interaction of rad2∆ and hif1∆ mutations has a synergistic character, as in the case of the interaction of hsm3∆ and rad2∆ mutations. Earlier, we had shown that rad52∆ mutation suppressed UV-induced mutagenesis in hsm3∆ significantly [38]. Therefore, we wished now to compare the UV-induced mutagenesis of hif1∆ mutant with that of rad52∆ mutant. rad52∆ mutation suppressed the UV-induced mutagenesis of hif1∆ mutant to wild-type strain level ( Figure 2B). In summary, hsm3∆ mutation may be considered a phenocopy of hif1∆ mutation.
However, the rad52∆ mutation suppresses hsm3and hif1-dependent UV-induced mutagenesis in different ways. In the hif1∆ mutant, the rad52∆ mutation suppresses the frequency of UV-induced mutations to the wild-type level, while in the hsm3∆ mutant the rad52∆ mutation slightly reduces the frequency of UV-induced mutations [38]. There was also a difference in the frequency of UV-induced mutagenesis in single mutants hsm3∆ and hif1∆ (Figure 2A). HAT1: Hat1p is the catalytic subunit of HAT complex. Therefore, we first tested for the genetic properties of hat1∆ mutant. The spontaneous mortality of single hat1∆ and double hat1∆ rad52∆ mutants compared to that of the wild-type strain. rad52∆ and hat1∆ single mutations increased the percentage of lethal clones to 10.1 ± 3.15 and 11.4 ± 3.97% respectively ( Table 2). Double mutants hat1∆ rad52∆ did not change the frequency of spontaneous cell death compared to single mutants (Table 2). Thus, there is an epistatic interaction of hat1∆ mutation with mutations that block the recombination repair. This conclusion is supported by data comparing the UV sensitivity of a single mutant hat1∆ with the UV sensitivity of double mutants hat1∆ rad52∆. Both strains, studied, showed the same UV sensitivity ( Figure 3A). Thus, in hat1∆ mutant, the recombination repair pathway of spontaneous and UV-induced lesions is destroyed.
Next, we tested the UV sensitivity of wild-type strain versus a single hat1Δ mutant. A single mutant showed a sensitivity to UV comparable to the sensitivity of a wild-type strain ( Figure 3B). Therefore, mutant hat1Δ does not affect NER. This conclusion is confirmed by the results obtained for the interaction of hat1Δ mutation and rad1Δ, which blocks NER. The double rad1Δ hat1Δ mutant showed UV sensitivity equal to a single rad1Δ mutant ( Figure 3C). Early the genetic effects of mutations in HAT1 gene have not been adequately studied. That is why we studied the effect of the hat1Δ mutation on the mutation process. From the data of Table 3, it can be seen that the presence of hat1Δ mutation results in a slight change in the rate of replicative spontaneous mutations compared to wild-type Next, we tested the UV sensitivity of wild-type strain versus a single hat1∆ mutant. A single mutant showed a sensitivity to UV comparable to the sensitivity of a wild-type strain ( Figure 3B). Therefore, mutant hat1∆ does not affect NER. This conclusion is confirmed by the results obtained for the interaction of hat1∆ mutation and rad1∆, which blocks NER. The double rad1∆ hat1∆ mutant showed UV sensitivity equal to a single rad1∆ mutant ( Figure 3C).
Early the genetic effects of mutations in HAT1 gene have not been adequately studied. That is why we studied the effect of the hat1∆ mutation on the mutation process. From the data of Table 3, it can be seen that the presence of hat1∆ mutation results in a slight change in the rate of replicative spontaneous mutations compared to wild-type cells. However, in the ordered seeding test, hat1∆ mutation leads to a sharp increase in the rate of spontaneous mutagenesis ( Table 3).
The frequency of direct mutations, in the loci of ADE4-ADE8, induced by UV rays in a wild-type strain and a single hat1∆ mutant was measured. The data presented in Figure 3B suggests that hat1∆ mutation does not affect the frequency of UV-induced mutagenesis.
As can be seen from Figure 3C single rad1∆ mutant and the double hat1∆ rad1∆ mutant have approximately the same level of mutagenesis. Single rad52∆ mutant shows greater UV-mutability compared to wild-type strain [38]. Double hat1∆ rad52∆ mutant at low doses showed the level of mutagenesis characteristic of a single hat1∆ mutant ( Figure 3A). Thus, in this test, hat1∆ mutation at low doses epistatizes to rad52∆ mutation.
It is known that the Hat1 subunit is catalytic in the HAT1 complex. In connection to this, we studied the epistatic interaction of hat1∆ mutation with mutations in genes, coding for other subunits of the complex. As shown in Table 3, the spontaneous mutation rates in double hat1∆ hsm3∆ and hat1∆ hif1∆ mutants do not differ from the spontaneous mutation rate in single hat1∆ mutant. These data corroborate previous results that hat1∆ epistatized to both single hsm3∆ and hif1∆ mutants. Figure 4 shows the dependence of the mutagenesis frequency on the dose of UV rays for single hat1∆, hsm3∆ and hif1∆ mutants, double hat1∆ hif1∆, hat1∆ hsm3∆, hsm3∆ hif1∆ mutants and triple hat1∆ hsm3∆ hif1∆ mutant. As can be seen from this figure, hat1∆ mutation epistatizes to all studied mutations.
cells. However, in the ordered seeding test, hat1Δ mutation leads to a sharp incre the rate of spontaneous mutagenesis ( Table 3).
The frequency of direct mutations, in the loci of ADE4-ADE8, induced by UV r a wild-type strain and a single hat1Δ mutant was measured. The data presented in F 3B suggests that hat1Δ mutation does not affect the frequency of UV-in mutagenesis.
As can be seen from Figure 3C single rad1Δ mutant and the double hat1Δ mutant have approximately the same level of mutagenesis. Single rad52Δ mutant greater UV-mutability compared to wild-type strain [38]. Double hat1Δ rad52Δ mu low doses showed the level of mutagenesis characteristic of a single hat1Δ mutant (F 3A). Thus, in this test, hat1Δ mutation at low doses epistatizes to rad52Δ mutation.
It is known that the Hat1 subunit is catalytic in the HAT1 complex. In connect this, we studied the epistatic interaction of hat1Δ mutation with mutations in coding for other subunits of the complex. As shown in Table 3, the spontaneous mu rates in double hat1Δ hsm3Δ and hat1Δ hif1Δ mutants do not differ from the sponta mutation rate in single hat1Δ mutant. These data corroborate previous results that epistatized to both single hsm3Δ and hif1Δ mutants. Figure 4 shows the dependence of the mutagenesis frequency on the dose of U for single hat1Δ, hsm3Δ and hif1Δ mutants, double hat1Δ hif1Δ, hat1Δ hsm3Δ, hsm3Δ mutants and triple hat1Δ hsm3Δ hif1Δ mutant. As can be seen from this figure, mutation epistatizes to all studied mutations. Taken together, our results argue that hsm3Δ and hif1Δ mutants have the phenotypes, and that hat1Δ mutation epistatizes to these mutations in all the used Thus, Hsm3 protein may be a new subunit of NuB4 complex.
Interaction between RAD30 and HSM3: Evidence has been obtained showin HSM3 and HIM1 genes play a role in stabilizing the D-loops [42,44]. Earlier, we sh that after the destruction of the D-loop in the him1 mutant, Polη fills the remainin (44). Based on these data, we hypothesized that the cause of hsm3-mediated UV-in mutagenesis as well as him1-dependent UV-mutagenesis is the replacement of Polδ highly erroneous Polη. To test this assumption, we have studied UV-induced mutag in rad30 and rad30 hsm3 mutants. rad30Δ single mutant showed UV-induced mutag as the wild-type strain ( Figure 5). At low doses, double mutant showed the same le Taken together, our results argue that hsm3∆ and hif1∆ mutants have the same phenotypes, and that hat1∆ mutation epistatizes to these mutations in all the used tests. Thus, Hsm3 protein may be a new subunit of NuB4 complex.
Interaction between RAD30 and HSM3: Evidence has been obtained showing that HSM3 and HIM1 genes play a role in stabilizing the D-loops [42,44]. Earlier, we showed that after the destruction of the D-loop in the him1 mutant, Polη fills the remaining gap (44). Based on these data, we hypothesized that the cause of hsm3-mediated UV-induced mutagenesis as well as him1-dependent UV-mutagenesis is the replacement of Polδ with highly erroneous Polη. To test this assumption, we have studied UV-induced mutagenesis in rad30 and rad30 hsm3 mutants. rad30∆ single mutant showed UV-induced mutagenesis as the wild-type strain ( Figure 5). At low doses, double mutant showed the same level of UV-induced mutagenesis as single rad30∆ mutant. However, at high doses, UV-induced mutagenesis in the double mutant was noticeably lower than in the single rad30∆. At the same time, double mutant showed high UV resistance than the single rad30∆. We observed the same tendency in the case of him1∆ rad30∆ mutant [44]. Thus, we can conclude that, in during PRR the Polη in hsm3∆ mutant carries out reparative synthesis in unfilled gaps. UV-induced mutagenesis as single rad30Δ mutant. However, at high doses, UV-induced mutagenesis in the double mutant was noticeably lower than in the single rad30Δ. At the same time, double mutant showed high UV resistance than the single rad30Δ. We observed the same tendency in the case of him1Δ rad30Δ mutant [44]. Thus, we can conclude that, in during PRR the Polη in hsm3Δ mutant carries out reparative synthesis in unfilled gaps. UV-induced mutations at bipyrimidine sites during TLS (TransLesion Synthesis) arise as a result of bypassing DNA damage. Mutations at non-bipyrimidine sites frequently occur on an intact template during polymerase Polη repair synthesis. It is known that the CAN1 gene sequence contains 77% bipyrimidine sites and 23% nonbipyrimidine sites (NBP) [45]. In order to find out the ratio of UV-induced mutations in non-and bipyrimidine sites in mutant rad30, mutation spectra were determined at the CAN1 locus in hsm3Δ strain. We used the same scheme and experimental conditions as in the work with him1Δ mutant [44]. 100 can R mutants were isolated after UV irradiation at a dose 84 J/m 2 .
The spectrum of mutations obtained by us in hsm3Δ mutant practically obtained in the work [44]. The UV-induced spectra generated in hsm3Δ background does not differ from the mutation spectra in him1Δ mutant and was characterized by a predominance of single base substitutions (Table 4). In NBP sites, the frequency of UV-induction mutations also practically does not differ between hsm3Δ (21 × 10 −5 ) and him1Δ (19 × 10 −5 ) and was significantly different from him1Δ rad30Δ strains (1 × 10 −5 ). Taken together, data obtained suggests a key role for Polη in hsm3-dependent mutagenesis, especially at NBP sites. UV-induced mutations at bipyrimidine sites during TLS (TransLesion Synthesis) arise as a result of bypassing DNA damage. Mutations at non-bipyrimidine sites frequently occur on an intact template during polymerase Polη repair synthesis. It is known that the CAN1 gene sequence contains 77% bipyrimidine sites and 23% non-bipyrimidine sites (NBP) [45]. In order to find out the ratio of UV-induced mutations in non-and bipyrimidine sites in mutant rad30, mutation spectra were determined at the CAN1 locus in hsm3∆ strain. We used the same scheme and experimental conditions as in the work with him1∆ mutant [44]. 100 can R mutants were isolated after UV irradiation at a dose 84 J/m 2 .
The spectrum of mutations obtained by us in hsm3∆ mutant practically obtained in the work [44]. The UV-induced spectra generated in hsm3∆ background does not differ from the mutation spectra in him1∆ mutant and was characterized by a predominance of single base substitutions (Table 4). In NBP sites, the frequency of UV-induction mutations also practically does not differ between hsm3∆ (21 × 10 −5 ) and him1∆ (19 × 10 −5 ) and was significantly different from him1∆ rad30∆ strains (1 × 10 −5 ). Taken together, data obtained suggests a key role for Polη in hsm3-dependent mutagenesis, especially at NBP sites.
RNR3 expression in hsm3, hif1, and hat1 mutants: Earlier, we showed that the reason for the change of polymerases in him1∆ mutant is a significant decrease in the level of dNTPs in mutant cells [44]. dNTP levels show a three-to five-fold increase in response to DNA damage relative to a normal S-phase, through the check-point-dependent induction of RNR genes, the allosteric regulation of RNR activity and the degradation of the Rnr1 inhibitor Sml1 [46][47][48]. To determine the role of the dNTPs pool in hsm3∆-dependent mutagenesis, we deleted the SML1 gene in wild-type and hsm3∆ mutant strains. SML1 gene encodes a specific suppressor of the RNR1 gene (RNR3 homologue). Deletion of the SML1 gene lowers the level of UV-induced mutagenesis in comparison with the wild-type strain ( Figure 6). Thus, high level of dNTP pool suppresses hsm3-dependent mutagenesis. hsm3, hif1, and hat1 mutants: Earlier, we showed that the reason for the change of polymerases in him1Δ mutant is a significant decrease in the level of dNTPs in mutant cells [44]. dNTP levels show a three-to five-fold increase in response to DNA damage relative to a normal S-phase, through the check-point-dependent induction of RNR genes, the allosteric regulation of RNR activity and the degradation of the Rnr1 inhibitor Sml1 [46][47][48]. To determine the role of the dNTPs pool in hsm3Δdependent mutagenesis, we deleted the SML1 gene in wild-type and hsm3Δ mutant strains. SML1 gene encodes a specific suppressor of the RNR1 gene (RNR3 homologue). Deletion of the SML1 gene lowers the level of UV-induced mutagenesis in comparison with the wild-type strain ( Figure 6). Thus, high level of dNTP pool suppresses hsm3dependent mutagenesis. To test these results that dNTP concentration regulates UV-induced mutagenesis, we studied the expression of RNR3 gene in the hsm3Δ mutant after UV irradiation. We measured the mRNA RNR3 gene levels in the wild type, hsm3Δ, hif1Δ, hsm3Δ hif1Δ, hsm3Δ hat1Δ, hif1Δ hat1Δ, and hat1Δ mutant cells 2 h after irradiated with UV light. The mRNA level in wild-type cells increased almost three times, while in mutant cells the increase did not reach 30% (Figure 7). Thus, hsm3Δ and hif1Δ mutations suppress the efficiency of the induction expression of RNR genes after UV irradiation. The consequence of the suppression of the expression of RNR genes will be a decrease in the dNTP concentration. Taken together, the results confirm the hypothesis that that suppression of UV-induced expression of RNR genes stimulate Polη recruitment to fill the gaps. Polη is highly erroneous polymerase and this is the cause of the increased UV-induced mutagenesis in hsm3Δ and hif1Δ mutants.

RNR3 expression in
As can be seen in Figure 7A, the level of expression of the RNR3 gene in hif1Δ hsm3Δ double mutant drops below the level of expression of this gene in wild type cells without irradiating. This result allows us to conclude that inactivation of both accessory subunits To test these results that dNTP concentration regulates UV-induced mutagenesis, we studied the expression of RNR3 gene in the hsm3∆ mutant after UV irradiation. We measured the mRNA RNR3 gene levels in the wild type, hsm3∆, hif1∆, hsm3∆ hif1∆, hsm3∆ hat1∆, hif1∆ hat1∆, and hat1∆ mutant cells 2 h after irradiated with UV light. The mRNA level in wild-type cells increased almost three times, while in mutant cells the increase did not reach 30% (Figure 7). Thus, hsm3∆ and hif1∆ mutations suppress the efficiency of the induction expression of RNR genes after UV irradiation. The consequence of the suppression of the expression of RNR genes will be a decrease in the dNTP concentration. Taken together, the results confirm the hypothesis that that suppression of UV-induced expression of RNR genes stimulate Polη recruitment to fill the gaps. Polη is highly erroneous polymerase and this is the cause of the increased UV-induced mutagenesis in hsm3∆ and hif1∆ mutants.
Genes 2021, 12, x FOR PEER REVIEW 13 of 18 was the cause of the suppression of mutagenesis in the double mutant. To test this assumption, we decided to use the dun1Δ mutant. In dun1 mutant cells, there is no increase in the expression of RNR genes after DNA damage [45,49]. We deleted the DUN1 gene in strains of wild-type, hif1Δ and hsm3Δ mutants. dun1Δ mutation significantly increases the sensitivity of yeast cells to UV radiation (Figure 8). At the same time, the double dun1Δ hsm3Δ and dun1Δ hif1Δ mutants did not differ from the single dun1Δ mutant under these conditions. Single dun1Δ mutant decreases the frequency of UV-induced mutagenesis compared to a wild-type strain ( Figure 8). Simultaneously, the double dun1Δ hsm3Δ and dun1Δ hif1Δ mutants does not differ practically from the single dun1Δ mutant according to the frequency of UV-induced mutagenesis. Taken together, these results show that a sharp decrease in the dNTP concentration suppress hsm3-and hif1-dependent mutagenesis.  As can be seen in Figure 7A, the level of expression of the RNR3 gene in hif1∆ hsm3∆ double mutant drops below the level of expression of this gene in wild type cells without irradiating. This result allows us to conclude that inactivation of both accessory subunits NuB4 complex Hsm3 and Hif1 completely suppresses UV-induced expression of RNR complex genes. It is possible that such a sharp decrease in the expression of the RNR genes was the cause of the suppression of mutagenesis in the double mutant. To test this assumption, we decided to use the dun1∆ mutant.
In dun1 mutant cells, there is no increase in the expression of RNR genes after DNA damage [45,49]. We deleted the DUN1 gene in strains of wild-type, hif1∆ and hsm3∆ mutants. dun1∆ mutation significantly increases the sensitivity of yeast cells to UV radiation (Figure 8). At the same time, the double dun1∆ hsm3∆ and dun1∆ hif1∆ mutants did not differ from the single dun1∆ mutant under these conditions. Single dun1∆ mutant decreases the frequency of UV-induced mutagenesis compared to a wild-type strain (Figure 8). Simultaneously, the double dun1∆ hsm3∆ and dun1∆ hif1∆ mutants does not differ practically from the single dun1∆ mutant according to the frequency of UV-induced mutagenesis. Taken together, these results show that a sharp decrease in the dNTP concentration suppress hsm3and hif1-dependent mutagenesis.
As seen from Figure 3B, hat1∆ mutation does not significantly affect the frequency of UV-induced mutagenesis. This is surprising, since mutations in the genes encoding the two subunits of the NuB4 complex increase the frequency of UV-induced mutagenesis. To examine the role hat1∆ mutation in RNR3 regulation, we measured the expression level of the RNR3 gene in hat1 mutant before and after UV irradiation and the double hsm3∆ hat1∆, hif1∆, hat1∆ mutants after UV irradiation. Figure 7A shows that hat1∆ mutation significantly increases the expression of the RNR complex genes both before and after irradiation. We have shown that hat1∆ mutation epistats to hsm3∆ and hif1∆ mutations, also increases the expression of the RNR complex genes after UV irradiation, as in the single hat1∆ mutant ( Figure 7B). This result explains the absence of increased mutagenesis in hat1∆ mutant and once again proves the key role of a decreased level of dNTP concentration in hsm3and hif1-specific mutagenesis.
( Figure 8). Simultaneously, the double dun1Δ hsm3Δ and dun1Δ hif1Δ mutants does not differ practically from the single dun1Δ mutant according to the frequency of UV-induced mutagenesis. Taken together, these results show that a sharp decrease in the dNTP concentration suppress hsm3-and hif1-dependent mutagenesis.

Discussion
Histone chaperone proteins have key roles in eukaryotic chromatin dynamics [50,51]. These proteins have been implicated in a wide range of processes including buffering of soluble H3-H4-complex [52], mediating H4 acetylation in the context of HAT1-complex [16,17]. In spite of these roles, histone chaperone proteins in various chromatin related processes, underlying mechanistic details are unclear. Here, we have shown that Hsm3 protein may be a new subunit of NuB4 complex.
Earlier in our laboratory, extensive research was carried out on the genetic properties of hsm3∆ mutation [25][26][27][28][29][30][31]. We have shown that proteins Mms2, Xrs2, Srs2, Mph1, Mms4, involved in the error-free branch of the PRR, have a crucial function in hsm3-dependent UV mutagenesis [38,43]. These results strongly suggest that the HSM3 gene is involved in the error-free branch of damage bypass.
Several studies previously established that Hsm3 physically interacts with Hat2, Hif1, and histone H4 [22]. In this regard, we carried out a comparative study of the genetic properties of hsm3∆ and hif1∆ mutations. In all tests carried out, both mutations showed the same properties. hsm3∆ and hif1∆ mutations did not affect the radiosensitivity of yeast cells, equally increased the frequency of UV-induced mutagenesis and the rate of spontaneous reparative mutagenesis. Both mutations do not affect spontaneous cell death, but they suppress spontaneous death of recombination-deficient mutants lowering the level of spontaneous death in double mutants to the level characteristic of single hsm3∆ and hif1∆ mutants (Table 2). In the same time rad52∆ mutation suppressed the hif1and hsm3specific UV-induced mutagenesis. The interaction of rad2∆ and hif1∆ and hsm3∆ mutations in UV-induced mutagenesis has a synergistic character. The data obtained allowed us to conclude that hsm3∆ mutation may be considered a phenocopy of hif1∆ mutation.
To date, there is no genetic evidence for the participation of Hsm3 protein in the NuB4 complex in the literature. The data obtained during the experiments on the study of UVinduced mutagenesis in hat1∆, hif1∆, and hsm3∆ mutants showed that such participation is possible (Figure 4). It can be seen from Figure 4 that the level of mutagenesis in hif1∆ and hsm3∆ mutants is the same and significantly exceeds that of hat1∆ mutant. In the double mutants, the mutation of hat1∆ epistatizes to both hif1∆ and hsm3∆ mutations. This conclusion is supported by the data of the epistatic analysis of hat1∆, hsm3∆, hif1∆, and rad2∆ mutations, as well as the data on the rate of spontaneous mutagenesis (Table 3). In all tests, hat1∆ mutation epistatized to hif1∆ and hsm3∆ mutations, thus confirming the same result.
When using the Coulson method, the cells were grown on a rich medium; the generation time for such growth was short and the amount of spontaneous damage that generated mutations during replication was relatively small. That is, the frequency of spontaneous errors of DNA polymerases on an intact template made the most contribution to the total frequency of spontaneous mutagenesis. When using the ordered seeding test, the cells were plated on a medium containing an antibiotic at a sublethal dose, which greatly increased the time of one generation (~10 divisions over 14 days) and, as a consequence, the amount of spontaneous damage. In wild-type cells, normally functioning repair systems can effectively do with this relatively small amount of spontaneous damage. However, in cells with faulty repair systems, the amount of spontaneous damage that turn out in the replication fork will be higher compared to the amount in wild-type cells. As a result, the proportion of spontaneous mutations arising in the repair process of lesions will prevail over the proportion of mutations arising from the replication of an intact template. hat1∆ mutation troubles the repair process and we observed the growth of spontaneous mutagenesis in the ordered seeding test and a slight change in the level of mutagenesis in the Coulson test. In the ordered seeding test, hsm3∆ and hif1∆, mutants showed a much higher frequency of spontaneous mutations than the mutant hat1∆, which epistatized to hsm3∆ and hif1∆ mutations in the double hat1∆ hsm3∆ and hat1∆ hif1∆ mutants. Taken together, the results obtained allow us to conclude that the HSM3 gene product is a new subunit of NuB4 complex.
The inactivation of two anti-recombination helicases Srs2 and Mph1 and Mms4 subunit of endonuclease terminating DNA synthesis in the D-loop suppresses hsm3-specific mutagenesis [38,43]. Both of these helicases and Mus81/Mms4 endonuclease decrease the average length of the synthesized DNA region [53][54][55][56][57][58]. Consequently, their inactivation will lead to an increase in the length of the newly synthesized DNA in the D-loop, and this event is the reason for the suppression of hsm3and hif1-specific mutagenesis. From this, it follows that the change of polymerases occurs after the destruction of the D-loop and the gap is filled with an erroneous polymerase. We observed the same events in him1∆ mutant [44].
Earlier, we have shown that mismatch repair plays a role in hsm3-dependent mutagenesis [30]. The mutation frequency in hsm3∆ pms1∆ double mutant was significantly higher than in both single mutants. This data shows that mismatch repair substrates arose in the cells of hsm3∆ mutant as a result of the attraction of erroneous DNA polymerases. Consistent with this conclusion, we found that Polη inactivation completely blocks hsm3dependent UV mutagenesis. This conclusion is also supported by data showed that Polη-dependent mutagenesis in NBP sites occurs significantly more frequently in hsm3∆ mutant than in the double him1∆ rad30∆ mutant.
We showed that mutations in HSM3 gene suppresses UV-induced expression of RNR genes.
Genetic data we obtained are consistent with the results of a study of physical interactions subunits of NuB4 complex with Hsm3p [22,24], indicating that the product of HSM3 gene can function as part of a NuB4 complex (Figure 9).
showed that Polη-dependent mutagenesis in NBP sites occurs significantly more frequently in hsm3Δ mutant than in the double him1Δ rad30Δ mutant.
We showed that mutations in HSM3 gene suppresses UV-induced expression of RNR genes.
Genetic data we obtained are consistent with the results of a study of physical interactions subunits of NuB4 complex with Hsm3p [22,24], indicating that the product of HSM3 gene can function as part of a NuB4 complex (Figure 9). . Figure 9. A model describing the potential mechanism of action of the Hsm3p as part of NuB4 complex. This model describes an approximate mechanism of action of Hsm3p, as a subunit of the histone acetyltransferases complex NuB4. Most likely, Hsm3p joins in the Hat1p/Hat2p complex, as well as Hif1p.