The presence of a G-quadruplex prone sequence upstream of a minimal promoter increases transcriptional activity in the yeast Saccharomyces cerevisiae Open Access

Although the shape of the DNA molecule is primarily epitomized in the double helix, the ENCODE project has shown that DNA can form noncanonical secondary structures with biological significance [1]. These structures include the so-called G-quadruplexes (G4s) that are formed from sequences rich in guanine bases. G4s are constituted of four guanine bases arranged in a square planar conformation (G-tetrad) held together by Hoogsteen hydrogen bonding and further stabilized by potassium or sodium ions [2,3]. G4s can be presented in various topologies including antiparallel, parallel, and hybrid structures, depending on the relative orientation of the DNA strand within the structure; intermolecular G4 structures can also be formed when more than one DNA strand creates the final structure [4]. G4s have been shown to be involved in processes such as DNA replication, gene transcription, translation, and the maintenance of genome stability; they occur in specific sequences such as telomeres and promoter regions of oncogenes, and also on 5′ and 3′ untranslated regions (UTRs) of mRNAs [2,5,6]. Interestingly, bioinformatics analysis of the binding sites of several transcription factors showed their potential to interact with G4s [6,7] and provided the basis for the assumption that positive or negative regulations occur between transcription factors and G4 structures [8]. Moreover, the increased frequency of G4 motifs in gene promoters compared with the rest of the genome suggests a significant contribution to the regulation of the expression of those genes [9]. Experiments carried out with a single chain antibody specific for G4 structures showed the involvement of G4s not only during the initiation of transcription, but also during its termination [8]. The first verified evidence of G4s influence on gene expression was demonstrated for the MYC oncogene, where mutations of a G4 in the promoter region affected MYC expression in vivo [10,11]. Subsequent studies addressed the regulation of transcription through G4 ligands (e.g., TMPyP4 and many others), resulting in a comparable decrease in transcription of MYC [11], KRAS [12] and KIT [13] oncogenes. Therefore, G4 targeting is suggested for cancer therapy [14,15].

The P53 family of transcription factors comprises the structurally related P53, P63, and P73 proteins that share an N-terminal transactivation domain (TA), a central sequence-specific DNA-binding domain (DBD) and an oligomerization domain (OD) at the C-terminus. All three proteins act mainly as tetramers to induce expression of a plethora of target genes involved in different cellular pathways, including cell proliferation, apoptosis, DNA repair, angiogenesis, metabolism and differentiation [16–18]. The three family members are also characterized by similar gene structures that produce groups of mRNAs controlled by separate promoters and encoding proteins with alternative N-terminal regions [19,20]. While the variants generated from the distal promoter contain the complete TA domain (TA-isoforms) and are transactivation competent, those generated from internal promoters lack the full TA domain (ΔN-isoforms) but still bind DNA, showing both specific transactivation ability and dominant negative activity. To further complicate the P53 family landscape, the transcripts are also subject to alternative splicing of the 3′ portion, giving rise to a combinatorial variety of P53, P63 and P73 specific isoforms [21,22].

Transcriptional regulation by P53 family proteins is mainly achieved through binding to a degenerate DNA motif known as the P53 response element (RE), consisting of two decameric half-sites separated by a short spacer (RRRCWWGYYY-n-RRRCWWGYYY, where R stands for a purine, W for A/T, Y for a pyrimidine and n for spacer) [23,24]. However, while the binding affinities of P53 family transcription factors are related to the primary sequence features of the RE, other factors that can modulate binding and transactivation are related to DNA accessibility within chromatin and to various features of the so-called indirect-readouts or shape-readouts; these are dependent on DNA structural features that can be influenced by trans-factors including proteins, non-coding RNAs and epigenetic changes that impact nucleosome density and positioning [24].

Interestingly, local unfolding events within the DNA double helix are able to stimulate the formation of G4 structures, potentially contributing to orchestrate P53 family cell-type specific transcriptomes [25]. Based on the critical role of these DNA structural elements in the modulation of transcriptional activity, we previously evaluated the influence of a G4 prone sequence from KSHV (Kaposi sarcoma-associated herpes virus) in proximity to a P53 RE from the BBC3 (PUMA) target gene on the transactivation potential of ΔN- and TA-variants of P53 family α isoforms [25,26]. Here, by using the same yeast-based assay, we study the effect of different G4 prone sequences on both basal and P53 family dependent expression of the Luciferase 1 reporter gene (LUC1). To this aim, different G4 prone sequences were modeled to vary the chance of intramolecular G4 formation and studied in vitro for the different propensity of forming G4 structures by Thioflavin T binding and circular dichroism analyses. The sequences were then placed downstream and adjacent from the PUMA RE, constructing otherwise isogenic yeast reporter strains. The impact of various G4 prone sequences was measured after transforming cells with centromeric inducible expression vectors and shifting them to galactose-containing medium to modulate the levels of P53, P63, and P73 (wild-type TA α isoforms). The results highlight that a higher propensity to form G4 structures upstream of the promoter correlates with higher transcriptional activity of both basal and transcription factor-dependent levels.

Synthetic oligonucleotides were purchased from Sigma-Aldrich and diluted with ultrapure water to a final concentration of 100 μM (Table 1). The PUMA sequence was derived from the BBC3 (PUMA) P53 target gene; oligonucleotide sequences with different G4 formation potential were designed to decrease the chance of intramolecular G4 formation by mutating the KSHV G-quadruplex sequence [25]. KSHV-Mut2.0 and KSHV-Mut1.5 sequences were designed by the G4 Killer program [27] to change G4 Hunter score below 2.0 and 1.5, respectively; the sequences of KSHV-1NO, KSHV-2NO and KSHV-3NO were designed by sequential substitution of G bases in guanine repeats. The propensities of G4 formation were predicted for these sequences by the G4 Hunter program and measured as G4 Hunter scores (Table 1) [28].

The thioflavin T (ThT) binding test was performed according to the procedure described in [29]. Synthetic oligonucleotides at a final concentration of 2 μM in 100 mM Tris-HCl (pH 7.5) or 100 mM Tris-HCl (pH 7.5) containing 100 mM KCl were denatured (95°C for 5 min) and gradually cooled to room temperature. Oligonucleotides were mixed with ThT in a 2:1 molar ratio and measured in a 384-well titration microplate reader at room temperature in three replicates (excitation at 425 nm). Fluorescence emission was detected at 490 nm; the measured fluorescence intensities of individual oligonucleotides were compared to the fluorescence intensity of the buffer with ThT without DNA (I/I₀).

Synthetic oligonucleotides at final concentration 3 µM in 10 mM Tris-HCl or 10 mM Tris-HCl containing 100 mM KCl were denatured (95°C for 5 min) and gradually cooled to room temperature. Circular Dichroism (CD) allows characterization of the topology of the resulting secondary structure by comparison with model spectra [30,31]. Spectra were measured on a Jasco 815 dichrograph in 1 cm tapered quartz cuvettes that were placed in a thermostatically controlled holder at 20°C. Four scans of each sample were taken at a scan rate of 100 nm·min⁻¹ with a data pitch of 0.5 nm in the wavelength range 200–330 nm, averaged, and the resulting spectra were smoothed using the Savitzky-Golay smoothing algorithm with a 15-point window. The CD signal was expressed as the difference in the molar absorption coefficient Δε of the left- and right-hand polarized light. CD measurements were taken also after 24 h.

Saccharomyces cerevisiae reporter strains that contain, in addition to the P53 RE from the BBC3 (PUMA) target gene (5′-CTGCAAGTCCTGACTTGTCC-3′), different G4 prone sequences (Table 1), were already available [25] or newly created by the Delitto Perfetto approach [32]. G4 prone sequences were located downstream of the P53 RE and upstream of the minimal CYC1-derived promoter controlling the LUC1 reporter gene, cloned at the yeast ADE2 locus on chromosome XV [33] (see sequences and chromosomal locus description in Figure S1). The following isogenic yeast strains were used: yLFM-PUMA (PUMA), yLFM-PUMA-KSHV (P-K), yLFM-PUMA-KSHV-1NO (P-K-1NO), yLFM-PUMA-KSHV-2NO (P-K-2NO), yLFM-PUMA-KSHV-Mut2.0 (P-K-Mut2.0), yLFM-PUMA-KSHV-Mut1.5 (P-K-Mut1.5) and yLFM-PUMA-KSHV-3NO (P-K-3NO). Vectors based on pTSG plasmid containing the coding sequences for wild-type P53 family TA α isoforms (i.e., P53, P63, and P73) expressed under a yeast-inducible GAL1,10 promoter were already available along with the empty vector pRS314 (TRP1 selection marker) [34]. Yeast cells were transformed by the lithium acetate method as previously described [35]. Yeast transformants were selected on plates lacking tryptophan and containing high amounts of adenine (200 mg/l), given the deletion of the ADE2 open reading frame and promoter in all strains, and incubated for 3 days at 30°C. Colonies were then patched onto the same selective plates to expand them prior to reporter assays.

Yeast transformants (empty and wild-type P53 family TA α isoforms) were resuspended using a 96-well plate in tryptophan-selective liquid medium containing raffinose as carbon source and then diluted in selective media supplemented with galactose at final concentrations of 0.016% or 1%. The shift from glucose containing plates to raffinose plus galactose media causes dose-dependent activation of the GAL1,10 promoter to induce P53 family protein expression at moderate and high levels [33]. Reporter expression is assayed after 6 h of incubation at 30°C. Luciferase assays were performed using a white 96-well plate, where 20 μl of yeast culture was mixed in each well with an equal volume of 2× Passive Lysis Buffer (Promega) to permeabilize cells by 15 min of shaking at room temperature; then the firefly luciferase substrate (20 μl) was added (Bright-Glo™ Luciferase Assay Kit, Promega). Light unit values were normalized to the OD₆₀₀ value of each culture measured from the 96-well culture plates. Results were expressed by (i) relative light units (RLU, i.e., Firefly units/OD), (ii) fold changes using the RLU values obtained from transformants with the empty pRS314 vector as reference, and (iii) relative activities as ratio of fold change. At least three independent cultures were measured for each experiment,

Ordinary one-way ANOVA followed by comparisons test were performed using GraphPad Prism version 9 for Mac (GraphPad Software, San Diego, California, U.S.A.).

First, we assessed G4 formation in the oligonucleotides under study (Table 1) using the ThT assay; as the G4 structure is stabilized by potassium ions, we used buffers with and without KCl addition. In both buffers, oligonucleotides with potential to form G4s achieved a fluorescence intensity several times higher than the PUMA oligonucleotide (negative control), whose fluorescence was only that of the ThT background signal value (Figure 1A,B and Supplementary Table S1); furthermore, fluorescent signals were higher in buffer with potassium ions (Figure 1A,B and Supplementary Table S1).

The ThT fluorescence for KSHV-1NO, KSHV-2NO and KSHV-3NO oligonucleotides decreased in accordance with the decrease of G4 Hunter score (Figure 1A,B and Supplementary Table S1); the increase in the fluorescent signal in the presence of potassium ions was most pronounced for the KSHV-2NO oligonucleotide, while the lowest increase was observed for KSHV-1NO. KSHV-3NO, characterized by the lowest G4 Hunter score, showed approximately 1.5 times higher fluorescence intensity in the presence of KCl compared with without KCl; this increase by a G4 sequence that is designed to strongly reduce the formation of the intramolecular G4 structure is probably associated with the ability of potassium ions to facilitate the formation of intermolecular G-quadruplex structures.

KSHV-Mut2.0 and KSHV-Mut1.5 oligonucleotides showed comparable fluorescence intensity in buffer without potassium ions (Figure 1A and Supplementary Table S1). A higher fluorescence intensity in KCl-containing medium was achieved only by KSHV-Mut2.0, which confirmed the formation and stabilization of G4 structures; conversely, KSHV-Mut1.5 showed a decrease in signal, with a value comparable with KSHV-3NO (Figure 1B and Supplementary Table S1). This result may suggest a destabilization of the emerging G4 structures or the formation of different loops through the contribution of electrostatic interactions with a less stable conformation.

Next, we characterized G-quadruplex formation using CD spectroscopy; measurements were also taken after 24 h to determine changes in conformation over time (Figure 2 and Supplementary Figure S2). PUMA oligonucleotide spectra (Figure 2A) did not show any typical peaks for G4 even after the addition of potassium ions, as apparent from the absence of the G4 characteristic peak in the 210 nm region [36]; moreover, the spectra were nearly identical in both types of buffers, reaching positive peaks at 220 and 280 nm and a negative peak at 250 nm. Similarly, KSHV-3NO with the lowest G4 Hunter score (0.909) did not form a G4 structure even in the presence of potassium ions (Figure 2G). In contrast with PUMA and KSHV-3NO oligonucleotides, the KSHV sequence showed characteristics typical for a parallel G-quadruplex, with positive peaks at 210 and 260 nm and a negative peak at 240 nm; moreover, a small peak around 295 nm suggested that mixed or antiparallel G4 structures were also possible for this sequence (Figure 2B). The strong formation of G4 structures in the KSHV sequence was also supported by the presence of the typical peaks described above in the absence of potassium ions.

The spectra of KSHV-1NO (Figure 2E) and KSHV-2NO (Figure 2F) suggested that these sequences could form parallel G4 structures, especially in buffer containing potassium ions. The CD spectra of KSHV-Mut2.0 (Figure 2C) and KSHV-Mut1.5 (Figure 2D) reached positive peaks around 216 and 264 nm and a negative peak at 242 nm, indicating the possible presence of a parallel G4 structure without the addition of potassium ions; in potassium buffer, the latter oligonucleotide formed a hybrid G4 structure characterized by positive peaks around 210, 260 and 295 nm.

After 24 h, there was a slight increase in signal intensity without any change in topology for KSHV-1NO, KSHV-2NO, KSHV-Mut2.0 and KSHV-Mut1.5 in KCl buffer; for the remaining motifs, there was a decrease in signal and in some cases deviations in the measured spectra, but no evidence of a change in conformation with time (Supplementary Figure S2). Taken together, the results confirm the differential propensity of oligonucleotides we selected to form G4 structures in vitro, revealing consistency with their selection based on the G4 Hunter score.

The same panel of G4 prone sequences assayed in vitro were placed downstream and adjacent to P53 PUMA RE in isogenic yeast reporter strains and analyzed by a functional assay. We began by examining the impact of the sequences on the basal expression of the LUC1 reporter 6 h after shifting the yeast cells to 0.016% and 1% galactose containing medium.

By plotting the RLU data and ordering the G4 prone sequences by decreasing G4 Hunter score, a proportional trend was observed. G4 prone sequences KSHV and KSHV-1NO led to higher transcription and this effect was progressively reduced by decreasing the strength of the G4 sequence measured by G4 Hunter (Figure 3), being partially affected also by the type of G4 prone motif. Indeed, transcription in P-K-1NO, P-K-2NO and P-K-3NO strains decreased; similarly, this effect was evident with P-K-Mut2.0 and P-K-Mut1.5 strains. However, the P-K-3NO strain, characterized by the lowest G4 Hunter score (0.909) did not show the lowest RLU value, and was statistically different from the PUMA control strain. Conversely, transcription in P-K-Mut1.5 strain (G4 Hunter score 1.091) was not or only weakly statistically different from PUMA at 0.016% and 1% galactose, respectively; finally, transcription in P-K and P-K-1NO strains was comparable and both were higher than the PUMA strain (P<0.0001).

These results suggested that the basal reporter transcription can be stimulated by the presence of a G4 sequence in S. cerevisiae in relation to the propensity to form G4 structures.

Next, we examined the potential impact of G4 prone sequences on transcription where sequence-specific transcription factors bind upstream of the promoter and the G4 site, stimulating LUC1 reporter expression. We exploited wild-type human P53 family proteins (i.e., TA P53, P63 and P73 α isoforms) under an inducible promoter to obtain moderate or high-level expression. As expected, LUC1 was enhanced at the lower level of galactose inducer, especially by P53 protein with respect to basal transcription (Figure 4). The presence of G4 prone sequences with different scores had a weak effect on modulation of P53 transactivation (Figure 4A,D); conversely, a trend is present indicating that P63- and P73-dependent reporter transcription was stimulated by the presence of G4 sequences (Figure 4B,E: P63; Figure 4C,F: P73). Based on these results, we conclude that a strong transcription factor such as wild-type P53 is less influenced by G4 prone sequences compared to the weaker P63 and P73 proteins.

Then, by plotting the data as fold change, i.e., using the basal transactivation levels to normalize the impact of P53 family proteins expression in each strain, as previously done in our transactivation studies [25,26], lower fold change values for the G4-containing strains compared to the PUMA control were evident (Supplementary Figure S3). Overall, the results were consistent with the view that G4 prone sequences are associated with higher basal transcription, a feature that may favor the transactivation function of the weak transcription factors P63 and P73. This effect becomes more evident when focusing on the relative activity of P63 and P73 transcription factors with respect to P53 (Figure 5). In general, the presence of G4 sequences with higher score reduces the gap in transactivation between P53 and the other two family members P63 and P73.

Transcription is regulated by several mechanisms, one of the most important being protein binding to promoters and enhancers [37]. Sequence-specific transcription factors interact with their target binding sites and their specificity and activity can be influenced by the presence of non-B and higher order DNA structures [38–41]. A prominent category of such structural elements includes DNA G-quadruplexes (G4s) consisting of secondary structures formed by stacked G-tetrads [9,42]. G4 prone sequences are present in all organisms including S. cerevisiae [43] and human [44–46]. Recently, G4 structures were confirmed and visualized within cells using conformation antibodies or small molecule ligands, and their presence has been correlated with transcriptional regulation [47,48]. Many studies support the view that G4s are favored by some transcription factors, affecting transcriptional response [49,50]. While several G4 binding proteins were evaluated in animals, including humans, the knowledge of G4-binding proteins in yeast is limited, and the influence of G4s on yeast transcription has not been deeply evaluated.

In the present study, our objective was to develop a proof-of-concept framework to investigate how features of different G4 prone sequences impact both basal and induced transcription in a defined, isogenic yeast model. While our approach relies on an artificial promoter-reporter construct, it has the advantage of testing defined variants of G4, embedded at the same position in a chromosomal locus upstream of a minimal promoter driving the LUC1 reporter [25,26]. Further upstream of the G4 motif is a transcription factor RE from the BBC3 (PUMA) gene that is targeted by P53 family proteins (Supplementary Figure S1). Exploiting a tunable promoter, we were able to obtain results with a defined matrix of variables consisting of (i) six different G4-forming proneness or a control sequence that does not form a G4 structure (i.e., PUMA) (Table 1), and (ii) different levels of P63, P73 or P53 (TA α variant) that differ in their binding affinity to the RE. The choice of P53 family transcription factors to investigate the impact of G4 prone sequences on transactivation was linked to our previous studies in which their transcriptional responses to many different variants of binding sites in the same chromosomal locus and yeast reporter system have been characterized [33]. Contrary to previous studies where one G4-forming sequence (i.e., KSHV) was studied by placing it either upstream or downstream of the P53 response element [25,26], here we investigated how changes in G-quadruplex propensity affect transactivation by exploiting a panel of six isogenic yeast reporter strains (Supplementary Figure S1). Interestingly, our results highlighted that an increase in basal transactivation is correlated with higher G-quadruplex propensity (Figure 6, panel A); this crucial observation was not obvious from previous results. Moreover, the possibility to test the entire family of P53 transcription factor was an added value of our proof-of-concept matrix; in fact, P53 family proteins can recognize binding sites with similar but not identical features. Also, the three proteins differ in the strength of transactivation, which is in part related to differences in quaternary structures and the presence of different combinations of domains [22,51,52]. In this regard, the presence of a G4 prone sequence in proximity to the P53 RE seems to support the activity of the weaker P63 and P73 transcription factors, possibly by facilitating their accessibility to the P53 RE or by increasing their interaction with the basal transcription machinery (Figure 6).

Moreover, we chose the PUMA derived RE due to its intermediate transactivation potential, reasoning that it could represent a sensitive tool to monitor the impact of changes in the surrounding sequences [25,33,53]. We have shown previously that high-affinity consensus P53 binding sites lead to high-level transactivation even when P53 protein is kept at minimum levels, preventing the tuning of promoter responses [54,55]. Indeed, the PUMA RE features three mismatches from the optimal P53 consensus site and does not contain the optimal, more flexible, CATG element in the core CWWG motifs; at the same time, it is responsive also to P63 and P73. Lastly, despite the S. cerevisiae genome being characterized by a low GC content and a low propensity to adopt G4s [43], a recent study identified 37 genes whose promoter contains a G4 [56]. Moreover, several yeast proteins are able to interact with G4s. Interestingly, binding to G4 DNA was demonstrated for the yeast transcription factor Msn2 with corresponding increased activity, while disruption of the G4 motif led to reduced transactivation [56]; G4 specificity was also shown for Nsr1 and Sub1 yeast proteins [57,58]. Therefore, the defined presence of G4-prone sequences in an isogenic yeast system can serve as a tool to analyze G4-dependent regulation of yeast transcription.

The six G4 sequences selected in the present study were modeled on the positive control sequence derived from the KSHV genome sequence [25]; specific changes were introduced to decrease the propensity to form G4s by the G4 Killer program [27] or by sequential substitution of G bases in guanine repeats. The resulting sequences were classified according to the G4 Hunter algorithm [28], a widely used tool for G4 prediction that takes into account G-richness and G-skewness of DNA or RNA sequences and provides a quadruplex propensity score.

The selected G4-prone sequences were biophysically characterized by ThT assays and CD spectroscopic analyses. ThT is a fluorescent probe that binds G4s more strongly than duplex or single stranded DNA, and for which a procedure for the rapid detection of G4 structures has been developed and verified [59]. Using the ThT assay, we confirmed the formation of G4s not only in the previously analyzed KSHV sequence [25] but also in the oligonucleotides we selected and characterized by a lower G4 Hunter score (Figure 1 and Supplementary Table 1). However, some deviations from the bioinformatic prediction of G4 formation were observed, mainly due to the possibility of parallel G4 formation in short oligonucleotides. Indeed, the value of the G4 Hunter score is based on the number of concurrent G bases (G-runs) and the total length of the sequence. Although the presence of G-runs is essential for G4 formation, differences in bioinformatics-determined PQS (potential quadruplex-forming sequences) and preferences for G4 formation in vitro have been demonstrated [60]. The results may have been influenced by the nature of ThT binding to DNA. A significant increase in fluorescence emission can occur due to inhibition of rotation of the ThT inner segments; binding of ThT to different DNA structures leads to different degrees of inhibition of torsional motion [61]. Three different binding modes to DNA have been described for ThT: binding to DNA cavities, intercalation between DNA bases, and external binding to DNA phosphate groups; ThT bound in each mode has a different yield of fluorescence [62].

The results of CD spectroscopy in the presence of potassium confirmed (i) the stabilization of parallel G4 conformation in KSHV, KSHV-1NO, KSHV-2NO oligonucleotides, (ii) an antiparallel conformation in KSHV-Mut1.5 sequence and (iii) a hybrid conformation in KSHV-Mut2.0 motif. That a parallel G4 conformation is represented to a higher extent in KSHV-1NO and KSHV-2NO (Figure 2C,D) than in the positive control KSHV oligonucleotide (Figure 2B), where the positive peak at 295 nm in the CD spectra partially indicates the formation of hybrid G4, is consistent with the results of the ThT assay. Indeed, ThT was reported to possess high specificity for the parallel G4 conformation [63,64]. The change in the G4 conformation of KSHV-Mut1.5 also corresponds to the ThT assay results, where a decrease in fluorescence intensity was observed in the potassium ion environment, indicating ThT binding to the less preferred G4 structure.

Transactivation data in our experimental system indicate that the presence of these G4 prone sequences enhances basal transcription proportional to their potential to form secondary structures (Figure 3). Higher levels of transactivation are also apparent when the transcription rates are enhanced by expression of the sequence-specific P53 family of transcription factors, but in this case an inverse relation to affinity and transactivation potential is evident (Figure 4).

Transcription factor-induced activity is typically expressed in reporter assays as fold change of transactivation, to focus on the direct impact of a trans-factor on gene expression. Considering that the promoter-reporter system used reached a maximal rate of transactivation, the relative fold change due to the P53 family proteins is lower when a G4 prone sequence is embedded near the promoter site, due to the effect of these sequence in increasing the basal, P53-independent transcription. Fold change graphs show a negative impact of G4 prone sequences particularly for P53 (Supplementary Figure S3); indeed, this negative effect is proportional to the G4 Hunter score, with the exception of KSHV-3NO. However, we emphasize that the fold change view is partially misleading, since the presence of G4s increase overall promoter transcription in all cases, particularly the basal activity.

In general, P53 exhibits a much stronger transactivation potential than P73 and P63 as TA α variant in yeast [54,65]. We exploited this feature, along with the inducible and tunable nature of the expression vector, to ask whether the impact of studied G4 prone sequences could be more relevant in cooperating with the weaker transcription factors P63 and P73. This seemed to be the case and led to a significant reduction in the relative activity of P53 over that of P63 and P73 with the presence of G4 forming sequence at the promoter site, and inversely proportional to the G4 Hunter score (Figure 5).

While we have no formal proof that G4 sequences are stably formed at the endogenous yeast locus where they were cloned, the impact of those sequences on transactivation broadly followed the predictions from the G4 Hunter score. However, there were some exceptions, and the KSHV-1NO sequence showed an equivalent or even a slightly higher transactivation compared to the highest-scoring KHSV sequence. This is in part consistent with the results of ThT measurements performed in the presence of KCl. Moreover, the KSHV-3NO sequence showed higher basal transactivation than the higher scoring KSHV-Mut1.5 motif, although in vitro experiments did not show a propensity for the former to form G4s. One limitation from our in-cellulo approach is represented by the inability to fully understand the contribution of a G4 prone sequence as primary sequence motifs or the effect at the level of local DNA structures. In other words, we cannot exclude that, by replacing the G content with A stretches in the oligonucleotides, new binding sites for yeast resident transcription factors are formed that can lead to higher basal transactivation independently from the formation of a G4 structure.

In conclusion, our proof-of-concept experiment modeled the impact of defined changes in G4 prone sequences at a yeast chromatin locus on both basal and induced transactivation, exploiting a well-established isogenic reporter assay. That local G4 structures influence the functionality of P53 family proteins could have important implications for the functional classification of pathogenic P53 or P63 missense mutant alleles that retain partial transactivation capacity, therefore influencing their relative activity on specific target genes.

All data are available in the article and supplementary materials.

The authors declare that there are no competing interests associated with the manuscript.

This work was supported by the Czech Science Foundation [grant number 22-21903S (to V.B.)] and the Italian Ministry of Health 5 × 1000 funds 2020 (to P.M.).

Libuše Kratochvilová: Data curation, Formal analysis, Investigation, Writing—original draft. Matúš Vojsovič: Data curation, Formal analysis, Investigation. Natália Valková: Data curation, Methodology. Lucie Šislerová: Validation. Zeinab El Rashed: Validation, Investigation, Visualization, Writing—review & editing. Alberto Inga: Visualization, Methodology, Writing—original draft, Writing—review & editing. Paola Monti: Data curation, Investigation, Visualization, Writing—original draft, Project administration, Writing—review & editing. Václav Brázda: Conceptualization, Project administration, Writing—review & editing.

The authors thank Philip Coates for critical reading and English proofreading.

CD

circular dichroism

DBD

DNA-binding domain

G4

G-quadruplex

OD

oligomerization domain

RE

response element

UTR

untranslated region