While the SBF and MBF transcription factors have been heavily studied in the context of the mitotic cell cycle, their involvement in regulating the meiotic transcriptional program is not well understood. To understand the regulation and function of these complexes during meiosis, we first monitored the levels of SBF and MBF subunits throughout a meiotic time course. Cells were first grown in rich media overnight and were then transferred to pre-sporulation media. After additional overnight growth, cells were shifted to sporulation media (SPO) to induce meiosis, and samples were taken hourly for protein extraction and immunoblotting to monitor the abundance of each subunit. Unlike the mitotic G1/S transition (Kelliher et al., 2018), meiotic entry resulted in ∼30% decrease in Swi4 levels after 2 hours, while Mbp1 and Swi6 levels were increased (Figure 1B and 1C). These data indicate that Swi4 is the sole subunit within SBF and MBF whose level declines during meiotic entry and are consistent with a published mass spectrometry dataset (Cheng et al., 2018).

We next examined the expression of a set of well-characterized transcripts that are regulated by SBF or MBF (Iyer et al., 2001; Simon et al., 2001; Bean et al., 2005; Harris et al., 2013; Smolka et al., 2012) using a published meiotic RNA-seq dataset (Brar et al., 2012). This analysis revealed that compared to mitotically dividing cells, most of the SBF-specific target genes had either low or no expression during early meiosis (Figure 1D). In contrast, most of the MBF targets displayed increased expression upon meiotic entry (Figure 1D).

Based on our findings thus far, we propose that Swi4 levels are downregulated in meiosis to ensure that SBF targets, including CLN1 and CLN2, are turned off during meiotic entry. If so, then higher Swi4 levels should lead to increased expression of SBF targets. To test this possibility, we overexpressed SWI4 by placing it under the regulation of the ATG8 promoter (pATG8-SWI4), which is highly expressed in meiosis (Brar et al., 2012). The steady-state level of pATG8-driven Swi4 was 5 times higher than wild type (Figure S1A and S1B). Using reverse transcription coupled with quantitative polymerase chain reaction (RT-qPCR), we measured the transcript levels of well-characterized SBF targets and observed a significant increase in the expression of CLN1 and CLN2 in pATG8-SWI4 cells relative to wild type (Figure 1E, p = 0.0351 [CLN1], p = 0.0013 [CLN2], two-tailed t-test). In contrast, the MBF-specific targets CDC21 and RNR1 remained similar (Figure 1E, p = 0.8488 [CDC21], p = 0.0859 [RNR1], two-tailed t-test). These findings indicate that upregulation of SWI4 is sufficient to induce the expression of SBF-specific targets CLN1 and CLN2 in meiosis without affecting MBF-specific targets.

To test whether the downregulation of SWI4 is functionally important for meiotic progression, we used time-lapse fluorescence microscopy and visualized the kinetics of meiotic divisions in pATG8-SWI4 cells relative to wild type. By tracking the endogenous histone H2B fused to the red fluorescent protein mCherry (Htb1-mCherry), we found that SWI4 overexpression caused a significant delay in meiotic progression (Figure 1F, p = 0.0045, Mann-Whitney test). We conclude that downregulation of SWI4 is necessary for timely meiotic progression.

Since Swi4 abundance, and by inference SBF activity, was decreased during the mitosis-to-meiosis transition, we hypothesized that the meiotic progression delay observed in pATG8-SWI4 cells was due to meiotic entry defects. To test this we monitored Ime1, a meiosis-specific transcription factor and a master regulator of meiotic entry (Kassir et al., 1988; Honigberg, 2003; reviewed in van Werven and Amon, 2011). To quantify the bulk levels of Ime1 protein in meiosis, we performed immunoblotting. During meiotic entry (2 h in SPO), when Swi4 abundance was elevated, there was a 50% reduction in Ime1 levels in pATG8-SWI4 cells compared to wild type (Figure S1C and S1D). In parallel, we monitored meiotic entry on a single-cell basis by measuring the localization of endogenous Ime1 carrying an N-terminal green fluorescent protein tag (GFP-IME1; Moretto 2018) and Htb1-mCherry. Compared to wild type where > 90% of the cells had nuclear Ime1 following meiotic entry (2 h in SPO), only ∼50% of pATG8-SWI4 cells had nuclear Ime1, which was significantly lower (Figure 2A and 2B, p = 0.0169, two-tailed t-test).

Given that the increase in Swi4 levels coincided with a decrease in Ime1 protein expression and nuclear localization, this raised the possibility that Swi4 interferes with Ime1 function. To further investigate the relationship between these two transcription factors at a single-cell level, we generated strains carrying endogenous fluorescent protein tags for each transcription factor (GFP-IME1, SWI4-mCherry). Using DAPI as a nuclear marker, we measured the mean nuclear intensity of each transcription factor before (0 h) and after meiotic entry (4 h) (see Materials and Methods for details on image quantification). In wild type, most cells exhibited decreased nuclear Swi4 and increased nuclear Ime1 upon meiotic entry (Figure 2C-E). Conversely, in the pATG8-SWI4 mutant, there was a significant shift in the fraction of cells with higher nuclear Swi4 levels (Figure 2D and 2E, p < 0.0001, Mann-Whitney test,). Additionally, the cells with increased nuclear Swi4 had reduced levels of nuclear Ime1 (Figure 2E, p < 0.0001, Mann-Whitney test). This inverse relationship between Swi4 and Ime1 nuclear localization further indicates that higher levels of Swi4 are antagonistic to Ime1 function.

Ime1 is necessary for the expression of early meiotic genes. Therefore, we next characterized the changes in the early meiotic transcriptome upon SWI4 overexpression by mRNA-sequencing (mRNA-seq). Using DESeq2 (Love et al., 2014), we identified differentially expressed genes in pATG8-SWI4 mutant compared to wild type during meiotic entry (2 h in SPO). To investigate general pathways being affected by increased Swi4 levels in meiosis, we ran gene ontology analysis of statistically significant (padj < 0.05) output from DESeq2, which revealed that the genes with significantly increased expression were involved in mitosis while those with significantly decreased expression were involved in meiotic processes (Figure S2A and S2B). Regarding the SBF targets, we noticed increased expression of many genes, including CLN1 and CLN2 (Figure 2F, Supplemental Table 1). We focused on a cluster of previously identified early meiotic genes involved in DNA replication and recombination (Brar et al., 2012) and observed that more than 50% had a significant decrease in their expression upon SWI4 overexpression (padj < 0.05, Figure 2F, Supplemental Table 1). Finally, gene set enrichment analysis was performed to test whether expression of either the early meiotic gene set or SBF target gene set is enriched in wild type or pATG8-SWI4 mutant cells (Subramanian et al., 2005). This gene set level analysis revealed significant enrichment of SBF target gene expression (NES = 2.88, p < 0.001, Figure 2G), as well as significant disenrichment of early meiotic gene expression (NES =-2.88, p < 0.001, Figure 2G) in the pATG8-SWI4 mutant. Altogether, these findings demonstrate that the increased levels of Swi4 during transition from mitotic to meiotic cell fate abruptly activates SBF and disrupts the early meiotic transcriptome, highlighting the importance of SWI4 regulation.

As an orthogonal approach, we performed live-cell imaging of Rec8, endogenously tagged with GFP (Rec8-GFP), which is a meiosis-specific cohesin subunit and a direct transcriptional target of Ime1 (Primig et al., 2000; Klein et al., 1999). Htb1-mCherry was used as a nuclear marker. This analysis revealed a significant delay in Rec8-GFP nuclear appearance in the pATG8-SWI4 strain compared to wild type (Figure 2H and 2I, p = 0.0005, Mann-Whitney test). Furthermore, sporulation efficiency was decreased by ∼20% in the pATG8-SWI4 strain relative to wild type (Supplemental Table 2). These findings are consistent with the mRNA-seq data and further underscore the importance of SWI4 downregulation in establishing a robust meiotic cell fate.

Overexpression of the G1 cyclins CLN1, CLN2 and CLN3 has been previously shown to inhibit meiotic entry (Colomina et al., 1999). Given that CLN1 and CLN2 are transcriptional targets of SBF, we next determined whether the meiotic progression delay in pATG8-SWI4 cells could be due to increased G1 cyclin protein levels. We first examined Cln1 and Cln2 protein levels using epitope-tagged alleles at their endogenous loci (CLN1-3V5 or CLN2-3V5) in wild-type and pATG8-SWI4 cells. In response to SWI4 overexpression, we observed up to 2-fold and 10-fold increase in Cln1 and Cln2 protein levels, respectively, corroborating the mRNA-seq data (Figure 3A-D).

To determine whether these two G1 cyclins are functionally responsible for the meiotic progression delay observed in response to SBF misregulation, we performed time-lapse fluorescence microscopy in cells carrying Rec8-GFP and Htb1-mCherry. We found that deletion of either CLN1 or CLN2 significantly rescued the meiotic progression delay in the pATG8-SWI4 mutant (p = 0.0111 [cln1Δ], p = 0.0478 [cln2Δ], Mann-Whitney test) (Figure 3E and 3F). The kinetics of meiotic entry in the absence of either Cln1 or Cln2 closely resembled that of wild type (Figure S3A), indicating that the observed rescue was specific to the meiotic defect resulting from elevated Swi4 levels. However, pATG8-SWI4 cells lacking either CLN1 or CLN2 were still delayed compared to wild type, suggesting redundancy. We were unable to examine meiotic progression in the cln1Δ cln2Δ double mutant due to its severe sickness in the SK1 background. Nevertheless, our analyses establish a causal link demonstrating that both CLN1 and CLN2 contribute to the meiotic defects arising from SBF misregulation.

Given the partial rescue of the meiotic delay in pATG8-SWI4 mutants by the deletion of individual G1 cyclins, we next investigated how the G1 cyclins interfere with meiosis. To this end, we generated transgenes that expressed either CLN1 or CLN2 under the control of the pATG8 promoter for meiotic overexpression. While wild-type cells successfully completed meiosis with more than 94% sporulation efficiency, CLN2 overexpression resulted in only 8.5% of cells forming gametes. Overexpression of CLN1 also resulted in a meiotic defect, albeit to a lesser extent than the pATG8-CLN2 mutant (65% sporulation efficiency, Figure 4A).

Since the meiotic defect was more pronounced in response to CLN2 overexpression, we decided to use the pATG8-CLN2 mutant to explore how G1 cyclins counteract meiosis. To assess meiotic entry, we performed fixed-cell imaging for GFP-Ime1 and found that when CLN2 was overexpressed, only 5% of the pATG8-CLN2 cells displayed nuclear GFP-Ime1 signal during early meiosis as opposed to >84% of wild-type cells (Figure 4B and 4C, 2 h in SPO). In parallel, we measured IME1 transcript and Ime1 protein levels. In both cases, we observed ∼30% decrease in abundance in the pATG8-CLN2 mutant relative to wild type (Figure S4A-C, 2 h in SPO). These data raise the possibility that the meiotic entry defect observed in the pATG8-CLN2 mutant arises from downregulation of IME1 expression.

To test whether the increased levels of Ime1 can rescue the meiotic defect of pATG8-CLN2 mutant, we replaced the endogenous IME1 promoter with a copper-inducible CUP1 promoter (Figure 4D). We adapted a previously well-characterized overexpression allele of IME1, pCUP1-IME1, (Berchowitz et al., 2013; Chia and van Werven, 2016) and included a functional, N-terminal GFP tag to track Ime1’s subcellular localization (pCUP1-GFP-IME1). Use of the CUP1 promoter was successful in elevating IME1 transcript and protein levels in the presence of CLN2 overexpression (Figure S4D-F). Despite the rescue of IME1 expression, gamete formation was still severely perturbed (Figure 4A). Using a single z-slice to measure mean nuclear intensity, we noticed that the intensity of the nuclear Ime1 signal was significantly lower in the pATG8-CLN2; pCUP1-GFP-IME1 cells compared to the pCUP1-GFP-IME1 control (Figure 4E and 4F, p < 0.0001, Mann-Whitney test). This finding indicates that rescue of IME1 expression did not also rescue sporulation, further suggesting a defect in Ime1 nuclear localization. To test this possibility, we utilized a nanobody trap strategy (Fridy et al., 2014) where we C-terminally fused a single-domain anti-GFP antibody to Pus1, a constitutively nuclear localized protein (PUS1-⍺GFP, Figure 4G). In this background, control strains carrying a GFP-IME1 allele sporulated efficiently, demonstrating that tethering of Ime1 to Pus1 does not interfere with Ime1 function (Figure 4A). Furthermore, mean nuclear intensity of GFP-Ime1 was indistinguishable between PUS1-⍺GFP and pATG8-CLN2; PUS1-⍺GFP, indicating that nuclear localization was fully rescued (Figure 4H and 4I, p = 0.6563 [PUS1-⍺GFP vs. pATG8-CLN2; PUS1-⍺GFP], Mann-Whitney test). Surprisingly, these cells still failed to undergo meiosis (Figure 4A), suggesting that G1 cyclins interfere with Ime1 function at an additional step beyond misregulating its expression and localization. Alternatively, G1 cyclins could disrupt a different meiotic factor in addition to Ime1.

To induce early meiotic genes, Ime1 must interact with another transcription factor called Ume6 (Rubin-Bejerano et al., 1996). Since Ime1 itself does not possess a DNA-binding domain, its binding to Ume6 is essential for targeting Ime1 to early meiotic gene promoters (Smith et al., 1993; Rubin-Bejerano et al., 1996). To address whether the G1 cyclins might disrupt the interaction between Ime1 and Ume6, we fused the anti-GFP nanobody to Ume6 (UME6-⍺GFP) in the pATG8-CLN2 strain carrying a GFP tagged Ime1 (Figure 4J). This nanobody trap should lead to constitutive tethering of Ime1 to Ume6, as evidenced by the rescue of Ime1 nuclear localization (Figure 4K and 4L, p = 0.035 [wild type vs. pATG8-CLN2 UME6-⍺GFP] Mann-Whitney test). Under these conditions, the sporulation defect of pATG8-CLN2 mutant was rescued, reaching similar levels to wild type (Figure 4A). Since IME1 is expressed from its endogenous promoter in the pATG8-CLN2; UME6-⍺GFP strain, these data suggest that overexpression of G1 cyclins results in meiotic failure due to reduced Ime1-Ume6 interaction.

Our findings thus far highlight the biological significance of restricting SBF activity during the transition from mitotic to meiotic cell fate. Among the SBF targets, G1 cyclins pose a major block to meiotic entry by interfering with Ime1 function, critically at the level of Ime1-Ume6 interaction. While our findings emphasize the importance of SWI4 regulation to ensure timely SBF activity, the question remains as to how SBF activity is downregulated during transition from mitosis to meiosis.

Given the importance of SWI4 downregulation in restricting SBF activity during mitotic to meiotic cell fate transition, we next investigated the mechanism of SWI4 downregulation. A previous study identified a long undecoded transcript isoform (LUTI) expressed from the SWI4 locus in meiotic cells (Brar et al., 2012, Figure 5A). When SWI4^LUTI is expressed, the canonical protein-coding SWI4 transcript, SWI4^canon, is downregulated (Tresenrider et al., 2021, Figure 5B), suggesting that LUTI expression restricts Swi4 protein levels and thus SBF activity during meiotic entry.

To further investigate LUTI-based repression of SWI4, we examined the relationship between SWI4^LUTIand SWI4^canon transcripts using single molecule RNA fluorescence in situ hybridization (smFISH: Chen et al., 2017, 2018; Raj et al., 2008). Two distinct probes were used: one, conjugated to Quasar 670 (Q670) and complementary to the SWI4 coding sequence (CDS) and the other, conjugated to CAL Fluor Red 590 (CF590) and unique to the 5′ extended LUTI sequence. Accordingly, a spot where the two probe sets colocalized indicated a SWI4^LUTI transcript, whereas a spot marked with Q670 probe alone highlighted a SWI4^canon transcript. We used a well-established meiotic cell synchronization system (Berchowitz et al., 2013) to investigate the precise temporal expression of these two mRNA isoforms. In comparison to premeiotic state, meiotic cells displayed a significant increase in SWI4^LUTI transcripts (p < 0.0001, Mann-Whitney test) as well as a significant decrease in SWI4^canon transcripts (p = 0.0007, Mann-Whitney test) (Figure 5C and 5D), thus confirming their inverse expression pattern.

To characterize the functional contribution of the LUTI to SWI4 downregulation, we eliminated SWI4^LUTI production (ΔLUTI) by deleting its promoter (Tresenrider et al., 2021) and used RNA blotting to visualize the SWI4 mRNA isoforms. In wild-type cells, SWI4^LUTI was readily detectable after 1 hour in SPO, corresponding to early meiotic entry (Figure 5E). During the time points when SWI4^LUTI was highly expressed, SWI4^canontranscript levels were lower, consistent with the smFISH data. In ΔLUTI cells, we observed an increase in the abundance of SWI4^canon mRNA (Figure 5E). This finding corroborates previous reports, where LUTI expression leads to co-transcriptional silencing of the canonical gene promoter, thereby silencing expression of the protein-coding transcript (Chen et al., 2017; Chia et al., 2017; Van Dalfsen et al., 2018; Tresenrider et al., 2021; Vander Wende et al., 2022). Finally, by immunoblotting, we observed an increase in Swi4 protein abundance in the ΔLUTI mutant compared to wild type (Figure 5E and 5F), further indicating that SWI4^LUTIexpression downregulates Swi4 protein levels in meiosis.

SWI4^LUTI translation occurs within the upstream ORFs (uORFs) of its 5′ leader sequence (Brar et al., 2012). If the uORFs repress productive translation of the SWI4 CDS contained within the LUTI, then their removal should result in increased Swi4 protein levels. To determine whether the uORFs inhibit translation of the SWI4 CDS, we mutated the start codon of all seven uORFs within SWI4^LUTI from ATG to ATC (ΔuORF) and measured Swi4 protein levels using immunoblotting. Compared to wild type, the ΔuORF mutant had higher levels of Swi4 protein (Figure 5G and 5H). RNA blotting confirmed that SWI4^canon transcript levels remained similar between wild type and ΔuORF mutant, whereas SWI4^LUTIlevels were higher in the ΔuORF mutant compared to wild type (Figure 5G), likely resulting from increased transcript stability due to bypass of nonsense mediated decay (Tresenrider et al., 2021). Our findings indicate that the uORFs within the LUTI are translated at the expense of SWI4 CDS, thus halting Swi4 synthesis during early meiosis.

Our previous findings highlight an antagonistic relationship between Swi4 and Ime1 nuclear localization, whereby overexpression of SWI4 leads to a decrease in Ime1 nuclear localization (Figure 2E). Given the previous finding that Ime1 activates transcription of LUTIs in early meiosis (Tresenrider et al., 2021), we were curious whether the reverse regulation could also occur, where Ime1 causes downregulation of Swi4 through activating SWI4^LUTItranscription. To test this possibility, we used an inducible allele of IME1 (pCUP1-IME1) (Chia and van Werven, 2016) and measured Swi4 protein abundance. In the absence of IME1 induction, Swi4 levels remained constant. However, after 1 hour of IME1 induction, Swi4 levels started to decrease dramatically (Figure 5I and 5J). Therefore, reduction in Swi4 levels is dependent on Ime1 rather than being driven by nutrient deprivation in SPO, a condition that is known to trigger autophagy (Abeliovich and Klionsky, 2001). To further assess whether SWI4^LUTI is regulated by the Ime1-Ume6 transcription factor complex, we analyzed a published Ume6 ChIP-seq dataset (Tresenrider et al., 2021) and found evidence for Ume6 binding at the SWI4^LUTI promoter (Figure 5K). Additionally, analysis of an mRNA-seq dataset from UME6-T99N (Tresenrider et al., 2021), an allele of UME6 that can no longer interact with Ime1 revealed a dramatic reduction in SWI4^LUTIexpression in meiotic conditions (Figure S5). Together, these data support the notion that Ime1-Ume6 complex induces the expression of SWI4^LUTI, which in turn inhibits Swi4 protein synthesis through the combined act of transcriptional and translational interference.

Since the removal of SWI4^LUTI resulted in an increase in Swi4 levels, we next wanted to investigate how the loss of LUTI-based regulation affects meiotic entry. We performed mRNA-seq in ΔLUTI mutant or wild-type cells during meiotic entry (2 h in SPO) and used DESeq2 to identify differentially expressed genes. To our surprise, there was no obvious increase in the expression of many SBF targets (Figure 6A). Namely, CLN1 and CLN2 were both expressed at similar levels to wild type upon loss of SWI4^LUTI. Therefore, it appears that disruption of the LUTI-based regulation alone is not sufficient to reactivate SBF targets.

Since transition from the mitotic to meiotic program is a critical cell fate decision, it is likely that additional players are in place to restrict SBF activity. In this case, even when the LUTI-based regulation fails, SBF would remain largely inactive due to a backup mechanism. In support of this notion and consistent with a previous report (Argüello-Miranda et al., 2018), the SBF inhibitor Whi5 was expressed during early meiosis (0 h in SPO) and localized to the nucleus (Figure S6A). During the mitotic G1 phase, Whi5 associates with promoter-bound SBF, thereby preventing the transcription of SBF target genes (De Bruin et al., 2004; Costanzo et al., 2004). Shortly before the G1/S transition, phosphorylation of Whi5 via the Cln3/CDK pathway activates SBF by promoting nuclear export of Whi5 (De Bruin et al., 2004; Costanzo et al., 2004). However, in nutrient-deprived conditions that favor meiosis such as nitrogen limitation, CLN3 is translationally repressed and any Cln3 protein that is expressed is also unstable (Parviz and Heideman, 1998; Gallego et al., 1997). Without Cln3/CDK, Whi5 is expected to inhibit SBF in a constitutive manner.

To determine whether Whi5 and SWI4^LUTI act in parallel to restrict SBF activity, we removed Whi5 from the nucleus using the anchor-away method, which enables compartment-specific depletion of a target protein via inducible dimerization (Haruki et al., 2008). Whi5 was tagged with FRB and ribosomal protein Rpl13a was tagged with FKBP12. Upon addition of rapamycin, Rpl13a-FKBP12 and Whi5-FRB formed a heterodimer, leading to successful nuclear exclusion of Whi5 (Figure 6B). Using Whi5 anchor away (WHI5-AA) alone or in combination with removal of LUTI-based SWI4 regulation (ΔLUTI), we performed mRNA-seq to measure expression of SBF targets and early meiotic genes. Similar to the ΔLUTI single mutant, WHI5-AA alone did not significantly change the expression of SBF targets or early meiotic genes relative to wild type (Figure 6C). However, loss of both modes of regulation resulted in increased expression of SBF targets as well as a concomitant decrease in early meiotic transcripts compared to wild type (Figure 6D, Supplemental Table 3). As expected, Swi4 levels were similar to wild type in WHI5-AA but were elevated in ΔLUTI and ΔLUTI; WHI5-AA mutants (Figure S6B). Gene set enrichment analysis revealed a significant enrichment of the SBF regulon (NES = 1.95, p < 0.001) as well as significant disenrichment of the early meiotic genes (NES =-3.39, p < 0.001) in the ΔLUTI; WHI5-AA double mutant (Figure S6C).

Finally, we monitored meiotic entry using time-lapse fluorescence microscopy in strains carrying endogenously tagged Rec8-GFP in conjunction with Htb1-mCherry. In agreement with our mRNA-seq data, loss of either LUTI-or Whi5-based repression of SBF alone was not sufficient to cause a delay in meiotic entry (Figure 6E). However, simultaneous perturbation of both pathways led to a significant delay in meiotic entry (Figure 6E, p = 0.0112, Mann-Whitney test). These data further support the notion that SWI4^LUTI is integrated into a larger regulatory network to regulate SBF activity during meiotic entry, which includes Whi5-mediated repression of SBF.

It has long been known that entry into the mitotic cell cycle and meiotic differentiation are regulated differently. During the mitotic cell cycle, budding yeast cells must reach a critical size before commitment to division in late G1 phase, an event termed the “Start” (Hartwell et al., 1974; Johnston et al., 1977). Start depends on the activation of the G1/S regulon by SBF and MBF, which regulate ∼200 genes that function in polarized growth, macromolecular biosynthesis, DNA replication, and repair among other critical processes (reviewed in Jorgensen and Tyers, 2004). In contrast, entry into meiosis requires the transcriptional activator, Ime1, which is exclusively induced in diploid cells through the integration of several extrinsic cues including nitrogen levels, carbon source and extracellular pH (reviewed in van Werven and Amon, 2011). Ime1 initiates the first transcriptional wave of early meiotic genes required for DNA replication, recombination, and chromosome morphogenesis. Before our study, it was unclear whether and how SBF-MBF activity is regulated during the transition from mitosis to meiosis, despite the transcriptome-wide differences between these two developmental programs. We found that during meiotic entry, the SBF-specific Swi4 subunit is downregulated, which indicates that the activity of SBF, but not MBF, is constrained. Consistently, SBF-specific targets display either low or no expression during early meiosis, while MBF targets are upregulated upon meiotic entry. MBF targets are involved in processes shared between mitosis and meiosis including DNA replication and repair, whereas SBF targets are predominantly involved in processes unique to mitosis such as budding and cell wall biosynthesis. This functional specialization might have enabled adaptations to accommodate specific cell fates. Likewise, SBF-specific targets that are detrimental to meiosis, such as the G1 cyclins (Colomina et al., 1999, and this study), might contribute to the mutually exclusive nature of the mitotic and meiotic programs.

Absolute measurements of the concentrations of SBF and MBF in single cells during mitotic G1 demonstrated that these transcription factors are sub-saturating with respect to their target promoters in small cells, but that SBF and MBF levels increase as cells grow, suggesting that their abundance is a limiting factor in activating the G1/S regulon (Dorsey et al., 2018). Corroborating these findings, we show that overexpression of SWI4 as well as the simultaneous disruption of the two pathways that normally restrict SBF activity at meiotic entry – namely LUTI and Whi5 – result in abrupt activation of the SBF regulon and a concomitant decrease in early meiotic gene expression. However, MBF targets do not seem to be affected in response to increased SBF activity. This is surprising given that SBF and MBF share a common subunit, Swi6, and that an upsurge in Swi4 abundance might be expected to titrate Swi6 away from MBF. It is possible that meiosis-specific regulators like Ime1 and Ime2 compensate to maintain MBF target expression (Brush et al., 2012). Future studies are necessary to dissect the meiotic roles of MBF and potential compensatory mechanisms.

We found that the meiotic entry delay due to untimely SBF activation can be partially rescued by loss of either CLN1 or CLN2, demonstrating that both cyclins are responsible for the meiotic defects associated with SBF misregulation. In addition to CLN1 and CLN2, improper activation of SBF also leads to upregulation of PCL1 during meiotic entry (Figure 2F and 6B). PCL1 encodes a cyclin that interacts with the Pho85 CDK and is involved in the regulation of polarized cell growth and morphogenesis as well as progression through the cell cycle (Espinoza et al., 2016). Whether PCL1 misexpression contributes to the SBF-associated meiotic entry defects is an interesting avenue for future studies.

We and others have demonstrated that elevated levels of G1 cyclins inhibit meiotic entry. Therefore, it is crucial to dissect the mechanisms that govern the downregulation of G1 cyclins during this process. Among the G1 cyclins that are known to inhibit meiotic entry (Colomina et al., 1999), only the mechanism of CLN3 restriction was previously known (Parviz and Heideman, 1998; Gallego et al., 1997). Our study provides a mechanistic understanding of how CLN1 and CLN2 remain quiescent during meiotic entry through the combined act of LUTI-based and Whi5-dependent inhibition of SBF. Whether additional mechanisms play a role in restricting SBF and/or CLN1-CLN2 remains to be determined.

Our findings also shed light on the mechanism by which G1 cyclins prevent meiotic entry. Previous work demonstrated that G1 cyclin overexpression leads to downregulation of IME1 expression and inhibition of Ime1 nuclear localization (Colomina et al., 1999). Even though we observed similar defects in IME1 in response to CLN2 overexpression, increasing IME1 expression or targeting Ime1 to the nucleus did not result in successful meiosis (Figure 4). Instead, we found that restoring the interaction between Ime1 and its cofactor Ume6 was sufficient to bypass the meiotic blockage exerted by Cln/CDK. Collectively, our analyses suggest that the primary reason why G1 cyclins cause a meiotic failure is due to a deficiency in Ime1-Ume6 interaction. A previous study reported that G1 cyclins do not affect the interaction between Ime1 and Ume6 (Colomina et al., 1999). However, this conclusion was based on the use of a truncated Ime1 protein that only contained the Ume6 interaction domain. Therefore CLN-dependent regulations on full-length Ime1 might be missed in this context.

Future work could be aimed at dissecting how G1 cyclins affect the interaction between Ime1 and Ume6 and whether their impact on Ime1’s subcellular localization is primarily due to G1 cyclin-dependent changes in Ime1-Ume6 interaction. Since there are no CDK consensus phosphorylation sites on Ime1 itself, other players are likely to be involved. Rim11 and Rim15 are potential candidates since these two kinases have been implicated in Ime1-Ume6 phosphorylation as well as regulation of Ime1 localization and its interaction with Ume6 (Rubin-Bejerano et al., 1996; Vidan and Mitchell, 1997; Bowdish et al., 1994). Interestingly, Rim15 contains CDK consensus phosphorylation sites (Holt et al., 2009; Moreno-Torres et al., 2017; Breitkreutz et al., 2010), Furthermore, Cln2/CDK activity has been previously shown to inhibit Rim15 nuclear localization (Talarek et al., 2010), thereby making it an attractive candidate for further investigation.

Swi4 and Ime1 regulate each other in an antagonistic manner (Figure 6F), as supported by several observations. First, overexpression of SWI4 leads to a reduction in Ime1 nuclear localization (Figure 2B). Second, premature activation of SBF through SWI4 overexpression or simultaneous disruption of SWI4^LUTIand Whi5 pathways results in a significant delay in the chromosomal localization of Rec8, a direct transcriptional target of Ime1 (Figure 2H, 2I and 6E). Additionally, many early meiotic genes that are also transcriptional targets of Ime1 are downregulated under the same conditions (Figure 2F and 6D). Third, IME1 is inhibited by the G1 cyclins, which are themselves targets of Swi4/SBF (Figure 4). The reverse regulation, where Ime1 inhibits Swi4, is also in place. This occurs through the Ime1-dependent expression of SWI4 ^LUTI, which is necessary to downregulate Swi4 levels (Figure 5, also reported in Tresenrider et al., 2021). The antagonistic relationship between Swi4 and Ime1 is further evidenced by their mutually exclusive pattern of nuclear localization at the single-cell level (Figure 2E). In summary, our study highlights the multiple ways in which Swi4 and Ime1 regulate each other, which likely plays a crucial role in cellular decision making between the mitotic and meiotic transcriptional programs.

While LUTI-based regulation is both necessary and sufficient to downregulate Swi4 levels, disrupting it alone is not enough to activate the SBF regulon. We found that the LUTI-based mechanism works together with the Whi5 pathway to inhibit SBF activity (Figure 6). As a result, only when both mechanisms are simultaneously disrupted, do cells exhibit abrupt activation of SBF targets and subsequent meiotic entry defects. This two-pronged inhibition of SBF activity is reminiscent of how meiotic cells prevent microtubule-kinetochore interactions during prophase I (Miller et al., 2012). Specifically, LUTI-based regulation represses the expression of a limiting outer kinetochore subunit, Ndc80 (Chen et al., 2017). Even when the outer kinetochore assembles upon disruption of NDC80^LUTI, microtubules are still unable to engage with the kinetochores due to restriction of Clb/CDK activity. The loss of both regulations results in meiotic chromosome segregation defects (Chen et al., 2017). Integrating LUTI-based repression into larger regulatory networks likely ensures robustness in cellular decision-making, providing a fail-safe system. However, combinatorial use of the LUTI-based mechanism with other regulatory pathways poses a challenge to uncover their functional importance. Despite 8% of the yeast genes being subject to LUTI-based regulation during meiosis (Cheng et al., 2018), the biological significance of only two LUTIs, NDC80^LUTI and SWI4^LUTI, have been uncovered thus far. Therefore, systematic studies aiming to dissect the biological roles of LUTIs in meiosis and beyond would benefit from simultaneous perturbation approaches, including synthetic genetic interactions.

Cell state changes are primarily governed by transcription factors. Our study highlights how transition from mitotic to meiotic cell fate is ensured by inhibition of the SBF transcription factor through two distinct mechanisms. The SBF regulon includes G1 cyclins, whose untimely expression blocks meiotic entry by disrupting the interaction between Ime1 and Ume6, a key transcriptional co-activator of meiosis. SBF and Ime1-Ume6 antagonize each other, thereby helping establish a mutually exclusive state between the mitotic and meiotic transcriptional programs. Mammalian cells have functional homologs to both sets of transcription factors. First, similar to Ume6, MEIOSIN binds to the promoters of early meiotic genes and recruits Ime1-like STRA8 to activate gene expression (Anderson et al., 2008; Kojima et al., 2019; Baltus et al., 2006; Ishiguro et al., 2020). Second, SBF and MBF counterparts in mammalian cells are known as E2Fs, nine of which control the G1/S transition (reviewed in Van Den Heuvel and Dyson, 2008). Each E2F regulates a distinct sets of genes (Gaubatz et al., 2000). Furthermore, some members of the E2F family are involved in tissue-specific regulation of cell fate (Julian et al., 2016). However, the molecular mechanisms driving the rewiring of these complex gene networks are not well understood. Our investigation could help shed light on the regulatory mechanisms governing the interplay between mitotic and meiotic transcription factors outside of yeast.

All strains used in this study were derived from the SK1 background. Detailed information about the strain genotypes can be found in Supplemental Table 4.

For SWI4 (pUB1587) 1200 bases upstream of the ORF, SWI4 ORF, C-terminal 3V5 epitope tag, and SWI4 3’UTR (1000 bases downstream of ORF) were cloned into a LEU2 single integration vector by Gibson assembly (NEB) (Gibson et al., 2009). LUTI promoter deletion (ΔLUTI) strain (pUB1588) was similarly constructed with-1200 to-934 bases upstream from ORF deleted. For the uORF mutant (ΔuORF) strain (pUB1734) all seven ATG uORFs were mutated using gBlocks with the ATG>ATC mutations and cloned by Gibson assembly into 3V5-tagged LEU2 single integration vector. In all strains described above, the endogenous SWI4 gene is deleted using Pringle-based insertion of KanMX6 marker (Bähler et al., 1998).

For overexpression of SWI4 (pUB1585), CLN1 (pUB2144), and CLN2 (pUB1899), the relevant gene carrying a 3V5 epitope tag was cloned downstream of the ATG8 promoter, which is highly expressed during meiosis (Brar et al., 2012). A LEU2 single integration vector was used for cloning the fragments by Gibson assembly (NEB). In the strains carrying overexpression transgenes, the wild-type alleles at the endogenous loci remained intact. pCUP1-GFP-IME1 allele was made with Pringle-based insertion of the pCUP1 promoter upstream of GFP-IME1. PUS1-⍺GFP and UME6-⍺GFP were made by Pringle-based insertion using a plasmid (pUB1707) gifted by the Lackner Lab. For WHI5-AA the WHI5-mCherry-FRB allele was made by Pringle-based insertion using a plasmid (pUB595) gifted by the Nasmyth lab. All plasmid sequences were confirmed by sequencing.

All single integration plasmids were digested with Pme1 before transformation. Proper integration was confirmed by PCR. The plasmids constructed in this study are listed in Supplemental Table 5.

For meiotic experiments using pCUP1-IME1/pCUP1-IME4 or pCUP-IME1 system, cells were synchronized as described in (Chia and van Werven, 2016). Briefly, after 24 h of growth in YPD at room temperature, saturated cultures (OD₆₀₀ > 10) were diluted to an OD₆₀₀ of 0.25 in BYTA (1% yeast extract, 2% bacto tryptone, 1% potassium acetate, and 1.02% potassium phthalate) for 16-18 h of growth at 30°C (OD₆₀₀ of > 5). Cells were washed with water twice before final resuspension in SPO with 0.5% potassium acetate to an OD₆₀₀ of 1.85. After 2 h in SPO, IME1 and IME4 were induced with 50 µM CuSO₄. Sporulation efficiency was measured after 24 h in SPO. Anchor away meiotic experiments was performed as described above with final 1 µM rapamycin (Millipore) added to BYTA 30 minutes before transfer to SPO and again 1 µM rapamycin added to SPO media.

In all other meiotic experiments, cells were prepared as in (Carlile and Amon, 2008). Briefly, after 24 h of growth in YPD at room temperature, saturated cultures (OD₆₀₀ > 10) were diluted to an OD₆₀₀ of 0.25 and inoculated in BYTA (1% yeast extract, 2% bacto tryptone, 1% potassium acetate, and 1.02% potassium phthalate) for 16-18 h at 30°C (OD₆₀₀ of > 5). Cells were washed with water twice before final resuspension in SPO with 0.5% potassium acetate to an OD₆₀₀ of 1.85. Sporulation efficiency was counted after 24 h in SPO.

UME6-⍺GFP meiotic experiments were prepared as in (Chia and van Werven, 2016). Briefly, after 18 h of growth in YPD at room temperature, saturated cultures (OD₆₀₀ > 10) were diluted to an OD₆₀₀ of 0.2 in reduced YPD (1% yeast extract, 2% peptone, 2% dextrose, uracil [24 mg/L]). Reduced YPD was used instead of BYTA to prevent cells from prematurely entering meiosis in BYTA due to the Ime1-Ume6 interaction from the GFP nanobody. Cells were grown for ∼6 h at 30°C until they reached an OD₆₀₀ between 0.5 and 1.0. Cultures were then back diluted to OD₆₀₀ of 0.1 and grown for 18 h at 30°C. Cells were washed with water twice before final resuspension in SPO with 0.5% potassium acetate to an OD₆₀₀ of 1.85. Sporulation efficiency was counted after 24 h in SPO.

RNA extraction was performed as described in (Tresenrider et al., 2021). Briefly, ∼4 OD unit of cells were pelleted by centrifugation for 1 minute at 20,000 rcf and snap frozen in liquid nitrogen. Cells were thawed on ice and resuspended in TES buffer (10 mM Tris pH 7.5, 10 mM EDTA, 0.5% SDS). An equal volume of Acid Phenol:Chloroform:Isoamyl alcohol (125:24:1; pH 4.7) was added to cells and incubated at 65°C for 45 min shaking at 1400 rpm. Aqueous phase was transferred to a new tube with chloroform, vortexed for 30 sec, separated by centrifugation, and precipitated in isopropanol and sodium acetate overnight at-20°C. Pellets were washed with 80% ethanol and resuspended in DEPC water for 10 min at 37°C. Total RNA was quantified using a nanodrop.

RNA-seq libraries were generated with the NEXTflex^TM Rapid Directional mRNA-Seq Kit (NOVA-5138, Perkin Elmer). 10 µg of total RNA was used as input for all libraries. AMPure XP beads (A63881, Beckman Coulter) were used to select fragments between 200-500 bp. Libraries were quantified using the Agilent 4200 TapeStation (Agilent Technologies, Inc). Samples were submitted for 100 bp SE sequencing by the Vincent J. Coates Genomics Sequencing Laboratory with a NovaSeq SP 100SR.

Hisat2 (Kim et al., 2019) was used to align reads to map sequences to SK1 PacBio genome. Quantification of RNA as transcripts per million was done StringTie (Pertea et al., 2015). Fold-change quantification was performed by DESeq2 using default options (version 1.34.0, (Love et al., 2014)).

For Gene Set Enrichment Analysis (GSEA) v4.3.2 was used to compare TPM values for different gene sets. The early meiotic gene set was created from Figure 2 of (Brar et al., 2012). The SBF regulon was from Figure 3 of (Iyer et al., 2001). GSEA was performed on the desktop app with default settings expect “Collapse/Remap to gene symbols” was set to “No_Collapse” and “Permutation type” was set to “gene_set”.

For Gene Ontology (GO) Analysis, SGD Gene Ontology Slim Term Mapper was used for GO analysis using Yeast Go-Slim: process GO Set. Used output from DE-Seq2 analysis with a cutoff of padj (P value) < 0.05.

The two overlapping targets (SWE1 and TOS4) between the early meiotic gene set and SBF targets are not plotted on volcano plot or used for GSEA analysis.

5 μg of isolated total RNA was treated with DNase (TURBO DNA-free^TM Kit). cDNA was reverse transcribed following the Superscript III kit (ThermoFisher Scientific). Quantification was performed with Absolute Blue qPCR Mix (ThermoFisher Scientific). Meiotic samples were normalized to PFY1, and mitotic samples were normalized to ACT1. Oligonucleotides are listed in Supplemental Table 6.

∼4 OD₆₀₀ of cells were collected and pelleted by centrifugation for 1 min at 20,000 rcf. Pellet was resuspended in 5% (w/v) TCA for at least 15 min at 4°C. Cells were washed with TE50 (50 mM Tris-HCl [pH 7.5], 1 mM EDTA) and then with 100% acetone. The cell pellet was dried overnight and then lysed with glass beads in lysis buffer (Tris-HCl [pH 7.5], 1 mM EDTA, 2.75 mM DTT, protease inhibitor cocktail (cOmplete EDTA-free [Roche]). Next 3x SDS sample buffer (187.5 mM Tris [pH 6.8], 6% β-mercaptoethanol, 30% glycerol, 9% SDS, 0.05% bromophenol blue) was added and the cell lysate was boiled for 5 min at 95°C. Protein was separated by PAGE using 4-12% Bis-Tris Bolt gels (Thermo Fisher) and transferred onto 0.45 μm nitrocellulose membranes.

Cln1-3V5, Cln2-3V5 and GFP-GFP tagged proteins were all transferred onto 0.45 μm nitrocellulose membranes using a semi-dry transfer apparatus (Trans-Blot Turbo System (Bio-rad)). Swi4-3V5 or untagged Swi4, Swi6, and Mbp1 proteins were all transferred onto 0.45 μm nitrocellulose membranes using a PROTEAN Tetra tank (BioRAD) filled with 25 mM Tris, 192 mM glycine, and 7.5% methanol. All blots were incubated at room temperature with Odyssey Blocking Buffer (PBS) (LI-COR Biosciences).

Immunoblotting for Cln1-3V5, Cln2-3V5, and GFP-Ime1 was performed as previously described in (Tresenrider et al., 2021). Briefly, mouse α-V5 antibody (R960-25, Thermo Fisher) or mouse α-GFP antibody (632381, Takara) were diluted 1:2000 in Odyssey Blocking Buffer (PBS) (LI-COR Biosciences) with 0.01% Tween. Rabbit α-hexokinase (α-Hxk2) antibody (H2035, US Biological) was diluted to 1:20,000. Secondary antibodies used were α-mouse antibody conjugated to IRDye 800CW (926-32212, LI-COR Biosciences) and α-rabbit antibody conjugated to IRDye 680RD (926-68071, LI-COR Biosciences). Secondary antibodies were diluted to 1:20,000 in Odyssey Blocking Buffer (PBS) with 0.01% Tween.

For immunoblotting of Swi4, Swi6, and Mbp1. antibodies specific to each subunit were a generous gift from the Andrews and deBruin labs (Andrews and Herskowitz, 1989; Harris et al., 2013). α-Swi4, α-Swi6, α-Mbp1 antibodies were each diluted to 1:2000 in Odyssey Blocking Buffer (PBS) (LI-COR Biosciences) with 0.01% Tween. Secondary antibodies included a α-rabbit antibody conjugated to IRDye 800CW (926-32213, LI-COR Biosciences) and a α-rabbit antibody conjugated to IRDye 680RD (926-68071, LI-COR Biosciences).

Odyssey system (LI-COR Biosciences) was used to image the blots, and Image Studio Lite (LI-COR Biosciences) was used for image quantification.

For each blot, 10 µg of total RNA was dried in a Savant Speed Vac (SPD111V). RNA was then resuspended and denatured in glyoxal/DMSO mix (1M deionized glyoxal, 50% v/v DMSO, 10 mM sodium phosphate (NaPi) buffer [pH 6.8]) at 70°C for 10 min. RNA sample was loaded into an agarose gel (1.1% [w/v] agarose in 0.01 M NaPi buffer) with loading dye (10% v/v glycerol, 2 mM NaPi buffer [pH 6.8], ∼0.25% w/v xylene cyanol, and orange G) and ran for 3 h at 100 V with a Variable Speed Pump (BioRad) to circulate buffer during the entire gel run.

RNA was transferred overnight to nylon membrane (Hybond-N+ [GE]) in SSC. Membrane was crosslinked using a Stratalinker UV Crosslinker (Stratagene). Ribosomal RNA (rRNA) bands were visualized with methylene blue staining and imaged on a Gel Doc XR+ Molecular Imager with Image Lab software (BioRad). Probe templates containing the T7 promoter were amplified using PCR. PCR product was concentrated with MinElute Spin Columns (Qiagen) and then used for in vitro transcription to generate a strand-specific RNA probe using a MaxiScript T7 Kit (Invitrogen) according to the manufacturer’s instructions, except cold UTP was replaced with α-P32 labeled UTP (PerkinElmer). Excess nucleotides were removed with NucAway Spin Columns (Invitrogen). Blots were blocked in ULTRAhyb Ultrasensitive Hybridization Buffer (Invitrogen) for 1 h and then incubated with the α-P32 labeled probe overnight at 68°C. Blots were then washed twice with low stringency wash buffer (2X SSC, 0.1% SDS) for 10 min and then washed twice with high stringency wash buffer (0.1X SSC, 0.1% SDS) for 15 min. Blots were then exposed overnight on a storage phosphor screen (Molecular Dynamics) and then imaged on a Typhoon phosphor-imaging system.

For time course imaging of cells expressing GFP-IME1, SWI4-mCherry, or WHI5-mCherry, 500 µL of meiotic culture was fixed with a final concentration of 3.7% formaldehyde (v/v) at room temperature for 15 min. Cells were then washed in 1 ml of 100 mM potassium phosphate [pH 6.4] and stored at 4°C in 20 µl of KPi Sorbitol solution overnight (100 mM potassium phosphate [pH 7.5], 1.2M sorbitol). Cells were mounted on a slide and imaged using DeltaVision Elite wide-field fluorescence microscope (GE Healthcare) with a 60x/1.516 oil immersion objective. Deconvolution of images was done with softWoRx imaging software (GE Life Sciences).

For live-cell imaging, cells at OD₆₀₀ of 1.85 in conditioned SPO (filter-sterilized SPO culture after ∼5 h in 30°C) were sonicated and transferred to a concanavalin A (Sigma) treated 96-well clear, flat bottom plate (Corning). Four z positions (2 μm step size) were acquired per XY position. Acquisition was performed in a temperature-controlled chamber at 30°C. Please refer to Supplemental Table 7 for acquisition settings.

All image analysis was performed with FIJI (Schindelin et al., 2012). Maximum z-projection are shown in figures and were modified using linear brightness and contrast adjustments in FIJI.

To quantify localization of GFP-Ime1 and Rec8-GFP, z-slices containing the nucleus were selected using the Htb1-mCherry signal. Max projection was created from these slices and GFP-Ime1 was scored double-blinded as nuclear or not nuclear.

For mean nuclear intensity of GFP-Ime1 and Swi4-mCherry, an individual z-slice containing the nucleus was selected and nuclear mask was generated using the Htb1-mCherry signal. The nuclear mask was then used to quantify the mean nuclear intensity of GFP-Ime1 signal.

smFISH was performed and quantified as previously described in (Chen et al., 2018). All probes were ordered from (Biosearch Technologies). Unique region of SWI4^LUTIwas visualized by twenty-eight 20-mer oligonucleotide probes coupled to CAL fluor Red 590. Thirty-eight 20-mer probes coupled to Quasar 670 dye targeted to SWI4^canon. ∼ 4 OD unit of cells were fixed in final 3% formaldehyde (v/v) and incubated at room temperature for 20 minutes. Fixed samples were moved to 4°C to continue fixing overnight. Cells were washed three times in cold Buffer B (0.1 M potassium phosphate [pH 7.5], 1.2 M sorbitol) and resuspended in digestion buffer (Buffer B, 200 mM Vanadyl ribonucleoside complex [VRC from NEB], zymolyse [zymolase 100T, MP Biomedicals]). Cells were digested at 30°C for 20 min and then gently washed with 1 mL of cold Buffer B and resuspended in 1 mL of 70% ethanol for 3.5-5 h.

Cells were then incubated in 1mL of 10% formamide wash buffer (10% formamide, 2X SSC) at room temperature for 15 minutes. For hybridization, each probe set was added (final concentration of 500 nM) to 20mM VRC and hybridization buffer (1% Dextran sulfate [EMD Millipore], 1 mg/mL E. coli tRNA [Sigma], 2 mM VRC, 0.2 mg/mL BSA, 1X SSC, 10% formamide in nuclease-free water). Hybridization was done overnight at 30°C. Samples were incubated in the dark for 30 min at 30°C in 1 mL of 10% formamide wash buffer. Buffer was then removed, and cells were stained with DAPI and resuspended in glucose-oxygen-scavenging buffer or GLOX buffer (10 mM Tris [pH 8.0], 2x SSC, 0.4% glucose) without enzymes. Before imaging, GLOX solution with enzyme (1% v/v catalase, 1% v/v glucose oxidase (Sigma), 2 mM Trolox (Sigma)) was added to sample.

Images were acquired with the DeltaVision microscope as described in previous section with filters: TRITC (EX542/27, EM597/45) for CAL Fluor Red 590 and CY5 (EX632/22, EM679/34) for Quasar 670. Series of z-stacks (15–25 slices) were acquired with a step size of 0.2 μm.

Matlab script (Chen et al., 2017, 2018) was run to quantify FISH spots with max intensity projection of z-stacks. The same “signal” and “SNR” thresholds were applied to all the images within a replicate.

All materials used in this study are available upon request from the corresponding author. Sequencing data generated in this study are available at NCBI GEO under the accession ID: GSE225963

The custom code used for the analysis is available in the following code repository: https://github.com/elifesciences-publications/Chen_Tresenrider_et_al_2017