Construction of a pheromone−inducible synthetic promoter

Starting from the AGA1 promoter, we first selected an alternative core promoter based on the CYC1 promoter (S3 Fig). Although the CYC1 core ensures a sizable inducibility, the expression dynamics are delayed by 5 minutes relative to pAGA1. This delay reinforces the idea that the core promoter identity contributes to define both the level and kinetics of gene expression [17]. In all our synthetic promoters, we will use the same pCYC1 core promoter allowing to focus our study on the role of the regulatory sequence in the dynamics of gene expression.

The promoter activity is controlled by the regulatory region, which can contain multiple TF binding sites. In pAGA1, we have identified three consensus PREs and multiple PRE-like (S3A Fig). We had previously determined that two PREs spaced by 29 bp were important for the fast and high induction of the promoter upon α-factor treatment [17]. To test if these two PRE sites spaced by 29 bp are sufficient to control gene induction, we placed them in a completely synthetic context. To do so, we have selected a CYC1 promoter, which was modified to decrease nucleosome binding [37] and where identified TF binding sites were mutated [29], as well as sequences that resembled potential PRE sites. Two PREs spaced by 29 bp placed in this synthetic context are not sufficient to induce expression (S3B and S3C Fig). The inducibility of the construct is recovered when the original sequence between the two PREs from pAGA1 is included. This difference was attributed to the presence of a PRE-like site located 3 bp away from the second consensus PRE [24]. Engineering a pCYC1 UAS containing two PRE-consensus sites 3 bp apart fused to the pCYC1 core, resulted in a rapid pheromone-inducible synthetic promoter (S3B, S3C and S3D Fig). This initial synthetic construct (pSYN_3TT- 2PRE spaced by 3 bp in Tail-to-Tail configuration) will be used as a reference for the systematic alterations that will be performed on the PRE sites to generate our library of Ste12-dependent promoters.

PRE site conformations

Our initial efforts to engineer a synthetic promoter have clearly demonstrated that while two PREs are required to induce gene expression, not all PRE conformations result in a functional promoter. Endogenous promoters harbor a wide diversity of PRE configurations, and it is difficult to predict which ones of these PRE or PRE-like sites contribute to the overall expression output by allowing formation of a Ste12 dimer. Dorrity et al. have described the canonical 3 bp tail-to-tail conformation (as found in our pSYN_3TT) as the most favorable binding in vitro for the Ste12 DNA-binding domain (DBD) fragment [24]. This PRE conformation is found on multiple endogenous promoters such as pAGA1, pSTE12, pKAR4 and pFAR1. However, many endogenous promoters don’t harbor this element. Therefore, it is likely that other conformations can promote a functional Ste12 binding. In addition, Su et al. demonstrated that a head-to-tail conformation could induce gene expression, while a head-to-head positioning prevented Ste12 binding when placed in close proximity [27].

To identify functional PRE conformations, we have systematically altered the arrangement of two PRE sites in our synthetic promoter construct (Fig 1C). With PRE in tail-to-tail orientation, the promoter becomes non-functional if the distance between the PREs is extended to 5, 7 or 10 bp. Interestingly, transcriptional activity is recovered when the spacing is set at 13 bp. This distance can even be further extended to 23 bp, resulting in a low and slow expression output (Fig 1D and 1E). These functional binding conformations can be rationalized by the fact that the DNA-helix turn corresponds to 10.5 bp [38]. Therefore, spacing the two PRE by 3, 13 or 23 bp, positions the two Ste12 proteins in a similar interaction geometry. The 13 bp tail-to-tail conformation is found in the promoter of SST2, however, the PRE spacings of 23 was not identified in the endogenous promoters we inspected.

If the orientation of the PRE is changed to head-to-tail or head-to-head conformations, the preferred distance between PREs becomes 5 bp (Figs 1C and S4). A spacing of 15 bp is also transcriptionally active, but other spacings result in no or very low expression output. The head-to-tail conformation spaced by 5 bp is readily found in numerous promoters (pAGA1, pFUS1, pPRM1, pFIG1), while the 15 bp seems less prevalent (pFUS2). In contrast, the 5 bp head-to-head conformation does not seem to be frequent because we did not identify it in the two dozen endogenous promoters investigated. We also verified by deleting STE12 that the induction of these synthetic promoters was strictly dependent on this transcription factor (S3E and S3F Fig).

Overall, these findings validate the use of synthetic promoters to identify functional binding site conformations. These results can be transferred to endogenous promoters to pinpoint the sites implicated in the mating-dependent induction. The promoter sequences can be analyzed to identify short-range interaction with specific spacings of 3 bp for tail-to-tail orientations and 5 bp for the other orientations while including a possible increment of 10 or 20 bp corresponding to one or two DNA helix turns.

Contribution of Ste12 activation domain to gene expression

Ste12 activates transcription via specific PRE conformations. This extended set of functional PRE pairs indicates a surprising flexibility from Ste12 to homodimerize (Fig 2A). In vitro data have shown that the DNA binding domain (DBD) of Ste12 can dimerize to bind to two PREs [24]. However, it is difficult to imagine how the Ste12 DBD can support the various dimerization conformations identified with our synthetic promoters, which include diverse orientations and distances. Since it has also been shown that the activation domain (AD) of Ste12 can multimerize [23], we wanted to determine if this flexible region of the protein contributes to the stabilization of some of the Ste12 binding conformations that we have identified, for instance, on PREs separated by larger distances.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Dominant role of Ste12 DNA binding domain for PRE conformation selection.

A. Schematic of the Ste12 transcription factor composed of a DNA-binding domain and a flexible activation domain. The small schemes describe the diversity of Ste12 homodimer interactions that could take place when varying the orientation or the distances of the PRE sites on the promoter. B. Scheme of the Ste12-EV chimeric transcription factor where the activation domain of Ste12 has been replaced by the estrogen receptor and the VP16 activation domain resulting in an β-estradiol responsive TF. C., D. Dynamics of nuclear relocation following stimulation with 1µM β-estradiol at time 0 for strains containing the pAGA1-dPSTR^Y (C) and various pSYN-dPSTR^R (D). The solid lines represent the median of the population and the shaded area, the 25- to 75- percentiles. E. Comparison of the inducibility of various PRE conformation with the Ste12 WT or the Ste12-EV (black borders). The Expression Outputs (EO) for the synthetic reporters induced by the Ste12 WT or the Ste12-EV were normalized relative to the EO of the reference pSYN_3TT. The bar represents the mean response of the replicates shown by the circles. A significant difference between the normalized EO of Ste12-WT and Ste12-EV based on the mean of the individual replicates is indicated by a star (t-test: p-val < 0.05).

https://doi.org/10.1371/journal.pgen.1011710.g002

To test the relative contributions of the DBD and the AD of Ste12 for the inducibility of our various synthetic promoters, we replaced the region coding for the activation domain in the endogenous STE12 locus by a fusion between the human estrogen receptor and the VP16 activation domain (EV) [39,40]. Using this construct (Fig 2B, Ste12-EV), Ste12-responsive genes will be activated by stimulating the cells with β-estradiol, which promotes the relocation of the Ste12-EV to the nucleus and by-passes the mating MAPK cascade. The strength of induction of the PRE containing promoters will be governed only by the binding of the Ste12 DBD. Indeed, no interaction from the EV domain is expected (S5A and S5B Fig). Thus, we can test if the absence of the Ste12 activation domain lowers the inducibility of our synthetic promoters.

Stimulating the cells expressing Ste12-EV with β-estradiol results in a potent, while relatively slow, activation of the pAGA1-dPSTR^Y and of the pSYN_3TT-dPSTR^R (Fig 2C and 2D). The time required to relocate Ste12 from the cytoplasm to the nucleus may contribute to this delay [40,41].

While the pSYN_3TT-dPSTR^R is strongly induced in the Ste12-EV background, synthetic promoters bearing no PRE or a single PRE fail to be induced (S5C and S5D Fig). Similarly, if the PREs are spaced by 40 bp, no relocation of the dPSTR^R can be observed. These control experiments demonstrate that the Ste12-EV chimeric transcription factor relies solely on the DBD domain of Ste12 for activation and if the promoter contains a single PRE or 2 PREs placed in an undesired configuration, the association of Ste12-EV to the DNA is too weak to promote transcription, despite the presence of the potent VP16 activation domain.

To determine if the activation domain of Ste12 plays a specific role in the capacity of Ste12 to promote transcription for a subset of PRE conformations, we compared the strength of induction of various synthetic promoters from wild-type Ste12 and from Ste12-EV (Fig 2E). The expression output was normalized relative to the induction of the pSYN_3TT-dPSTR^R. A significant decrease of 20–30% in the dPSTR^R normalized expression output was observed for the Ste12-EV when the two PREs are spaced by 13, 15 or 23 bp. Note that the normalized pAGA1-dPSTR^Y expression output and the fraction of responding cells remain similar between Ste12 and Ste12-EV (S5E and S5F Fig). These results suggest that the expression output is mostly dictated by the interaction between the PRE and the DBD. However, when the two PREs are spaced by more than 5 bp, the self-interacting properties of the activation domain of Ste12 contribute to promote the expression of the downstream gene, possibly by stabilizing the formation of the Ste12 dimer when the conformation of the two PREs is not optimal.

Affinity of the PRE sites

Suboptimal PRE-arrangements can prevent Ste12 from binding to the promoter and limit the expression output. However, even in the context of PRE-dimers with optimal arrangements, Ste12 binding on mating promoters is affected if the binding site carries point mutations (PRE-like) as is found in a majority of endogenous promoters which harbor a combination of a consensus PRE and a PRE-like site (pAGA1, pFIG1) [24]. A single base change to the TGAAAC consensus sequence can have a very different effect on the Ste12 affinity. While TaAAAC has only a 20% decrease in affinity, TGAgAC or TcAAAC result in a more than 95% reduction in competitive binding relative to the consensus sequence [27].

Therefore, to test the influence of the strength of the PRE site on the expression output, two PREs spaced by 3 base pairs were used and the sequence of one of the binding sites was modified. A clear decrease in the fraction of responding cells and in the level of expression are observed for the six PRE-like variants tested (Fig 3A and 3B). This is in line with previous measurements where the lowering of the affinity of the DNA-binding site results in a weaker gene expression output [27,30]. We can imagine that the lower affinity of the site decreases the residence time of Ste12 on the promoter and therefore limits the expression output. Importantly, however, the dynamics of gene expression is not influenced by the lowering of the binding site affinity and all the promoters tested here, despite their lower induction level, are induced rapidly (Fig 3D).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Influence of Kar4 deletion on low affinity PRE sites.

A. Dynamics of nuclear relocation for the dPSTR^R controlled by two PRE spaced by 3 bp in tail-to-tail orientation, where one of the PRE sequences was mutated to alter its binding affinity (colored lines: median of the population and shaded area: 25- to 75- percentile). The solid black line represents the median response of the consensus PRE (pSYN_3TT), while the dashed line represents the control promoter without PRE. B. Fraction of responding cells for the various PRE sequences. The darker portion of the bar represents the fraction of highly expressing cells (expression output > 50% of reference expression output) and the light bar the fraction of low-expressing cells (> 20% expression output < 50%). The marker represents the fraction of responding cells for individual replicates. The F indicates that the total fraction of expressing cells is significantly lower (t-test: p-val < 0.05) than the expression from the promoter with two consensus PRE sites. C. Dynamics of nuclear enrichment of WT (solid lines) and kar4∆ cells (dashed lines) with a synthetic promoter with 2 PREs spaced by 3 bp in tail-to-tail orientation and where one of the PRE sequences has been mutated to decrease the affinity to Ste12. D. Summary graph displaying the expression output, the speed and the fraction of responding cells for two PRE spaced by 3 bp in tail-to-tail orientation and where one of the PRE sequences has been mutated to decrease the affinity to Ste12 in WT and kar4∆ cells. The color of the marker indicates the difference in response time between the synthetic promoter and the reference pAGA1-dPSTR^Y. The size of the marker represents the fraction of responding cells. The expression output of individual replicates is indicated by small white dots. The dashed line represents the expression output and the dashed dotted line the expression threshold calculated based on the pSYN_3TT. The O and T indicate a significant difference between the mean of the replicates (t-test: p-val < 0.05) in the timing of induction (T) or in the expression output (O) between the WT and kar4∆ strains for the same promoter.

https://doi.org/10.1371/journal.pgen.1011710.g003

As hypothesized in a previous study, the association of Ste12 on promoters containing suboptimal PRE-arrangements seems to be enhanced by Kar4 [17]. Indeed, late pheromone responsive promoters like pFIG1 show minimal expression in kar4∆ cells, while fast ones are independent of Kar4. Therefore, to evaluate the contribution of Kar4 to the expression output of synthetic promoters containing PRE sites with different affinities for Ste12, we measured a set of promoters in kar4∆ cells. On promoters containing a PRE-like site, our measurements suggest that Kar4 has a dual role. On the one hand, Kar4 contributes to the rapid activation of Ste12-bound promoters, while, on the other hand, it limits the level of induction of the promoter (Figs 3C, 3D, S6A and S6B). A promoter with two PRE spaced by 13 bp in tail-to-tail orientation follows the same trend of slow but high induction in kar4∆ cells, while the promoters with 5 and 15 bp in head-to-tail configuration are minimally affected by the same deletion (S6C and S6D Fig). Moreover, the high induction observed in kar4∆ cells for the promoter with a PRE-like seem to act via the activation domain of Ste12 because in the Ste12-EV chimeric protein, we don’t see an effect of the deletion of KAR4 (S6E and S6F Fig).

Taken together, these results indicate that the combination of a PRE with a PRE-like on a promoter tends to decrease the transcriptional output compared to two PRE sites, while the speed of gene induction remains unaffected, thanks to the contribution of Kar4. However, the phenotype associated with the deletion of KAR4 is complex because in the absence of this gene, the dynamics of induction of our synthetic promoters are slowed down, while their expression output is increased.

Location of PRE sites on the promoter

In endogenous promoters, the distance of the PRE sites relative to the core promoter can vary greatly from -150 (pFUS1) to -440 pSTE12. To test the influence of this parameter, we have moved the two PREs from pSYN_3TT from their original position at -223 from the Start codon between -183 to -413 bp. Extending the distance between the Ste12 binding sites and the core promoter leads to a general decline of the expression output of the promoter (Fig 4A and 4B). A dip in expression output and fraction of responding cells is observed at -365 bp, which could be due to the presence of a loosely associated nucleosome. A similar decrease in expression has been observed for the Msn2 stress response TF where moving its binding sites away from the core promoter lowered the expression output of the promoter [42]. However, we demonstrate that this lower expression output is not correlated with a decrease in the speed of gene expression, since all the pSYN tested remain fast (Fig 4C). Many of these promoters are in fact slightly faster than the reference pSYN_3TT, possibly because they have an increased basal expression level. In these synthetic constructs, the binding dynamics of Ste12 should not be modified by the location of the PREs on the promoter. Thus, we speculate that the increased distance between the Ste12 binding sites and the core promoter precludes the ability of Ste12 from activating the general transcription factors via the Mediator.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Distance of the PREs to the core promoter impacts expression output in the synthetic promoter.

A. Dynamics of nuclear relocation for the dPSTR^R controlled by two PREs spaced by 3 bp in tail-to-tail orientation placed at different positions on the promoter (color lines represent the median of the population). The solid black line represents the median response of the PREs placed 223 bp downstream of the start codon (pSYN_3TT), while the dashed line represents the control promoter without PRE. B. Fraction of responding cells for the various PRE sequences. The darker portion of the bar represents the fraction of highly expressing cells (expression output > 50% of reference expression output) and the light bar the fraction of low-expressing cells (> 20% expression output < 50%). The marker represents the fraction of responding cells for individual replicates. The F indicates that the total fraction of expressing cells is significantly lower (t-test: p-val < 0.05) than the expression from the reference construct with the two PRE placed at -223 bp. C. Summary graph displaying the expression output, the speed and the fraction of responding cells for PRE placed at various distances from the Start codon. The color of the marker indicates the difference in response time between the synthetic promoter and the reference pAGA1-dPSTR^Y. The size of the marker represents the fraction of responding cells. The expression output of individual replicates is indicated by small white dots. The dashed line represents the expression output and the dashed dotted line the expression threshold calculated based on the pSYN_3TT. The O and T indicate a significant difference between the mean of the replicates (t-test: p-val < 0.05) in the timing of induction (T) or in the expression output (O) relative to the pSYN_3TT.

https://doi.org/10.1371/journal.pgen.1011710.g004

Interplay between Ste12 and nucleosomes

The modulation of the distance between core promoter and PRE was also tested using a modified AGA1 promoter where all the consensus binding sites were mutated, and a PRE-dimer was moved from -135 to -445 bp relative to the Start codon (Fig 5A). In this context, however, not only the expression levels but also the dynamics of induction were strongly influenced by the location of the PRE-sites (Fig 5B, 5C and 5D). We believe that this behavior can be explained by the position of the nucleosomes on the promoter. Indeed, the promoter activation is rapid and strong when the PREs fall in a nucleosome-depleted region (NDR). When the sites are located within a region protected by a nucleosome, the fraction of responding cells and the level and speed of induction are all attenuated. The most significant impact is observed with the PRE-dimers located at -330 bp, where Ste12 binding is conflicting with a nucleosome centered around -349 bp [43]. Interestingly, the fraction of expressing cells for this promoter is low (18%), but the few cells that express display a substantial nuclear enrichment of the dPSTR^R (Fig 5E). This stochastic activation suggests the presence of a dynamic interplay between the binding of the nucleosomes and the Ste12 dimer. When the nucleosome is bound, no transcription takes place. In some cells, the histones can be displaced by the binding of Ste12, which remains stably associated to the promoter to induce a sizable but delayed expression.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Varying the location of the PRE dimer in an endogenous promoter can influence both level and dynamics of induction.

A. Scheme of the modified pAGA1 promoters with consensus PRE sites indicated by elongated arrowheads and non-consensus PRE-like sites by small arrowheads. The regions protected by the nucleosomes positioned at -100, -349 and -516 are indicated by a lighter color. B. Dynamics of nuclear relocation for the dPSTR^R controlled for a selected set of modified pAGA1 promoters containing two PRE spaced by 3 bp in tail-to-tail orientation placed at different positions along the promoter (colored lines median of the population and shaded area 25- to 75- percentile). The solid black line represents the median response of the reference endogenous AGA1 promoter, while the dashed line represents a mutated pAGA1 where the three PRE and one PRE-like sites are mutated. C. Summary graph displaying the expression output, the speed and the fraction of responding cells for the complete set of the modified pAGA1 promoters. The color of the marker indicates the difference in response time between the synthetic promoter and the reference pAGA1-dPSTR^Y. The size of the marker represents the fraction of responding cells. The expression output of individual replicates is indicated by small white dots. The dashed line represents the expression output and the dashed dotted line the expression threshold calculated based on the pAGA1-dPSTR^R. The gray areas represent the regions of the promoters covered by the three nucleosomes. The O and T indicate a significant difference between the mean of the replicates (t-test: p-val < 0.05) in the timing of induction (T) or in the expression output (O) relative to the pAGA1 with 2 PRE positioned at -230. D. Histograms of the difference in response time between a few selected modified pAGA1-dPSTR^R and the reference pAGA1-dPSTR^Y in the cells that express both constructs. E. Thumbnail images of cells bearing the pAGA1-dPSTR^Y and the modified pAGA1-dPSTR^R with the PRE dimer positioned 330 base pairs before the Start codon. While all cells in the image relocate the dPSTR^Y, only one cell displays a relocation of the dPSTR^R. The scale bar represents 5 µm.

https://doi.org/10.1371/journal.pgen.1011710.g005

To test this model more directly, the promoter sequence was mutated to introduce a nucleosome disfavoring sequence (poly dA/dT, 20 bp long) in the vicinity of the Ste12 binding sites (Fig 6A). The fraction of expressing cells increased from 20% to 50% when the dA/dT element was placed 6 bp away from the PRE sites (Fig 6B and 6C). An alternative option to increase the binding efficiency of Ste12 in the nucleosome bound region is to use multiple PRE sites. We used conformations with 3 PREs present in PRM1 and FUS1. The multiple sites and the high A/T content of the element facilitated the association of Ste12 to the promoter and resulted in an increased fraction of transcribing cells (Fig 6B and 6C). However, for all these constructs, the dynamics of induction were slow compared to the endogenous pAGA1. Interestingly, the addition of the dA/dT sequence slightly accelerated the induction of the reporter, suggesting that in a fraction of the population Ste12 can associate to the promoter under basal conditions (Fig 6D).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Favoring nucleosome eviction increases expression output.

A. Scheme of the various promoters tested. Starting from the modified pAGA1 with 2 PREs in tail-to-tail orientation at -330, poly-dA/dT segments of 20 bp were generated by mutations 6 or 20 bp away from the 2 PREs. Alternatively, the two PREs were replaced by three PRE sites extracted from the FUS1 or PRM1 promoters. B. Dynamics of nuclear relocation for the dPSTR^R controlled by modified pAGA1 promoters. The four different modified promoters tested are shown in color. The solid black line represents the median response of the reference endogenous AGA1 promoter, while the dashed line represents the median of the population of cells with the 2 PREs in tail-to-tail orientation at -330. C. Summary graph displaying the expression output, the speed and the fraction of responding cells. The color of the marker indicates the difference in response time between the synthetic promoter and the reference pAGA1-dPSTR^Y. The size of the marker represents the fraction of responding cells. The dashed line represents the expression output and the dashed dotted line the expression threshold calculated based on the pAGA1-dPSTR^R. The O and T indicate a significant difference between the mean of the replicates (t-test: p-val < 0.05) in the timing of induction (T) or in the expression output (O) relative to the pAGA1 with 2 PRE positioned at -330. The inset represents the fraction of responding cells for each promoter. The dark portion of the bar represents the fraction of highly expressing cells (Expression output > 50% EO of the reference promoter) and the light bar the fraction of low-expressing cells (20% < EO < 50%). The marker represents the fraction of responding cells for individual replicates. The F indicates that the fraction of responding cells is significantly different relative to the pAGA1 with 2 PRE positioned at -330. D. Histograms of the response time difference measured between the modified pAGA1-dPSTR^R and the reference pAGA1-dPSTR^Y. The solid black line corresponds to the endogenous AGA1 promoter, while the dashed line represents the 2 PREs in tail-to-tail orientation at -330. The addition of the poly-dA/dT 6 bp away from the PRE dimer (light green) slightly accelerates the induction of the promoter. Placing the three PREs of FUS1 (blue) slows down the induction of the promoter. The T indicates that the histograms for the two mutated promoters are significantly different from the endogenous promoter using a Wilcoxon rank sum test. E. Basal association of Ste12 on various pAGA1 promoters evaluated by Ch-IP in a region between -327 and -205. The values are normalized relative to the no TAG control. The star indicates a significant difference (t-test: p-val < 0.05) relative to the promoter with 2 PRE at position -330. F. Fold induction in Ste12 binding measured by Ch-IP after 30 min treatment with pheromone. The star indicates a significant difference between the untreated controls and the α-factor stimulated samples (t-test: p-val < 0.05). G. Eviction of nucleosomes monitored by MNase protection assays after 30min pheromone treatment in the same region as in E (between -327 and -205). Nucleosome occupancy is normalized relative to the untreated pAGA1 endogenous promoter. A ✳ indicates a significant difference between the untreated control and the α-factor stimulated samples and the ⭐ indicates a significant difference between the various mutated promoters in the untreated conditions relative to the pAGA1 (t-test: p-val < 0.05).

https://doi.org/10.1371/journal.pgen.1011710.g006

To probe the association of Ste12 on the DNA, Chromatin Immuno-Precipitation (Ch-IP) was performed on some of these promoter variants by amplifying a region centered around -259 bp. The basal association of Ste12 is much higher in the endogenous AGA1 sequence than in our promoter variants (Fig 6E). In addition, we see a small but significant enrichment of Ste12 on the promoter when the dA/dT stretch is added 6 bp away from the two PREs compare located at -330. Upon pheromone treatment Ste12 becomes more than 10-fold enriched on the endogenous promoter while we don’t detect a significant enrichment of the TF on the 2 PRE placed at -330 (Fig 6F). In contrast, promoters modified with the 3 PREs using the FUS1 conformation or with the dA/dT stretch 6 bp away from the PREs, the enrichment of Ste12 is significant. In parallel, the association of histones in the same region of these promoters was evaluated by MNase protection assays. On the WT pAGA1, we observe a strong eviction of the nucleosomes 30 minutes after the stimulation of the cells with α-factor (Fig 6G). No significant eviction can be observed on the promoter with the 2 PRE at -330 bp, while the eviction is recovered when the dA/dT stretch is added 6 bp away from the two PREs or if three PREs are present. Both the recruitment of Ste12 and the eviction of the nucleosomes are in line with the higher expression output measured with the dPSTR^R. Note that under basal conditions, the nucleosome occupancy in this region is slightly lower for the promoter with 2 PRE at -330 bp, possibly because the position of the nucleosome is slightly shifted in this construct. These biochemical data confirm the hypothesis that the access to the PRE sites at -330 is limited for Ste12 under basal conditions which limits its ability to activate transcription from this site. However, the addition of a third PRE or the presence of a dA/dT stretch favors the access of Ste12 to the promoter by facilitating the eviction of the nucleosomes upon pheromone treatment.

The endogenous pAGA1 promoter contains three consensus PREs and at least five PRE-like sites (S7A Fig). The PRE/PRE-like sites spaced by 3 bp at position -220 alone seem sufficient to ensure a high induction level. Therefore, it is not clear why there are so many additional Ste12 binding sites present in this sequence. One of their roles might be to increase the local concentration of Ste12 in the vicinity of the locus. Interestingly, if the first PRE at -196 bp or its neighboring PRE-like at -185 bp are mutated, the level of induction of the promoter remains almost identical, while the dynamics of induction are delayed (S7B and S7C Fig). Since these two PREs are 5 bp apart in a head-to-tail conformation, it is likely that they allow the formation of a Ste12 dimer. One possible explanation for the delayed expression observed is that these two binding sites contribute to defining the nucleosome depleted region in the pAGA1 promoter which favors the basal association of Ste12 to the central PRE of the promoter. Regardless, together, these results demonstrate that nucleosomes prevent the basal association of Ste12 outside of the NDR. Sites present in a nucleosome protected region of the promoter will display slow induction kinetics and a low fraction of inducing cells. Additional PRE sites on endogenous promoters may contribute to shaping the NDR and thereby allowing indirectly an early activation of transcription.

Role of Kar4

Previous measurements have suggested that Kar4 interacts directly with the DNA and that late mating genes contain specific binding sites for Kar4 [18]. However, we favor a model where Kar4 is recruited to the promoters via its interaction with Ste12. Multiple evidences point in this direction. We have shown previously that Kar4 is recruited to the promoters of both early and late genes [17]. The yeast epigenome project has identified a binding site for Kar4 on DNA which corresponds to the one of Ste12 [43]. Recent point mutants of Kar4 which display mating deficiencies have a decreased interaction with Ste12 [45].

Our present data demonstrate that Kar4 has a dual effect on some of our synthetic promoters. First, deleting KAR4 results in a delayed induction for promoters with lower affinity Ste12 binding sites. This suggests that Kar4 contributes to the stabilization of Ste12 on these weaker sites under basal conditions such that upon stimulation of the pathway, these promoters can be activated promptly. In parallel, we observe an increase in the level of induction from these same promoters containing PRE-like sites in kar4∆ cells. This surprising behavior could act via the activation domain of Ste12 since we don’t observe a similar increase in cells expressing the Ste12-EV chimeric TF in the absence of KAR4.

Absence of Kar4 slows down the induction of intermediate genes (PRM1, FIG2) and is required for the activation of late genes such as FIG1 or KAR3 [17]. These late promoters bear PRE-like combined with PRE which are hidden by nucleosomes. In the absence of Kar4, the binding of Ste12 to the promoter is presumably too weak to be stabilized and displaced by the nucleosomes without resulting in productive transcription whenever it binds.

Dynamics of promoter induction

We identified two strategies to achieve a slow induction of the promoter: either by restricting the access of Ste12 to the promoter by nucleosomes or by using an unconventional PRE dimer (23 bp tail-to-tail or 15 bp head-to-tail or head-to-head). In the two dozen mating-dependent promoters that we have analyzed, we identified only one with 15 bp spacing. Thus, the nucleosome protection may be the preferred option, because it allows tuning both the speed and level of induction. Indeed, the large spacing between two PRE sites can only achieve a slow and low level of gene expression. In contrast, when the PRE dimer is moved within the pAGA1 promoter delayed but high induction can be generated. Positioning the PRE dimer under the nucleosome slows down the induction but also strictly reduces the fraction of responding cells. However, placing the binding sites closer to the edge of a nucleosome protected region seems to be sufficient to obtain a similar decrease in the speed of induction without affecting severely the expression level and, importantly, the fraction of responding cells. In the promoter of the late gene FIG1, a PRE dimer (5 bp, tail-to-head) essential for gene expression is also placed at the boundary of the sequence protected by the nucleosome [17]. This positioning of the PRE dimers delays the induction of FIG1 until it is required for the fusion of the two mating cells without compromising its robust expression when it is needed. Positioning the PREs further into the nucleosome protected region could lead to a stochastic activation of the protein which could be detrimental for the mating outcome.

The timing of gene expression is crucial for many processes, from the rapid induction of stress response genes to the controlled induction of proteins during the cell cycle. Cell-fate decisions are characterized with a chronology of gene expression which requires the ability to tune both the level and the dynamics of gene expression independently of each other. Regulating the basal association of Ste12 on promoters via the positioning of nucleosomes offers the opportunity to control the chronology of gene expression during the mating process. With our improved understanding of the regulation of the mating gene expression, we are in a better position to perturb the timing of key proteins implicated in mating to verify the importance of this chronology on the mating process.

Yeast strains and plasmids

The synthetic promoters were derived from a plasmid containing a slightly modified pAGA1-dPSTR^R plasmid [17] which contains a ClaI site at position 144 bp separating the regulatory region of AGA1 (-1000 to -150) and the core region (-150–0). First the core promoter was replaced by the core promoter from CYC1 (-180–0) cloned between ApaI and ClaI. Then the regulatory region (between ClaI and AatII) was replaced by fragments from a modified pCYC1 promoter (138 bp) which destabilizes the association of nucleosomes [37]. This fragment encodes various conformations of Ste12 binding sites. The distance to the Start mentioned in the text and figures corresponds to the distance to the end of the CYC1 or AGA1 promoters. The actual Start of the inducible dPSTR moiety is 53 bp downstream due to the presence of multiple restriction sites. The longer spacing between Start and PRE sites (Fig 4) were obtained by duplicating the CYC1 regulatory region. The core promoter and the multiple UAS sequences were synthesized as double stranded DNA fragments (IDT) or as plasmids (GeneScript). The list of synthetic sequences is provided in S1 Table. The plasmids obtained were verified by restriction digestion and sequencing and were transformed in the same reference strain from the W303 background with the histone Hta2 tagged with CFP (pGTH-CFP) [46] and containing the reference pAGA1-dPSTR^Y [17]. The strains used in this study are listed in S1 Table.

Gene deletions were performed using the pFA6a KAN cassette [47]. Correct insertion of the deletion cassette was verified by PCR on genomic DNA. The human Estradiol receptor and the VP16 activation domain (EV) [39,40] were cloned in a pGT-NAT plasmid between BamHI and NheI. The Ste12-EV chimeric transcription factor was generated by transforming the EV sequence and NAT cassette with primers containing homology to the sequence flanking the activation domain of Ste12 (from 647 bp to STOP codon) in strains containing the selected synthetic promoters controlling the dPSTR^R. The transformants were selected on SD-UHL + NAT plates. The correct insertion of the EV sequence in frame with the Ste12 ORF was controlled by sequencing PCR fragments obtained from genomic DNA. The synthetic transcription factor Z4-EV is placed under the control of the RPL30 promoter [48]. The Venus is control by a modified pGAL1 containing 6 Z4 binding sites. The pZ4_1BS and pZ4_2BS are based on the pCYC1 promoter containing either one or two Z4 binding sites and driving the expression of the dPSTR^R.

Typically, eight transformants were selected and screened visually to verify the expression level of the three fluorescent constructs (Hta2-CFP, pAGA1-dPSTR^Y, pSYN-dPSTR^R). Four transformants were then quantified in time-lapse experiments and used as biological replicates for the measurements of one promoter. Based on the consistency of the pAGA1-dPSTR^Y response and the behavior of the tested dPSTR^R, we excluded some of the replicates. If two or more replicates were not providing consistent results, the experiment was repeated with different transformants. In each experiment, the behavior of 500 cells is typically quantified, but some replicates contain only 200 cells and others up to 1000 individual cells. In the dPSTR nuclear relocation graphs and in the response time histograms, we represent the behavior of one representative replicate. In the bar graphs displaying the fraction of responding cells or in the summary graphs, the mean behavior of all the selected replicates is plotted.

Time-lapse microscopy

Yeast strains were grown overnight to saturation in SD-full medium (Complete CSM DCS0031, ForMedium) diluted in the morning in fresh SD-full and grown for at least 4 hours. The cultures were diluted to OD 0.04 and briefly sonicated and 200µl were loaded in the well of a 96-well plate (PS96B-G175, SwissCI) previously coated with Concanavalin A (L7647, Sigma-Aldrich). Cells settled in the well for 30 minutes before the start of the time-lapse.

Cells were imaged on an inverted wide-field epifluorescence microscope (Nikon, Ti2) enclosed in an incubation chamber set at 30° using a 40X Oil or a 40X AIR objective (Figs 2B to 2D, S3E, S3F, S5E, S5F and S6C to S6F). The fluorescence excitation is provided by a Lumencor Spectra 3 light source (LED intensity 50%). A CFP YFP RFP dichroic filter (F68-003) and appropriate emission filters were used to detect the fluorescence emission using a sCMOS camera (Hamamatsu, Fusion BT), with 50ms, 50ms and 100ms exposure time respectively. Up to 8 wells were imaged in parallel and 5 fields of view per well were monitored. Two brightfield images (one slightly out of focus) and the three fluorescent images were recorded every 5 minutes for 100 minutes. Before the third time point, 100µl of a 3µM α-factor solution in SD-full was added to each well to reach a final concentration of 1µM in the well. The α-factor is a gift from the Peter lab at the ETHZ. To induce the Ste12-EV construct, 100µl of a 3µM β-estradiol (Sigma-Aldrich, E2758-250MG) solution in SD-full was used to reach a final concentration of 1µM in the well.

Data analysis

Time-lapse images were segmented and single cell data extracted using the YestQuant platform [49]. The CFP image allowed to determine the position of the nuclei of each cell. Using this first object, the two brightfield images were used to identify the border of each cell and define the cytoplasm object by removing the pixels belonging to the nucleus.

The quantification of the single cell traces was performed with Matlab (The Mathworks, R2023b). Only cells tracked from the beginning to the end of the timelapse were kept for the analysis. The nuclear enrichment is calculated as the difference between the average fluorescence of the nucleus object and the cytoplasm object. The basal nuclear enrichment is measured as the mean of the three first time points. The Expression Output (EO) corresponds to the difference between the maximum of a single cell traces and the basal level (S2D Fig).

To determine if a cell is deemed transcribing or not, two criteria are used. First, the last 5 points of the trace minus the basal level has to be significantly higher than zero (sign-test. P-value 0.05). Second, the Expression output of the trace (maximum of the trace – basal level) has to exceed a threshold. To define the threshold, we select one strain as the reference and calculate the threshold as 20% of the mean Expression ouput from all cells. The strain bearing the pAGA1-dPSTR^R (Figs 5, 6, S2, S3 and S7) or the pSYN_3TT-dPSTR^R (Figs 1, 3, 4 and S4) were used as references. Transcribing cells are further differentiated in strong and weak expression if their output exceed 50% of the Expression output of the reference strain and as weakly expressing if the expression output falls between 20% to 50% of the reference (S2D, S2E and S2F Fig).

The response time is defined as the first time point after the trace overcomes 20% of its own expression output (S2D Fig). The difference in response time between the tested dPSTR^R and the internal reference construct pAGA1-dPSTR^Y is only calculated for cells considered as transcribing both reporters (S2G and S2H Fig).

ChIP assays

Yeast cultures were grown to early log phase (A₆₆₀ 0.4–0.6), then samples (50ml) were subjected to 1µM α-factor for 30 minutes. For crosslinking, yeast cells were treated with 1% formaldehyde for 20 minutes at room temperature. Glycine was added to a final concentration of 330 mM for 15 minutes. Cells were collected, washed four times with cold TBS (20 mM Tris-HCl, pH 7.5, 150 mM NaCl), and kept at −20 °C for further processing. Cell pellets were resuspended in 0.3 ml cold lysis buffer (50 mM HEPES-KOH, pH 7.5, 140 mM NaCl, 1 mM EDTA, 0.1% sodium deoxycholate, 1% Triton-X 100, 1 mM PMSF, 2 mM benzamidine, 2 µg/mL leupeptin, 2 µg/mL pepstatin, 2 µg/mL aprotinin). An equal volume of glass beads was added, and cells were disrupted by vortexing (with Vortex Genie) for 13 minutes at 4ºC. Glass beads were discarded and the crosslinked chromatin was sonicated with water bath sonicator (Bioruptor) to yield an average DNA fragment size of 350 bp (range, 100–850 bp). Finally, the samples were clarified by centrifugation at 16,000g for 5 minutes at 4°C. Supernatants were incubated with 50µL anti-HA 12CA5 monoclonal antibodies pre-coupled to pan mouse IgG Dynabeads (Invitrogen, 11042). After 120 minutes at 4°C on a rotator, beads were washed twice in 1mL lysis buffer, twice in 1mL lysis buffer with 500 mM NaCl, twice in 1mL washing buffer (10 mM Tris-HCl pH 8.0, 0.25 M LiCl, 1 mM EDTA, 0.5% N-P40, 0.5% sodium deoxycholate) and once in 1mL TE (10 mM Tris-HCl pH 8.0, 1 mM EDTA). Immunoprecipitated material was eluted twice from the beads by heating for 10 minutes at 65 °C in 50 µl elution buffer (25 mM Tris-HCl pH 7.5, 1 mM EDTA, 0.5% SDS). To reverse crosslinking, samples were adjusted to 0.3 ml with elution buffer and incubated overnight at 65°C. Proteins were digested by adding 0.5mg/ml Proteinase K (Novagen, 71049) for 1.5 hours at 37°C. DNA was extracted with phenol-chloroform-isoamyl alcohol (25:24:1) and chloroform. It was finally precipitated with 48% (v/v) of isopropanol and 90 mM NaCl for 2 hours at −20 °C in the presence of 20 µg glycogen, and resuspended in 30 µL of TE buffer. Quantitative PCR analysis on pAGA1-dPSTR-R used the following primers with locations indicated by the distance from the respective ATG initiation codon: AGA1 promoter (-310/-207); and TEL (telomeric region on the right arm of chromosome VI). Experiments were done on three independent chromatin preparations and quantitative PCR analysis was done in real time using an Applied Biosystems Via7. Immunoprecipitation efficiency was calculated in triplicate by normalizing the amount of PCR product in the immunoprecipitated sample by that in TEL sequence control. The binding data are presented as fold induction with respect to the non-treated condition, for basal binding of Ste12 the data are referenced to the untagged strain (no tag) which was set to 1.

MNase nucleosome mapping

Yeast spheroplast preparation and micrococcal nuclease digestions were performed as described previously with modifications [50,51]. Ste12-6xHA tagged strain was grown to early log phase (A660 0.4–0.6) and samples of 500 ml of culture were exposed to 1 µM α-factor for 30 minutes. The cells were cross-linked with 1% formaldehyde for 15 minutes at 30°C and the reaction was stopped with 125 mM glycine for minutes. Cells were washed and resupended in 1M sorbitol TE buffer before cell wall digestion with 100 T zymoliase (USB). Cells were then lysed and immediately digested with 60–240 mU/µl of micrococcal nuclease (Worthington Biochemical Corporation, Lakewood; NJ., USA). DNA was subjected to electrophoresis in a 1.5% (w/v) agarose gel and the band corresponding to the mononucleosome was cut and purified using a QIAquick gel extraction kit (Qiagen). DNA was used in a real-time PCR with specific tiled oligonucleotides covering the AGA1 promoter of the dPSTR^R (the endogenous AGA1 gene has been mutated). PCR quantification was referred to an internal loading control (telomeric region in chromosome 6) and nucleosome occupancy was normalized to 1 at the (-1) nucleosome region of the untreated condition. The degree of nucleosome eviction at the indicated region was set to 1 in the wild type strain in control conditions and used as a reference.