Pre-exposure to UV radiation enhances cell survival following a secondary UV exposure

We initially sought to determine if exposure to UV radiation affects cell survival following a subsequent UV exposure. A double-exposure protocol was developed, in which log phase yeast cells were split into two groups: a “ + UV” experimental group that received a 50 J/m2 exposure to UV radiation; and an unexposed “−UV” control group. Both groups were subsequently incubated for 1 h to allow for repair of DNA damage, and then cells were subjected to secondary UV exposures of varying dosages. Survival frequencies were determined by colony formation on agar plates. We observed that cells which experienced the pre-exposure (+ UV) displayed significantly higher survival rates than control cells following secondary exposures at most dosages (Fig. 1a), ranging from ~ twofold to 11-fold, depending on the secondary exposure dosage. These results indicate that a prior exposure to UV radiation triggers cellular changes that enable enhanced survival in response to subsequent UV exposures. We refer to this as a “UV hyper-resistance” phenotype (UVHR).

Fig. 1

UV exposure induces hyper-resistance to subsequent UV exposure (UVHR). Quantitative UV double-exposure survival assays were done on log-phase yeast cultures (strain BY4741) using a Philips 30 W G30T8 UV lamp at 254 nm. Cultures were serially diluted and plated in duplicate on YEPD agar, exposed to UV, and then incubated in the dark for ~ 6 days. Colonies were counted and used to calculate relative survival frequencies. Assays were done a minimum of three times, with mean ± 1 SE reported for each condition (*p < 0.01). A. Yeast cells were exposed to two sequential rounds of UV radiation, first on suspended cells at 50 J/m2 (+ UV), along with a mock-exposed control (−UV), followed by a 1 h incubation at 30 °C, and then plating and UV exposure to varying secondary dosages (as indicated on the X axis). B. Same as A, except with varying initial exposure dosages (25, 50, and 100 J/m2, respectively). *, statistically significant differences found at the indicated UV secondary dosages (p < 0.01 for all pairwise comparisons, with the following exceptions: at 50 J/m2, only + UV50 versus + UV100 and −UV versus + UV100 differences are significant; for all other comparisons, p > 0.05; at 100 J/m2, −UV versus + UV25, p > 0.05; at 150 and 200 J/m2, −UV versus + UV100, p = 0.03; at 200 J/m2, + UV25 versus + UV50 and + UV100, p > 0.05). C. Same as A, except with a 4 h incubation period. D. Same as A, except with varying incubation times, as indicated on X axis. The graph displays fold difference in survival (+ UV relative to −UV) based on secondary exposure survival at 200 J/m2. Values for 1- and 4-h incubation assays are derived from a and c. Values for the 7-, 10-, 16-, and 22-h incubations are derived from assays reported in Additional file 1: Figure S1

To further evaluate this response, we varied the intensity of the initial UV exposure to determine how UVHR is influenced by UV dosage. The same double-exposure protocol was followed, this time with initial exposure intensities of either 25 or 100 J/m2. A pre-exposure of 25 J/m2 resulted in significant increases in cell survival relative to unexposed cells, although the degree of enhanced survival was reduced relative to what had been observed with the 50 J/m2 pre-exposures. Significant increases in survival were observed only at secondary exposures of 150 and 200 J/m2, and the fold increase in survival was reduced at all secondary exposure intensities relative to the 50 J/m2 pre-exposure group at the same intensities (twofold and fourfold, respectively; Fig. 1b). In contrast, pre-exposure at 100 J/m2 led to decreases in survival for the + UV group relative to the −UV group for secondary exposures of 50, 100, and 150 J/m2, reversing the trend observed in the other two pre-exposure intensities. + UV cell survival dropped to ~ 0.5-fold relative to −UV cells at these dosages. While cell survival was statistically higher for pre-exposed cells that experienced a 200 J/m2 secondary exposure, this fold increase (2.5-fold) was smaller than the difference observed in the other pre-exposure intensity groups. Thus, UVHR is dependent on the intensity of the initial exposure, with a maximal effect observed at 50 J/m2.

UVHR is epigenetically inherited during mitosis

We sought to determine whether UVHR is a persistent and inheritable phenotype. To do so, we extended the incubation time between UV pre-exposure and secondary exposure to provide cells sufficient time to repair UV-induced damage and undergo subsequent rounds of mitotic reproduction before the secondary exposure. Using a 4 h incubation, allowing for ~ 1–2 mitotic cycles (based on observed ~ threefold increases in + UV cell culture density during the incubation period), we observed significant increases in survival of the + UV cells at all secondary exposure dosages, with larger fold increases in survival of + UV cells relative to −UV cells at each dosage compared to the original 1 h incubation (Fig. 1c). Under these conditions, we observed a ~ 200-fold increase in survival in the + UV group at the 200 J/m2 secondary exposure. UVHR was observed in cultures that were incubated up to 16 h following the initial exposure (Fig. 1d, Additional file 1: Figure S1). Fold increased survival in + UV cells relative to −UV cells gradually diminished with extended incubation time, reaching a low of twofold after 16 h (at 200 J/m2 secondary exposure), and completely disappearing at 22 h. Over a period of 16 h, it is estimated that the initial UV exposed cells have undergone ~ 7–8 rounds of cell division (based on observed ~ 100-fold to 200-fold increases in + UV cell culture density during the incubation period), indicating that the UVHR phenotype is inherited mitotically.

To further examine the inheritability of the UVHR phenotype, we considered the possibility that UVHR might be exclusively present in the pre-exposed mother cells, which, because of the budding mechanism for reproduction in S. cerevisiae, remain part of the secondarily exposed cell population (in gradually decreasing relative numbers). To determine if UVHR was specifically passed onto daughter cells, we used a labeling technique to remove mother cells from the population prior to the secondary exposure. Immediately following UV exposure, mother cells (in both the −UV and + UV cultures) were labeled with biotin, which covalently attaches to the cell wall [16]. Cells were then incubated for 4 h to allow for expansion of the cell populations. Yeast cell walls are synthesized de novo at the site of bud formation [16], thus the biotin label is exclusively retained by the mother cell, and not passed onto the daughter cell. After the incubation period, streptavidin-linked magnetic beads were added to cultures to extract the biotin-labeled mother cells from the population. The remaining daughter cells were then plated and exposed to the secondary UV dosage (200 J/m2) and evaluated for viability via colony formation. We observed that the pre-exposed daughter cells (+ UV) exhibited significantly increased survival following the second UV exposure relative to the unexposed (−UV) control population, with a 40-fold increase in survival (Fig. 2a). While the magnitude of the UVHR phenotype was reduced relative to what was previously described above in the unsorted populations under equivalent UV intensity/incubation time conditions, this difference is likely explained by the prolonged period between the two UV exposures and temperature differences (the biotin labeling/sorting process adds ~ 3 h to the incubation period, with ~ 2 h at 4 °C). These results indicate that the UVHR phenotype is passed onto subsequent cell generations, and not simply a product of the mother cell response to UV exposure.

Fig. 2

UVHR is epigenetically inherited. UV survival assays were done as described in Fig. 1, with variations as indicated below. A Mother cell-free population. Cells were labeled with biotin immediately after the initial UV exposure at 50 J/m2 (+ UV), or after mock exposure (−UV). After the 4 h incubation, labeled mother cells were removed from the population with streptavidin-linked magnetic beads. The remaining unlabeled daughter cells were subsequently diluted, plated, and exposed to UV at 200 J/m2. *p < 0.01. B Isolated colonies from a single UV exposure at 50 J/m2, followed by a 4 h incubation. Thirty isolates from UV exposed set (+ UV) and 10 isolates from mock-exposed controls (−UV) were restreaked and incubated for ~ 3 days, and then subsequently cultured, diluted, plated, and exposed to UV (200 J/m2). Each isolate was tested twice, and values are reported as means. Horizontal bars indicate the geometric mean for each set of isolates. p > 0.1

We also considered the possibility that the UVHR phenotype might be genetic, potentially via mutations caused by the initial UV exposure. To explore this, we isolated random colonies produced from cultures that had been pre-exposed to 50 J/m2 UV, followed by a 4 h incubation prior to plating. The resultant colonies were re-streaked and incubated on fresh agar plates for ~ 72 h, to ensure that cells had grown far past the ~ 16–22-h period during which the UVHR phenotype is observed. These isolates were then used to perform a standard single UV exposure survival assay alongside isolates derived from unexposed control cultures. If the UVHR phenotype is due to mutational changes, it was predicted that a subset of the + UV isolates would display persistent UVHR.

The majority of + UV isolates exhibited survival frequencies that were comparable to those seen in the −UV isolates (Fig. 2b). A small fraction of + UV isolates displayed modestly increased survival relative to the controls, with a maximum of ~ tenfold increased survival. In contrast, a slightly larger fraction of + UV isolates displayed an ~ tenfold decrease in survival. The collective survival average of the + UV isolates was slightly lower than that of the −UV isolates, but this difference was not statistically significant. These results suggest that only a small percentage of cells acquire mutations that confer a hyper-resistance phenotype, and these effects are largely offset on the population level by mutations that cause hyper-sensitivity to UV. Furthermore, the degree of hyper-resistance observed in these select isolates is much smaller than that observed in the collective population (tenfold versus 200-fold). While we cannot fully rule out the possibility that there might be rare mutant isolates that exhibit high hyper-resistance that could explain the effect observed at the population level, such mutants would have to possess extremely high degrees of UV resistance to account for the aggregate UVHR phenotype. Thus, the collective results argue that the UVHR phenotype is not genetically based, but rather is a product of an epigenetic mechanism.

Pre-exposure to UV radiation protects against subsequent UV damage

We subsequently addressed the question of the underlying molecular basis of the UVHR phenomenon. We initially considered the possibility that UVHR is a result of enhanced DNA repair during the secondary exposure to UV. UV radiation causes the formation of cyclopyrimidine dimers (CPDs), which are processed by a variety of DNA damage-induced repair processes, including nucleotide excision repair [10]. We speculated that repair processes might remain active after the repair of the initial damage, or that repair genes remain poised for hyper-activation during subsequent DNA damage events. If so, we anticipated that pre-exposed cells would be able to repair UV-induced CPDs more efficiently.

To test this hypothesis, samples of pre-exposed cells (and unexposed controls) were collected at various times following the secondary UV exposure. In these experiments, cells were either exposed (+ UV) or not exposed (−UV) to 50 J/m2 of UV radiation and incubated for 4 h (by which time CPDs were completely repaired). Both cultures were then exposed to 50 J/m2 of UV radiation. DNA was isolated from cells at various times before and after each exposure, and immunoblot assays were carried out using anti-CPD antibodies to quantify CPD levels (using equal quantities of DNA per slot blotted sample/culture condition). CPD levels were normalized to the damage levels detected immediately following the secondary exposure to determine the repair kinetics.

We found that the rate of removal of CPDs was indistinguishable between the −UV and + UV cultures (Fig. 3a, b). There was no significant difference in the fraction of relative CPDs remaining between the cells that were pre-exposed and those that were not over the subsequent 90 min evaluation period. These results indicate that altered repair kinetics are unlikely to be the cause of acquired UV hyper-resistance in yeast cells. However, we observed a pattern in the DNA samples collected immediately after the second UV exposure: the −UV cells displayed ~ 3.6-fold higher relative CPDs levels compared to the + UV cells (Fig. 3a, b). This observation suggests that the pre-exposed cells acquire less damage during the secondary UV exposure than the unexposed control cells.

Fig. 3

UV-exposed cells experience reduced DNA damage levels in response to subsequent UV exposure. Immunoblotting was done to assess UV-induced CPD formation and repair kinetics. Suspended yeast cells were UV exposed, as described in Fig. 1, and following indicated incubation periods, DNA was isolated from culture aliquots. Equal amounts of DNA were slot blotted and probed with anti-thymine dimer antibodies. Images were then evaluated by densitometry. Assays were done a minimum of three times, with mean ± 1 SE reported for each condition (*p < 0.01). A. UV repair kinetics. Log-phase cultures were initially exposed to UV at 50 J/m2 (+ UV), or mock exposed (−UV), followed by a 4 h incubation, and then a secondary UV exposure at 50 J/m2. Samples were collected at the indicated timepoints for DNA isolation and analysis. −UV and + UV images shown are from the same blot using identical exposure conditions. B. Densitometry analysis of the experiments presented in A, reported as the percentage of remaining CPD levels relative to the amount of CPDs present immediately after the second exposure. C. CPD levels acquired in pre-exposed (+ UV) and unexposed controls (−UV) during secondary UV exposures. Cultures were initially exposed to 50 J/m2, followed by a 4 h incubation, and then a secondary exposure at varying dosages (0–200 J/m2). Samples were collected immediately after the second exposure for DNA isolation and analysis. −UV and + UV images shown are from the same blot using identical exposure conditions. D. Densitometry analysis of the experiments presented in C, reported as CPD levels relative to the amount of CPDs present in the −UV/50 J/m2 samples

To further substantiate this finding, we repeated these experiments by varying the secondary exposure dosage levels. These experiments showed similar results: in general, the −UV samples contained relative CPD levels ~ 2–3 times higher than the + UV samples (Fig. 3c, d). This difference was statistically significant at secondary exposure dosages of 50, 100, and 150 J/m2, and approached significance at the 200 J/m2 dosage. These results suggest that the UVHR phenotype is the result of a protective mechanism that is employed after the initial UV exposure to insulate against the formation of CPDs.

Cell morphology changes following UV exposure may protect against subsequent exposure

Given the apparent DNA damage protection acquired during the initial UV exposures, we wished to gain further insights into the manner by which protection is achieved. We examined cells microscopically prior to and at various times after the initial UV exposure to determine if any obvious morphological changes occurred that correlated with the UVHR phenotype. Collected cells were stained with calcofluor white (CFW), a non-specific dye that binds to cellulose and chitin in yeast cell walls [17], and examined by fluorescent microscopy. Cells were digitally photographed and then analyzed to evaluate changes in CFW staining intensity and overall cell size.

We found that UV exposed cells displayed an increase in size, beginning ~ 2 h after UV exposure, and persisting for at least 10 h, across multiple cell divisions (Fig. 4). The cross-sectional area of UV-exposed cells reached a maximum of a twofold increase relative to unexposed cells 7 h following UV exposure, remaining ~ 1.5-fold higher at the 10 h mark (Fig. 4b). In addition, UV exposed cells displayed ~ 1.3-fold increased CFW fluorescence beginning at the 2 h timepoint, persisting until the 7 h mark (Fig. 4c). Thus, these collective results indicate that UV exposure causes an increase in cell size and altered cell wall characteristics. The timeframe during which these changes persist roughly correlates with the duration of the UVHR phenotype.

Fig. 4

UV-exposed cells experience persistent increases in cell size and changes in cell wall composition. Log-phase suspensions were exposed to UV at 50 J/m2, or mock exposed, followed by incubation, as described in Fig. 1. Samples were collected from exposed cultures and unexposed controls at indicated timepoints and stained with calcofluor white. Cells were visualized by fluorescent microscopy and analyzed for cross-sectional cell area and fluorescent intensity. A Representative images from unexposed (−UV) and exposed (+ UV) cultures, following a 4 h incubation (400X magnification). B Cross-sectional cell area, reported as means ± 1 SE, normalized relative to the corresponding −UV cell size for each condition (*p < 0.01). C Calcofluor white (CFW) staining intensity, reported as means ± 1 SE, normalized relative to the corresponding −UV CFW staining intensity for each condition (*p < 0.01)

In light of these observations, we executed a genetic screen to determine if UV-induced cell size changes might be responsible for the UVHR phenotype. We evaluated ~ 40 strains possessing knockouts of genes that have been previously demonstrated to have altered cell size, as reported in the Saccharomyces Genome Database [18]. We initially screened these strains for UV-induced size alterations, as described above, to identify mutants that displayed abnormal size changes in response to UV. About one quarter of these strains failed to enlarge in response to UV exposure (Additional file 1: Figure S2a; strains denoted with yellow boxes), and another quarter of these strains were found to be inherently large (in the absence of UV exposure), with minimal to no additional enlargement in response to UV (denoted with green boxes). To assess if the UV size response abnormalities had an impact on the UVHR phenotype, we subsequently evaluated cell survival of these two subgroups of mutant strains using the quantitative UV double-exposure assay described earlier (Additional file 1: Figure S2b). Regardless of the starting size of the cells or their lack of size changes in response to UV, we found that most of these mutant strains exhibited a normal UVHR phenotype. Only two strains (bem4 and fig 1) exhibited a partial reduction of the UVHR phenotype. These findings suggest that UV-induced cell size changes are not required for the UVHR phenotype.

However, we did note varying degrees of general survival in response to UV exposure across these strains. When examining UV survival in the absence of pre-exposure, we observed a modest, statistically significant linear relationship relative to cell size, with larger cells exhibiting higher survival frequencies (Fig. 5a). Interestingly, we found that increased survival is specifically observed in strains that are  > 1.25-fold larger compared to unexposed wildtype cells, corresponding roughly to the size of UV-exposed wildtype cells (binned group L2 in Fig. 5b). The survival frequency observed in these strains is comparable to that seen in UV-exposed wildtype cells. These results indicate that increased cell size serves as a protective mechanism against UV. Thus, while the UV-induced cell size increase is not explicitly required for UVHR, this induced morphological change may be a contributing component to the phenotype.

Fig. 5

Relationship between yeast cell size and UV resistance. A Strain cell size was correlated with UV survival, based on data reported in Additional file 1: Figure S2a, b. Survival frequencies are log-converted values, accompanied by best fit lines for the pre-exposed (+ UV; R2 = 0.05; p = 0.34) and unexposed conditions (−UV; R2 = 0.35; p = 0.003). Darkened data points indicate the wildtype strain. B Log-converted UV survival frequencies in pre-exposed (+ UV) and unexposed (−UV) strains, binned based on relative cell size ranges: S (smaller than −UV wildtype; 0.85–0.95 relative size); N (normal; comparable to −UV wildtype; 0.95–1.05); L1 (modestly larger than −UV wildtype; 1.05–1.25); L2 (much larger than −UV wildtype; comparable to + UV wildtype;  > 1.25). The lack of a bar for “S + UV” indicates that no strains were in the indicated size range. Values are means ± 1 SE. For data sets not sharing a letter on the graph, p < 0.01 (except S- versus L2-, p = 0.02)

To evaluate the role of UV-induced cell wall changes in the context of the UVHR phenotype, we similarly executed a genetic screen to identify genes required for cell wall synthesis/assembly/maintenance that might be required for UVHR. We executed a qualitative UV double-exposure assay on ~ 50 yeast strains that possessed knockouts of non-essential cell wall-related genes. In general, we found that all strains exhibited largely normal hyper-resistance to UV following an initial 50 J/m2 pre-exposure, with only minor differences observed between the mutant and wildtype strains (Additional file 1: Figure S3). Thus, while there is a correlation between the UVHR phenotype and changes in cell wall morphology, we cannot definitively demonstrate that UVHR is caused by cell wall changes.

Specific histone post-translational modifications are important for UVHR

Our final goal was to identify the inheritable epigenetic factors that propagate the UVHR phenotype. As addressed earlier, histone modifications have been identified as potential epigenetic mediators, and may be inherited through cell division. To pinpoint histone modifications that contribute to the UVHR phenotype, yeast strains lacking genes that encode for specific histone modification enzymes (or genes that regulate these modifications) were evaluated for UVHR phenotype using the qualitative UV double-exposure assay described above. Cells either received a primary 50 J/m2 exposure or were left unexposed, followed by a 4 h incubation period and subsequent UV exposures at higher dosages. Strains that displayed reduced UVHR were subsequently evaluated via a quantitative UV double-exposure assay to characterize the impact of the specific gene knockouts more precisely with respect to the UVHR phenotype.

While most of the mutant strains exhibited only very slight or no impact on UVHR (Additional file 1: Figures S4 and S5), several mutant strains were identified that displayed notable alterations to the UVHR phenotype (Fig. 6), focusing our attention on H3K56 acetylation (H3K56ac) and H3K4 methylation (H3K4me). With regards to H3K56ac, we observed that deletion of RTT109, which encodes for histone H3K56 acetyltransferase Rtt109 [19], resulted in an increase in resistance to the initial UV exposure, while concurrently eliminating subsequent hyper-resistance conferred by pre-exposure (Fig. 6a). To confirm the role of this modification in UVHR, we examined the response in the unacetylatable histone H3K56R mutant. Comparable to the rtt109 mutant, the H3K56R mutant displayed increased general resistance to UV relative to the wildtype strain in the absence of pre-exposure, albeit to a lower degree than the rtt109 strain (Fig. 6b), accompanied by a modest ~ 20-fold increased resistance in response to UV pre-exposure. Furthermore, we found that deletion of SPT10, which is required for cell cycle-specific acetylation of H3K56 of select histone genes [20, 21], resulted in a substantial reduction of the UVHR phenotype (Fig. 6c). Finally, while deletion of the individual HST3 and HST4 genes, which encode for redundant H3K56 deacetylases [22], had no impact on UVHR (Additional file 1: Figure S5), deletion of both genes resulted in a partial reduction in the UVHR phenotype (Fig. 6d). These collective results indicate that H3K56ac is important for the UVHR response.

Fig. 6

UV double-exposure survival analysis of histone modifier mutant strains. Assays were done as described in Fig. 1. A Histone H3K56 acetyltransferase rtt109 mutant. B Histone H3K56R mutant. C Histone H3K56 acetylation regulator spt10 mutant. D Histone H3K56 deacetylase hst3/hst4 double-mutant. E Histone H3K4 methyltransferase set1 mutant. F Histone H3K4 demethylase jhd2 mutant

To evaluate how H3K56ac is regulated in response to UV exposure, we employed western blotting to measure acetylation levels in UV-exposed cells after the 4 h incubation period. We found that UV-exposed cells experienced a ~ 2.5-fold increase in H3K56ac levels relative to pre-exposure levels (as well as compared to unexposed cells after the same incubation period; Fig. 7a, b). As expected, no acetylation was observed in the rtt109 strain in response to UV, indicating that Rtt109 is required for UV-induced acetylation (Fig. 7c, d). In the spt10 strain, H3K56ac levels were reduced to ~ 20% of wildtype levels in absence of UV. UV exposure caused a corresponding threefold increase in acetylation in the spt10 strain, but overall H3K56ac levels remained proportionately reduced relative to UV-exposed wildtype levels. Conversely, H3K56ac was elevated in the hst3/hst4 strain, regardless of UV exposure conditions, with levels comparable to those observed in the UV-exposed wildtype cells. These collective results suggest that increased H3K56 acetylation by Rtt109, regulated by Spt10, is an important feature of the UVHR phenotype.

Fig. 7

Histone H3K4 methylation (H3K4me) and H3K56 acetylation (H3K56ac) levels in UV-exposed cells. Log-phase cultures were exposed to UV at 50 J/m2, followed by a 4 h incubation. Samples were then collected, and isolated proteins were analyzed by western blot with histone modification-specific antibodies. Blots were subsequently analyzed by densitometry. Values were initially normalized relative to general H3 values to correct for histone level variation, and then subsequently normalized relative to the pre-exposure values for each modification (except panel D; normalized to unexposed wildtype cultures). Values are reported as means ± 1 SE reported for each condition. + UV, exposed cells; −UV, unexposed controls; PE, pre-exposure cells. A, B. H3K56 acetylation in wildtype cells. *p < 0.01. C, D. H3K56 acetylation in wildtype (WT), spt10, hst3/hst4, and rtt109 mutant strains. For data sets not sharing a letter on the graph, p < 0.01. E, F. Histone H3K4 methylation levels in wildtype cells. H3K4me1, monomethylation; H3K4me2, dimethylation; H3K4me3, trimethylation. *p < 0.01

We similarly found that deletion of genes involved in H3K4me influences UVHR. Deletion of SET1, which encodes for the H3K4 methyltransferase [23], had no impact on UVHR (Fig. 6e). However, deletion of JHD2, which encodes for an H3K4 demethylase [23], resulted in decreased general resistance to UV, while also reducing the UVHR response (Fig. 6f). These results indicate that while H3K4me is not required for UVHR, regulation of this modification via the Jhd2 demethylase is important for this phenotype. We subsequently examined UV-induced changes in H3K4me by western blot. We found that methylation levels at this site increased in response to UV during the subsequent post-exposure incubation period (Fig. 7e, f). However, the response was uniquely observed with respect to trimethylation, which increased ~ twofold, with no significant changes observed with respect to mono- or di-methylation. The collective results indicate that UV-induced changes in H3K4me3 levels are important for UVHR.