Transcription-dependent enrichment of the yeast FACT complex influences nucleosome dynamics on the RNA polymerase III-transcribed genes.

The FACT (FAcilitates Chromatin Transactions) complex influences transcription initiation and enables passage of RNA polymerase (pol) II through gene body nucleosomes during elongation. In the budding yeast, ~280 non-coding RNA genes highly transcribed in vivo by pol III are found in the nucleosome-free regions bordered by positioned nucleosomes. The downstream nucleosome dynamics was found to regulate transcription via controlling the gene terminator accessibility and hence, terminator-dependent pol III recycling. As opposed to the enrichment at the 5'-ends of pol II-transcribed genes, our genome-wide mapping found transcription-dependent enrichment of the FACT subunit Spt16 near the 3'-end of all pol III-transcribed genes. Spt16 physically associates with the pol III transcription complex and shows gene-specific occupancy levels on the individual genes. On the non-tRNA pol III-transcribed genes, Spt16 facilitates transcription by reducing the nucleosome occupany on the gene body. On the tRNA genes, it maintains the position of the nucleosome at the 3' gene-end and affects transcription in gene-specific manner. Under nutritional stress, Spt16 enrichment is abolished in the gene downstream region of all pol III-transcribed genes and reciprocally changed on the induced or repressed pol II-transcribed ESR genes. Under the heat and replicative stress, its occupancy on the pol III-transcribed genes increases significantly. Our results show that Spt16 elicits a differential, gene-specific and stress-responsive dynamics, which provides a novel stress-sensor mechanism of regulating transcription against external stress. By primarily influencing the nucleosomal organization, FACT links the downstream nucleosome dynamics to transcription and environmental stress on the pol III-transcribed genes. Published by Cold Spring Harbor Laboratory Press for the RNA Society.

INTRODUCTION

For a typical yeast gene transcribed by RNA polymerase (Pol) II, the 5′-end is characterized by a nucleosome-free region (NFR) immediate upstream of the transcription start site (TSS), flanked by two well-positioned nucleosomes (−1 and +1) (Lee et al. 2007; Mavrich et al. 2008; Rando and Chang 2009). In the budding yeast, a similar arrangement of nucleosomes is found on all the genes transcribed by Pol III, which reside in an NFR actively maintained by the chromatin remodeling complexes RSC and ISW (Parnell et al. 2008; Kumar and Bhargava 2013). The +1 nucleosome on the Pol II-transcribed genes shows a strong positioning (Jansen and Verstrepen 2011), whereas on the Pol III-transcribed genes, the upstream (US, −1) nucleosome shows a strong positioning between −70 and −220 bp (Kumar and Bhargava 2013). In contrast, the +1 (DS) nucleosome is positioned at various distances from the transcription termination site (TTS/terminator) of different Pol III-transcribed genes (Kumar and Bhargava 2013).

The transcription complex (TC) of Pol III consists of two general factors TFIIIC and TFIIIB apart from Pol III (Geiduschek and Kassavetis 2001). Pol III transcribes short, noncoding RNA genes, of which tRNAs form the bulk of its transcriptome. Depending on the promoter elements (typically intragenic), they are grouped into three different classes, represented by 5S rRNA (class I), tRNAs (class II), and U6 snRNA (class III) (Dieci et al. 2013a). The short length of their transcribed regions (generally ∼100 bp) allows physical transfer of Pol III from the terminator to the 5′ gene end at the end of each transcription cycle (Dieci and Sentenac 2003; Dieci et al. 2013b). This terminator-dependent facilitated recycling of Pol III (Dieci and Sentenac 1996; Leśniewska and Boguta 2017) could be one of the reasons for high transcription rate of these genes in vivo. Transcription by Pol III is repressed under a variety of stress conditions, including nutrient starvation (Clarke et al. 1996; Willis and Moir 2007; Moir and Willis 2013). In the budding yeast, ∼280 genes transcribed by Pol III are found scattered on different chromosomes, each with its own chromatin context (Hani and Feldmann 1998). There is increasing evidence in support of their chromatin-mediated and gene-specific regulation under different conditions (Bhargava 2013; Shukla and Bhargava 2018).

FACT complex (a heterodimer of Spt16–Pob3 in yeast) has been reported to show genetic interaction with the Pol II elongation factors, physically associate and travel with Pol II during transcription elongation (Squazzo et al. 2002; Belotserkovskaya et al. 2003; Mason and Struhl 2003; Saunders et al. 2003; Schwabish and Struhl 2004) and influence transcription by Pol I and Pol II (Formosa et al. 2002; Mason and Struhl 2003; Biswas et al. 2005; Birch et al. 2009). FACT has been reported to influence transcription by Pol II at different steps via different mechanisms. As a chaperone for dimers of the histone H2B with either H2A or its variants (Belotserkovskaya et al. 2003; Mahapatra et al. 2011; Formosa 2012; Jeronimo et al. 2015), FACT participates in histone exchange by removing/reinstating the H2A–H2B dimer from a nucleosome with the passage of the Pol II during elongation (Schwabish and Struhl 2004; Jamai et al. 2009). Yeast FACT is also reported to restore the histone density on the bodies of the genes, which is reduced during gene induction or transcription by Pol II (Schwabish and Struhl 2004; Jamai et al. 2009; Voth et al. 2014).

We had earlier reported a histone H2A.Z chaperone activity of FACT (Mahapatra et al. 2011) on the yeast Pol III-transcribed genes. Later, we found the FACT complex in our mass spectrometric data on the interacting partners of the yeast Pol III TC (Bhalla et al. 2019a). Barring a few genes (Birch et al. 2009; Mahapatra et al. 2011), the global presence of FACT on specifically the Pol III-transcribed genes has not been reported till recently (Martin et al. 2018). Our genome-wide mapping found that the FACT subunit Spt16 shows high enrichment at the 3′-ends of all the Pol III-transcribed genes from where it is lost under nutritional stress condition. Spt16 coimmunoprecipitates with the Pol III TC and affects Pol III transcription in a gene-specific manner. Its primary function on Pol III-transcribed genes is seen in maintaining the chromatin organization, whereby it depletes the gene-body nucleosomes and facilitates the transcription of the non-tRNA genes. On the tRNA genes, transcription elongation by Pol III leads to Spt16 enrichment in the DS region, which helps maintain the DS nucleosome position. Spt16 dynamics on Pol III-transcribed genes under starvation, heat and replication stress and on Pol II-transcribed ESR genes under nutritional stress indicates a role for it in the global, stress-responsive/sensing mechanism for cells under adverse growth conditions.

RESULTS

Spt16 interacts with the Pol III transcription complex

In our earlier attempts to capture the proteins interacting with the Pol III TC, we found the FACT complex as one of the interacting partners (Supplemental Table S1; Bhalla et al. 2019a) under both normal (active) and starvation (repressed) conditions. Yeast Nhp6 helps loading of the Spt16–Pob3 heterodimer of the FACT complex onto the nucleosome (Formosa et al. 2001). Out of the three baits, TFIIIC IP had the highest specific spectral counts (SSCs) of Nhp6, while TFIIIB had highest Pob3 SSCs (Supplemental Table S1), suggesting TFIIIC/TFIIIB may be involved in recruiting the FACT complex on the Pol III-transcribed genes. In coimmunoprecipitation (co-IP) experiments, Spt16 was efficiently pulled down by TFIIIB, TFIIIC, and Pol III in both active and repressed states (Fig. 1A–F). On the contrary, comparatively lower pull-down of the three in the reverse co-IPs (pull downs by the Spt16-HA; Fig. 1A–F) may be related to a smaller fraction of the total Spt16 pool available for them due to its distribution between Pol I-, Pol II-, and Pol III-TCs in vivo. While the co-IP of all three RNA polymerases with Spt16 bait has been reported earlier from HeLa cells (Birch et al. 2009), we did not find Pol III TC components among the pulled-down interactors of yeast Spt16–TAP used as bait in an earlier AP-MS data (Bedard et al. 2016). We validated the Pol III–Spt16 physical interaction further by in vitro pull-down experiments, using purified Pol III as bait. The Flag-tagged pure Pol III could pull down Spt16 specifically, from the crude lysate of cells carrying Myc-tagged Spt16 (lane 3, Fig. 1G), further confirming their physical interaction.

FIGURE 1.

Spt16 interacts with the Pol III TC. Immunoprecipitations (IPs) were performed from the extracts of the doubly tagged strains harboring tagged Brf1 or Tfc1 or Rpc160 paired with Spt16, grown under active (A) or repression (R) conditions. Samples were prepared, western transferred and analyzed as given in the Materials and Methods section. Representative results of both forward and reverse immunoblots (repeated with a minimum of three biologocal replicates) are shown in panels A,C,E. Panels B,D,F show quantification results from multiple western blots similar to the representative blots for co-IPs. Protein A/G beads without antibody were used in mock immunoprecipitations (M). Whole cell extract (Input) was used as a positive control. Subpanels are marked with antibody used for probing, and pull-down baits are indicated at the top of the panel. For ease of identification, IP and co-IP bands are marked in a lane. (A,B) Transcription factor TFIIIB subunit Brf1 and FACT subunit Spt16 coimmunoprecipitate. (C,D) Transcription factor TFIIIC subunit Tfc1 and FACT subunit Spt16 efficiently coimmunoprecipitate each other. (E,F) Reciprocal immunoprecipitations (IPs) of Rpc160-Myc (lanes 5,6) and Spt16-HA (lanes 7,8) show that they coimmunoprecipitate in both the conditions. Input (lanes 1,2) shows the whole cell extract, mock (lanes 3,4) shows IP without antibody addition. The blot was sequentially probed with bait (IP) and prey (co-IP) antibodies. Bands representing IP/co-IP are marked. (G) Spt16 pull-down by pure Flag-tagged Pol III from the crude lysate of cells carrying Myc-tagged Spt16. Input (lane 1) shows DNase I treated whole cell lysate. Mock (lane 2) shows Flag M2 agarose beads incubated with the cell lysate. IP (lane 3) shows pull-down of Spt16–Myc in the cell lysate by the Flag-tagged Pol III immobilized on Flag M2 agarose beads.

A comparison of active and repressed state IPs of the TC showed higher Spt16–Pob3 SSCs for Pol III and TFIIIB than for TFIIIC (Supplemental Fig. S1). Pol III is lost while TFIIIC and TFIIIB stay on the genes in the repressed state (Roberts et al. 2003). Therefore, increased pull down of Rpc160 by Spt16 (Fig. 1F) in the repressed state co-IP suggests Pol III–Spt16 interact even off the DNA in vivo. In contrast, the highest Spt16–Pob3 SSCs with TFIIIB as bait (Supplemental Table S1) but reduced pull down of TFIIIC and TFIIIB by Spt16 in the reverse co-IPs (Fig. 1B,D) suggest that under the repressed state, TFIIIB and TFIIIC do not associate with Spt16 stably off the DNA but may hold it on the gene under the repressed state. This implies a dynamic behavior of Spt16 on the gene body as discussed later.

Spt16 is enriched at the 3′ end of Pol III-transcribed genes

Our genome-wide ChIP-seq mapping of Spt16 gave data of high quality and reproducibility between replicates (Supplemental File S1; Supplemental Fig. S1A–C). The data showed good correlation with several previously published Spt16 occupancy data on the Pol II-transcribed ORFs (Supplemental Fig. S2A). The highest Spt16 levels on the Pol II-transcribed genes in our data were found on the +1 nucleosome (Fig. 2A), with the average Spt16 profile similar to the distribution reported earlier (Supplemental File S1; Mayer et al. 2010; True et al. 2016; Martin et al. 2018). A similar level occupancy of Spt16 at the +40 bp position (marked by an asterisk) on the tRNA genes (Fig. 2B) in the NFR (Kumar and Bhargava 2013) is found overlapping the ∼150 bp bootprint of TFIIIB and TFIIIC (Nagarajavel et al. 2013). However, Spt16 and Pol III occupancy levels (peaks marked by asterisks, Fig. 2B) on the tRNA gene body show only a moderately positive correlation (r = 0.45).

FIGURE 2.

Spt16 is enriched at the DS region of Pol III-transcribed genes. Spt16 ChIP-seq data were acquired and analyzed as detailed in Materials and Methods. Occupancies are plotted as normalized tag counts per million reads, all aligned to the TSS. (A) Spt16 occupancy profile on the Pol II-transcribed genes found in this study agrees with earlier reports (True et al. 2016; Martin et al. 2018) showing a distribution that follows the nucleosomal profile (MNase-seq data from Weiner et al. 2010). (B) Average profiles of Spt16 (this study) and Pol III (Kumar and Bhargava 2013) occupancies are compared on the tRNA genes. An asterisk (color coded) marks the Pol III and Spt16 peaks at the tRNA gene body where Spt16 occupancy is similar to that at the 5′-ends of the Pol II-transcribed genes. (C) Relative Spt16 occupancy in the unique regions flanking some of the tRNA genes. Spt16–TAP occupancy was measured by performing ChIP and the real time PCR, using TELVIR as normalizer. All genes show very low US levels. The P-values are (*) 0.0093, (**) <0.00022, and (***) <0.00008. (D) Spt16 occupancy on the non-tRNA Pol III-transcribed RPR1 and SCR1 genes. Occupancy is given as the normalized, tag counts per million reads in a window of −800 bp to +1000 bp with respect to the gene TSS. The color-coded asterisks mark the US and DS peaks of Spt16 on both the genes. (E) Schematic representation of the positions of the amplicons used in the real time PCR estimations on three non-tRNA genes. Bent arrow depicts position of the TSS. Gray boxes mark the mature transcript position. Short bars represent the amplicons. Names of the amplicons at the two ends of each gene are given below the short bars. (F) Spt16 occupancy at the 5′- and 3′-ends of the non-tRNA genes as measured by the ChIP-real time PCR method, using TELVIR as normalizer. (*) P = 0.006; (**) P < 0.0001.

The most striking feature of the Spt16 ocupancy profile was its highest levels in the DS nucleosome region at the tDNA loci (Fig. 2B; Supplemental Fig. S2B); approximately twofold higher than the levels on the tRNA gene body or ORFs (cf. Fig. 2A,B). Though Pol II is found near many tRNA genes (Supplemental Fig. S2C; Bhalla et al. 2019b); the average Spt16 profile was not influenced by the presence of Pol II-transcribed genes within 150 or 300 bp of both the gene ends (Supplemental Fig. S2D). Being outside the tRNA gene body, the highest association of Spt16 in the DS region of the Pol III-transcribed genes is specific and not because of the earlier reported hyper-ChIPability of a set of the tRNA genes (Teytelman et al. 2013). This is corroborated by the largely overlapping Spt16 enrichment profiles on the tRNA genes classified as hyper-ChIPable or otherwise (Supplemental File S1; Supplemental Fig. S2E).

We used the ChIP and real time PCR method for measuring the Spt16 occupancies. Our measurements of Spt16 occupancies on different tRNA genes (Fig. 2C) found different Spt16 levels on each of them and validated the higher enrichment in the DS than the US region. Since the US Spt16 levels are found close to the background (Fig. 2C), Spt16 is most likely not found in the US region (Supplemental File S1; Supplemental Fig. S1E,F). Spt16 levels on the Pol III-transcribed non-tRNA genes were also found higher at the 3′-ends than at the 5′-ends (marked by the asterisks) in our ChIP-seq data (Fig. 2D; Supplemental File S1). ChIP and real time PCR measurements (Fig. 2E,F) validated these level differences.

Spt16 influences Pol III transcription in a gene-specific manner

Similar to earlier reports (Mayer et al. 2010; Liu et al. 2014; Chang et al. 2018; Pathak et al. 2018), the highest enrichment of Spt16 on the Pol II-transcribed genes in our data is seen on those transcribed at highest frequency (Supplemental Fig. S3A). We used the available data on nascent transcript levels of the 274 tRNA genes (Jordán-Pla et al. 2015) to find the correlations, if any between the Pol III transcription and Spt16 occupancy. Classification of all the tRNA genes according to their transcription activity into three groups did not show any major difference in Spt16 occupancy between the three groups (Fig. 3A). No correlation between the mean occupancies of Spt16 on tDNAs and of Pol III on the nascent transcripts, individually or family-wise (Supplemental Fig. S3B,C) could be seen.

FIGURE 3.

Spt16 shows gene-specific influence on Pol III transcription. (A) Spt16 levels are not correlated with transcription from the tRNA genes. Average occupancy profiles of Spt16 on the three groups of tRNA genes classified according to their transcription activity, using the nascent transcription rates (nTRs) of individual genes from the data of Pol III nascentome (Jordán-Pla et al. 2015), kindly provided by A. Jordán-Pla. (B) A typical agarose gel profile of total RNA extracted as detailed in the Materials and Methods section is shown. Total RNA is resolved on the 2.8% agarose gel and stained with ethidium bromide. (C) ChIP-real time PCR measurements of the Pol III (Rpc160-myc) occupancy at some of the genes was made at both permissive and nonpermissive temperatures of growth. Occupancies were calculated by the % input method and normalized with the TELVIR region. Values for the SNR6 (U6 snRNA) TATA were divided by 2 to fit on the scale with others. (D) Measurements of RNA levels at 37°C in the wild-type (W) and spt16-197 (M) cells by northern blotting. Three biological replicates are shown. Pol II-transcribed SNR14 (U4 snRNA) gene was taken as positive control. Arrow marks the pre-tRNA, seen for some of the genes. (E) Quantification of total RNA levels at 37°C by qPCR in the spt16-197 cells normalized to wild-type RNA levels for selected Pol II- (U4 snRNA) and Pol III-transcribed genes. The P-values are (**) 0.0087, (*) 0.022619, (*) 0.011947, (***) 0.00011, and (****) 0.000002 in the order of their appearance on the graph panel.

We had earlier reported that FACT does not affect the transcription of a chromatinized tRNA gene in vitro (Mahapatra et al. 2011). However, Pol III-transcribed genes generally elicit gene-specific behavior under different conditions (Foretek et al. 2016; Turowski et al. 2016; Shukla and Bhargava 2018; Bhalla et al. 2019b). A gene-specific effect of Spt16 depletion, showing an increased or decreased transcription of some of the tested Pol II-transcribed genes was reported earlier (Jimeno-González et al. 2006). While Spt16 depletion was reported to reduce mammalian Pol I transcription (Birch et al. 2009), one of the well-characterized Spt16 amino-terminal, temperature-sensitive mutant spt16-197 is reported to show both positive and negative effects on yeast Pol II transcription (Malone et al. 1991; Cheung et al. 2008; Feng et al. 2016). We noticed that while the Pol I-transcribed rRNA levels decrease, the Pol III-transcribed 5S rRNA levels were unaffected at the nonpermissive temperature in both wild-type and mutant cells (Fig. 3B). In comparison, the similar total tRNA levels in the two cell types increase at 37°C, more so in the mutant cells (Supplemental Fig. S3D). As compared to the wild-type, gene-specific, low level increase or decrease of steady-state levels of different tRNAs was seen at 30°C (not shown) in the spt16-197 mutant cells, by using the tRNA-HySeq method (Arimbasseri et al. 2015) also.

It is generally difficult to detect and measure the primary transcripts of Pol III due to their quick processing and the sequence degeneracy of the tRNA isogenes. Therefore, Pol III occupancy is considered a measure of transcription. We found that higher temperature by itself does not affect the occupancy of the largest Pol III subunit Rpc160 on different tRNA genes in the wild-type cells (Fig. 3C). In comparison, Pol III occupancies could not be measured by ChIP in the spt16-197 cells due to very low but similar (close to the background and hence unreliable) Rpc160-myc signal at both the temperatures (Supplemental Fig. S3E). Rpc160-myc signal was found very low in westerns suggesting the myc epitope is most probably blocked by a cellular component in the mutant while total Rpc160 protein levels may be similar to wild-type cells (Supplemental File S1; Supplemental Fig. S3F,G). Our measurements of the tRNA levels by the northern blotting (Fig. 3D) or real time PCR method for some of the genes (Fig. 3E) at 37°C found highly reduced U4 (Pol II-transcribed) snRNA levels but no significant change in the tested tRNA levels in the spt16-197 mutant probably because of the high stability of the mature tRNA molecules (half-life of many hours, Hopper 2013). Although not accurately quantifiable, the similar levels of three (out of five tested genes) pre-tRNAs (Fig. 3D, marked by arrowheads) seen in northern blots indicate similar Pol III transcription in the wild-type and mutant cells. Similarly, out of the five tested tRNA genes, an intron-specific unique primer for the tP(UGG)O3 isogene measured its equal pre-tRNA levels in both the cell types (Fig. 3E). On the other hand, an increase of ∼2.5 fold in tT(CGU)K levels (Fig. 3E) indicates that in the spt16-197 cells, pre-tRNA levels and transcription of some tRNA genes increase, as suggested by the total tRNA level measurements (Fig. 3B). In comparison, the non-tRNA transcript levels were found significantly reduced in the spt16-197 mutant (Fig. 3D,E). U4 and U6 (Pol III-transcribed) snRNAs with reported half-lives of ∼24 h (Bordonné and Guthrie 1992), showed up to 80% loss (Fig. 3D,E). With no expected change for 4 h at 37°C (van Dijk et al. 2011), SCR1 levels decrease ∼40% (Fig. 3E), while despite similar stability as tRNAs (Lee et al. 1991), RPR1 showed ∼20% decrease (Fig. 3E) in the spt16-197 cells. The above changes of non-tRNA and tRNA levels reflect both positive and negative gene-specific regulatory effects of spt16-197 mutation on the transcription by Pol III.

Spt16 maintains the DS nucleosome position on tRNA genes

No effect on the majority of tRNA genes is consistent with an earlier study reporting that amino-terminal domain of Spt16 is reduntant for its transcription activation function but necessary for maintaining the chromatin repression of transcription on Pol II-transcribed genes (Evans et al. 1998). Yeast tRNA genes reside in the nucleosome-free regions (NFRs) flanked by the positioned nucleosomes on both sides (Kumar and Bhargava 2013). The DS enrichment of Spt16 is suggestive of its relevance to the DS nucleosome dynamics on the Pol III-transcribed genes, which was earlier reported to regulate Pol III transcription (Mahapatra et al. 2011). Comparative analysis of the nucleosome occupancy on tRNA genes in the wild-type and spt16-197 cells (True et al. 2016) revealed that the average occupancy of the US nucleosome is not affected but inactivation of Spt16 in the mutant results in a loss of nucleosome phasing specifically in the downstream region (Fig. 4A). We measured H3 levels on the same sets of DS region amplicons in both the cell types. The nonpermissive temperature did not cause any significant change in the DS nucleosome occupancy on the tested tRNA genes in either the wild-type (Fig. 4B) or the mutant cells (Supplemental Fig. S4A). As compared to the wild-type, a significant and gene-specific loss of DS H3 levels could be seen in the mutant cells at the nonpermissive temperature (Fig. 4C), while under the same conditions, Spt16 levels are significantly lower in the mutant cells (Fig. 4D). This loss could be due to the reported instability of the mutant Spt16 protein (Evans et al. 1998), which results in its lower levels in the spt16-197 cells (Supplemental Fig. S4B,C). In comparison, the total protein levels and DS occupancies of Spt16 on the tRNA genes in the wild-type and spt16-197 cells were found similar at the permissive temperature (Supplemental Fig. S4B–D).

FIGURE 4.

Spt16 maintains the DS nucleosome position at tDNA loci. All occupancies were measured by the ChIP and real time PCR (% Input) method. The dots mark nonsignificant changes. (A) Average nucleosomal (histone H3) occupancy profiles (True et al. 2016) aligned by the TSS of tRNA genes in the wild-type and spt16-197 cells. Profiles were normalized by setting the NFR minima to zero. (B) Comparison of relative H3 occupancies in the DS region of selected tRNA genes at 37°C, normalized with the respective H3 levels at 30°C in wild-type cells. (C) The relative H3 occupancy in the DS region of selected tRNA genes in the wild-type and spt16-197 cells show gene-specific changes with inactivation of Spt16 in spt16-197 cells at 37°C. The P-values are (*) 0.002582, (**) 0.000107, (***) 0.000079, and (****) <0.000001. (D) Occupancies of Spt16 in the DS region of the tRNA genes in the wild-type and mutant cells. The gene-specific occupancies show a significantly reduced level in the spt16-197 cells at 37°C on all the tested genes. The P-values are (***) 0.0008796, (**) 0.005094, (*) 0.017963, (**) 0.005608, and (**) 0.005002 in order of their appearance on the graph panel. (E,F) Nucleosome occupancy profile (H3 tag counts per million reads) on the tRNA genes (E) tP(UGG)O3 and (F) tT(CGU)K from the ChIP-seq data (True et al. 2016). The shaded box marks the gene body. (E) The US/DS nucleosome peaks are marked. (F) Asterisk marks the DS nucleosome peak. (G) Schematic representation of the positions of real time PCR amplicons on SCR1 and RPR1 genes. Bent arrow depicts the position of TSS. Gray boxes mark the transcribed region and numbers in parentheses denote the size of the primary transcript. Amplicon D in the case of RPR1 and A4 in the case of SCR1 map the 3′-end of the genes whereas A5 overlaps the strongly positioned DS nuclesosome on the SCR1 gene. (H,I) Nucleosome occupancies at the non-tRNA gene bodies increase in the mutant cells. Relative H3 levels at different positions on the (H) SCR1 gene and (I) RPR1 gene are shown. The P-values are for the amplicons (**) A1 = 0.000061, (****) A2 = 0.000002, (*) A3 = 0.000012, (***) A4 = 0.000004 in panel H and B (*) = 0.000053, (**) C = 0.000015 in panel I.

A loss of H3 density (Schwabish and Struhl 2004) and a shift in the position of the +1 nucleosome (True et al. 2016) on Pol II-transcribed genes in the spt16-197 cells was reported earlier. Spt16 depletion in another Spt16 mutant was reported to result in fuzziness and position shifts of the nucleosomes genome-wide (Feng et al. 2016). Accordingly, the reduced H3 levels (Fig. 4C) represent the delocalization and not a reduced occupancy of specifically the DS nucleosome in the spt16-197 mutant, which is reflected in the accompanying loss of the nucleosomal periodicity (Pusarla et al. 2007) further downstream (Fig. 4A). Therefore, the positioning of the DS nucleosome near the 3′-ends of the tRNA genes may be serving as a barrier for the statistical positioning of the nucleosomes further downstream (Pazin et al. 1997; Mavrich et al. 2008).

The available data on the tRNA genes (True et al. 2016) show that the DS nucleosome dynamics elicits a very subtle gene-specific effect of Spt16. For example, equal transcript levels of tP(UGG)O3 in both cell types (Fig. 3E) is matched by no change in the nucleosome profile on the gene (Fig. 4E) in the spt16-197 cells. On the tT(CGU)K gene in the spt16-197 cells, the DS nucleosome position is shifted ∼30 bp further away from the gene (Fig. 4F), which increases steric accessibility of the gene terminator. Consquential possibility of increased terminator-dependent Pol III recycling (Dieci and Sentenac 1996) may be the reason for the higher transcription/tRNA level for this gene (Fig. 3E). Similar to this, we had earlier seen a high increase in transcription of the tRNATyr gene in another spt16 mutant with the amino-terminal truncation, spt16-Δ922 (Mahapatra et al. 2011). A positioned DS nucleosome abutting the terminator on this gene becomes fuzzy in the spt16-197 cells (True et al. 2016). Thus, Spt16 affects tDNA transcription in coordination with the gene-specific DS nucleosome context on the tRNA genes. This complementary influence of Spt16 on transcription and the DS nucleosome mobility on the tRNA genes reveals a Spt16-dependent link between the two.

Spt16 depletes the 3′ end-proximal nucleosomes on the non-tRNA gene bodies

FACT was found necessary for chromatin organization of the Pol II-transcribed gene regions (Jimeno-González et al. 2006) and to repopulate the gene body nucleosomes after transcription by Pol II (Voth et al. 2014). SCR1 and RPR1, the two longest non-tRNA genes among the Pol III-transcribed genes, give the transcripts of the sizes 522 and 486 bases, respectively (Fig. 4G). Similar to tRNA genes, they have positioned nucleosomes at both the gene ends but unlike tRNA genes, their transcribed regions have low nucleosome occupancy specifically toward the 3′-ends (Supplemental Fig. S4E,F; Dewari and Bhargava 2014). We argued FACT may facilitate the transcription of these long Pol III-transcribed genes by negotiating with the gene body nucleosomes. Our measurements of nucleosome occupancy (histone H3 levels) at different locations on the transcribed regions of both the genes (Fig. 4G) showed significantly higher H3 levels on SCR1 and RPR1 gene bodies (Fig. 4H,I), denoting a gain of nucleosomes on their gene bodies in the spt16-197 cells. No change was seen at the gene-end amplicons A5 (SCR1, Fig. 4H) and A or D (RPR1, Fig. 4I) as no transcription activity is expected beyond the gene region. The presence of nucleosomes on the transcribed gene bodies is generally inhibitory for transcription elongation. The eviction of nucleosomes on the transcribed gene bodies of the SCR1 and RPR1 genes by Spt16 in the wild-type cells thus explains the observed decrease in transcript levels of both the genes in the spt16-197 mutant (Fig. 3E). Residing in an NFR (Kumar and Bhargava 2013) may alleviate the need of Spt16 on the tRNA gene bodies during transcription. In agreement, no H3 (nucleosome) gain was seen on the tDNA NFR in the spt16-197 cells, though the DS nucleosome is delocalized (Fig. 4A). We conclude that Spt16 enables disruption/eviction of the nucleosomes on the body of the SCR1 and RPR1 genes, facilitating their transcription by Pol III in turn.

Spt16 is lost from the DS region under nutritional stress

Transcription by Pol III is curtailed under most of the stress conditions (Clarke et al. 1996; Willis and Moir 2007 Moir and Willis 2013) with a loss of Pol III from all the target genes under nutritional stress (Roberts et al. 2003; Kumar and Bhargava 2013). The physical interaction of Pol III and Spt16 invokes a possibility of their similar dynamics on the tRNA genes under similar conditions. Under the starvation condition, the average Spt16 occupancy on the ORFs showed a loss at the 5′ end of the genes (Fig. 5A). On the tRNA genes, Spt16 in the DS region reduced to the US levels (Fig. 5B,C), while its levels on the gene body where TFIIIC and TFIIIB stay on (Roberts et al. 2003), do not change under starvation (Fig. 5B). The Spt16 loss from the DS region was not due to a change in total Spt16 protein levels; which did not reduce over a period of starvation up to 2 h (Fig. 5D). The ChIP and real time PCR quantifications on some of the genes (Fig. 5E) validated this loss.

FIGURE 5.

Spt16 is lost from the 3′-end of tRNA genes under nutrient starvation. Spt16 occupancy was measured in the wild-type cells grown under active (normal growth) or repressed (starvation) conditions. (A,B) Average Spt16 tag counts (normalized to per million reads) from our genome-wide ChIP-seq data; all aligned by the TSS are plotted. (A) On the Pol II-transcribed ORFs, reduced Spt16 levels are found near the 5′ gene ends under starvation condition. (B) Comparison of average Spt16 occupancy profiles on the tRNA genes under active and repressed conditions. (C) Heat maps of Spt16 occupancy within 1 kb upstream and downstream from the gene ends on all the tRNA genes (red block). Color gradient code is shown at the bottom. (D) Total protein levels of Spt16 do not change under repression for up to 2 h. TCA precipitated total proteins were resolved by SDS-PAGE, western transferred and probed with anti-Spt16 antibody. Actin was used as a loading control. (E) Relative Spt16–TAP occupancy under active or repressed conditions on some of the tRNA genes, as measured by ChIP and real time PCR. The P-values in order of their appearance on the graph are 0.072805, (**) 0.000743, (**) 0.000979, (*) 0.005788, (***) 0.000016, and (**) 0.000437.

Spt16 is lost from the DS regions of the three non-tRNA genes, the US region of SCR1 and from the transcribed region of the RPR1 gene under repression (Supplemental Fig. S5). The loss of Spt16 upon nutrient starvation affirms that the high Spt16 level in the DS region in the normal state is specific to the transcriptionally active state of the genes and suggests that Spt16 enrichment is associated with Pol III transcription on the genes.

Spt16 enrichment at the 3′ gene ends follows Pol III transcription

Our results invoke a possibility of Spt16 recruitment by TFIIIC on the gene body. At this point, we asked if Spt16 is recruited on the gene body, how does it get enriched in the DS region?

In a recent genome-wide study, FACT and Pol II have been reported to track together on the gene bodies (Vinayachandran et al. 2018). The physical association (Fig. 1) and cooccupancy on the gene body (Fig. 2B) of Spt16 and Pol III suggest that Spt16 associates with the transcribing Pol III, which delivers Spt16 to the 3′ gene end upon termination, where Spt16 levels build-up with multiple rounds of transcription. According to this hypothesis, reduced number of transcription cycles would result in reduced deposition of Spt16 at the 3′ end, as suggested by the higher Spt16 occupancy on the genes with higher rate of transcription by Pol II (Supplemental Fig. S3A; Pathak et al. 2018). To check for a similar possibility, we used cells carrying a mutant Pol III rpc160-112, which can form a productive TC and initiate normally but shows a reduced transcription elongation rate (Dieci et al. 1995). We monitored the Pol III and Spt16 occupancies on several tRNA (short gene length) and two non-tRNA (comparatively longer) genes at their 5′ and 3′ gene ends, respectively, where their highest levels are found (Fig. 6).

FIGURE 6.

DS enrichment of Spt16 depends on the transcription process. Cells harboring 3XHA-tagged either the wild-type (control) or mutated Rpc160 (rpc160-112) were used to follow the effect of slow elongation on Spt16 occupancy (using Spt16-specific antibody) on different Pol III-transcribed genes. Relative Pol III (top panels) and Spt16 (bottom panels) occupancies were measured at the 5′ and 3′ gene ends, respectively. All enrichments (% Input method) on different Pol III-transcribed tRNA or non-tRNA genes were measured by the ChIP-real time PCR method. (A,B) Pol III and Spt16 enrichments require normal Pol III transcription on tRNA genes. (A) Pol III occupancies in the upstream and 5′ gene end. For all the genes, P < 0.00001 and (B) Spt16 occupancy in the DS region of selected tRNA genes. For all the genes, P < 0.0001. (C,E) The Pol III occupancies at different locations over the (C) SCR1 and (E) RPR1 gene body in the Pol III mutant compared to the control cells. P < 0.00001 for A1, A2 (panel C), and (**) A (panel E); <0.00007 for (*) B (panel E) amplicons. (D) Spt16 occupancy at different positions on the SCR1 gene are lower than the wild-type levels in the Pol III mutant cells. The P-values are 0.0006 for (**) A1, 0.0139 for (*) A2, 0.0005 for (**) A3 and (**) A4, 0.0001 for (**) A5. (F) On the RPR1 gene, mutant Pol III pausing at the 5′-half of the gene body (panel E, at A and B) results in (F) lower Spt16 levels at the rest of the gene in the 3′-half region (P-value = 0.0023 for [*] A, 0.1133 for B and 0.0005 at [**] D).

On the tRNA genes

As compared to the control (rescued) cells, we found gene-specific and significantly higher Pol III levels in the US, 5′- as well as the 3′-end and DS regions of the tRNA genes in the rpc160-112 mutant cells (Fig. 6A; Supplemental Fig. S6A). This denotes a slow progress and stalling of the mutant Pol III due to the slow polymerization at its active centre (Dieci et al. 1995) and a slow clearance/release at the terminator. Pol III mutation does not affect the total Rpc160 or Spt16 protein levels (Supplemental Fig. S6B,C). Consistent with its absence in the US region of tRNA genes, Spt16 levels in the Pol III mutant cells remain significantly lower than the control background level (Supplemental Fig. S6D). In contrast, DS Spt16 levels were significantly reduced in the downstream region (Fig. 6B) of the tRNA genes. Similar to this, inhibition of transcription elongation by an exogenous inhibitor was found to result in loss of DS Spt16 on the tRNA genes (Martin et al. 2018).

The compromised Spt16 levels at the 3′-end as a consequence of the slow elongation by Pol III imply that the DS Spt16 enrichment on tRNA genes coincides with traversing of the gene by a normally transcribing Pol III. This conclusion is supported by the apparently different peak positions of Spt16 and Pol III (Fig. 2B) and differences in the occupancy profiles of Spt16 in the active and repressed states (Fig. 5B). Together, both the profiles corroborate the enrichment of Spt16 at the 3′ gene end as a result of its delivery there by Pol III, specifically in the transcriptionally active state.

On the non-tRNA SCR1 and RPR1 genes

Pol III shows a uniform distribution on the transcribed region of the SCR1 gene (Fig. 6C; Supplemental Fig. S6E), while Spt16 is found more enriched on the amplicons A4 and A5 in the wild-type cells (Fig. 6D). Similar to tRNA genes (Fig. 6A), Pol III elongation mutant showed an increase of Pol III levels at the 5′ end and body of the SCR1 gene (Fig. 6C). The similarly reduced levels of both wild-type and mutant Pol III at the amplicon A5 indicate that mutant Pol III does not accumulate beyond the transcribed region. It shows a processivity lower than the wild-type Pol III processivity (Dieci et al. 1995) and hence a gradual reduction in occupancy from the 5′ to the 3′ gene end of the SCR1 gene (Supplemental Fig. S6F). As compared to the wild-type, Spt16 occupancy reduced near the TSS (amplicon A1) and at every position on the gene body in the Pol III mutant cells (Fig. 6D). Smallest change was at the amplicon A2, a position after which the mutant Pol III levels stop increasing (cf. Fig. 6C,D). Spt16 levels at the 3′-end (amplicons A3, A4, A5) near the terminator showed a highly significant decrease on SCR1 in the Pol III mutant cell (Fig. 6D). As wild-type Pol III levels drop at the 3′ end, Spt16 levels increase there (Fig. 6C,D), suggesting their disengagement at the terminator. In comparison, loss of mutant Pol III levels at the 3′ gene end resulting in loss of Spt16 there (Fig. 6C,D) suggests that Spt16 enrichment at the 3′-end depends on its delivery there by Pol III.

As compared to SCR1, Pol III shows an unusual distribution on the RPR1 gene, which is suggestive of a Pol III pausing on the amplicon regions B and C near the 3′ gene end (cf. Supplemental Fig. S6E,G). The higher nucleosome density on the RPR1 than on the SCR1 gene body (Supplemental Fig. S4H vs Supplemental Fig. S4I) may be the reason for this difference between the two genes in the wild-type cells. Higher than wild-type levels of the mutant Pol III at the amplicons A and B on the RPR1 gene body (Fig. 6E) may be due to the longer RNA synthesis time and longer pausing of the rpc160-112 Pol III at the intrinsic pause sites (Dieci et al. 1995) as well as higher H3 levels (Fig. 4I) further downstream. Apart from enrichment in the DS region, Spt16 levels on the RPR1 gene body in the control cells show an overlapping peak and colocalization with Pol III at probably the intrinsic pause site (marked with asterisks, Supplemental Fig. S7G). As the mutant Pol III is strongly paused at the amplicon B (Fig. 6E), Spt16 levels show decrease on the amplicons C and D (Fig. 6F). This result clearly shows a dependence of Spt16 levels on the transcription elongation activity of Pol III and explains why Spt16 effect on transcript (Fig. 3G) or gene body nucleosome (Fig. 4H,I) levels is less for RPR1 than for the same on the SCR1 gene.

The dependence of Spt16 levels on the transcription elongation rate rather than the Pol III occupancy on the SCR1, RPR1, and tRNA genes supports a possibility that FACT joins and follows the Pol III engaged in elongation on the gene body. A comigration of both on the gene body during transcription is implicit, which helps build up higher Spt16 levels at the DS region of the Pol III-transcribed genes.

Spt16 loss from the DS region under starvation is a stress response

Above results show that Spt16 and Pol III occupancies are not correlated (Fig. 3) but the DS enrichment of Spt16 on the tRNA genes follows transcription by Pol III (Figs. 5, 6). Similarly, the occupancy levels of Pol III (Kumar and Bhargava 2013) and DS Spt16 (this study) in normal and starvation conditions were found only poorly correlated (Supplemental Fig. S7A). In order to resolve it further, we used a small molecule inhibitor (Wu et al. 2003) to cause only partial inhibition of transcription. Using the Pol III-specific LD50 dose of ML60218 (Millipore), in a time course study, we found Pol III occupancy is reduced by 50% in 2 h (Fig. 7A, P < 0.025), as expected. Therefore, the inhibitor either dislodges the bound Pol III or affects its reloading on the genes at the initiation step. In comparison, the DS Spt16 occupancy is maintained even after 2 h exposure to the inhibitor (Fig. 7B). Measurements of total Spt16 protein levels showed the relative levels do not change with inhibitor exposure time (Supplemental Fig. S7B,C). Together, all these results demonstrate that a partial loss of Pol III does not cause loss of Spt16 already bound to the gene at the 3′ end. Therefore, the loss of Spt16 with complete loss of Pol III under starvation (Fig. 5B,E) but no Spt16 loss with 50% reduced Pol III levels (Fig. 7C) suggest that the downstream effects of transcription inhibition by Maf1 during starvation (Pluta et al. 2001; Karkusiewicz et al. 2011; Boguta 2013) and by Pol III-specific inhibitor ML60218 are different.

FIGURE 7.

Occupancies of Spt16 change with the environmental conditions. (A,B) Spt16 levels do not change with the level of transcription. Median lethal dose (LD50) of Pol III-specific inhibitor (ML-60218, Millipore) was added in the YEPD media before cross-linking in the ChIP experiment and samples were monitored at 0, 60, and 120 min after the addition. (A) Relative Pol III occupancy at the 5′ end. For plotting on the same scale, values for the genes marked with dots were divided by two. (B) Relative Spt16 occupancy at the 3′-end of the selected tRNA genes, as determined by ChIP-qPCR are shown. (C) Comparison of average Spt16 profiles aligned by the gene body of Pol II-transcribed ESR genes under normal and nutrient starvation conditions. The gene body of each ORF is normalized to 2 kb in length. (D) DS Spt16 occupancy measurement on tRNA genes by ChIP and real time PCR method. Shifting of wild-type cells from permissive to nonpermissive temperature for growth causes a significant increase. The P-values are 0.03288 (*), 0.009197 (**), 0.010967 (*), 0.00676 (**), and 0.0099 (**) in the order of their appearance. (E) Wild-type cells were grown to 0.6 OD600nm in the YEPD and treated with 200 mM genotoxin hydroxy urea (HU) for 2 h before harvesting. DS Spt16 occupancy (normalized to ORF-free region) on the selected tRNA genes was measured by the ChIP and real time PCR. The P-values in the order of their appearance on the graph are (**) 0.008304, (***) 0.00003, (**) 0.004697, 0.057374, and (*) 0.012096.

The above results demonstrate that the DS Spt16 loss under starvation requires the cues specific to nutrient stress signaling. Spt16 dynamics on a set of Pol II-transcribed genes, identified as environmental stress response (ESR) genes in yeast (Gasch et al; 2000) corroborates the same (Fig. 7C). We found that with the loss of Spt16 from the normally Spt16-occupied active ESR genes in our data, the occupancy on the inducible genes goes up under nutrient starvation (Fig. 7C). Similar to nutritional stress, reciprocity of the FACT occupancy on the induced and repressed genes has been seen upon heat shock as well (Vinayachandran et al. 2018). Measurements of Spt16 occupancy under a few more stress conditions suggested a general stress-responsive role of Spt16. At the nonpermissive growth temperature, Pol III occupancy does not change (Fig. 3C) but Spt16 occupancy on tRNA genes in the wild-type cells increases (Fig. 7D). Simultaneously, the total cellular Spt16 protein levels increase upon shifting of wild-type cells to higher growth temperature (Supplemental Figs. S4C, S7D). The amino-terminal domain of Spt16 is required for cell growth under HU-induced replication stress (O'Donnell et al. 2004). Consistent with this, Spt16 occupancy on the tRNA genes goes up during replication stress (Fig. 7E), while Pol III levels are reported to go down significantly under replication stress (Bhalla et al. 2019b). The above results indicate Spt16 involvement in cellular stress-sensing/response mechanisms. Like a true stress sensor, Spt16 protein as well as its gene occupancy level changes in the cells upon exposure to different stress conditions.

Taken together, the genome-wide mapping in this study found that the FACT subunit Spt16 shows transcription elongation-dependent enrichment in the DS region and plays a dual role on the Pol III-transcribed genes. Though Spt16 interacts with the Pol III TC even off the DNA, its DS occupancy is not correlated with Pol III occupancy or transcription of the nucleosome-free tRNA genes. On the non-tRNA genes, Spt16 depletes the gene body nucleosomes and facilitates their transcription. Its gene-specific occupancies elicit different regulatory (positive, negative or neutral) effects on tRNA transcription as it maintains the position of the DS nucleosome, reported earlier to influence the transcription from a tRNA gene. Nutrient stress signaling and not merely transcription inhibition reduces Spt16 levels at the 3′-ends of tRNA genes under transcriptional repression due to starvation. In comparison, Spt16 occupancy shows both increase and decrease under starvation on the Pol II-transcribed ESR genes in a gene-specific manner. Thus, on the Pol III-transcribed genes, transcription and DS nucleosome dynamics are coordinated by FACT in a gene-specific and stress-responsive manner.

DISCUSSION

TFIIIC/TFIIIB recruit FACT on the Pol III-transcribed genes

FACT associates with Pol II-transcribed genes immediately downstream to the PIC (Mason and Struhl 2003) and influences transcription initiation. In this study, Spt16 is found on the tRNA gene bodies with Pol III, where TFIIIC is also bound. Yeast FACT requires Nhp6 as an anchor for binding of the Spt16–Pob3 to the nucleosomes (Ruone et al. 2003). The absence of Spt16 in the US region, higher Nhp6 association with TFIIIC than with TFIIIB or Pol III and the highest FACT association with TFIIIB in our mass-spectrometric data suggest that Spt16 is not recruited by its binding to the US nucleosome. It interacts with TFIIIB on the gene under repression and upon activation, with binding of Pol III and weakened TFIIIB-TFIIIC association (Ciesla et al. 2018), Spt16 is probably transferred to Pol III and TFIIIC on the gene body. The locations of TFIIIB (recruits Pol III) and TFIIIC (recruits TFIIIB) near the 5′ gene end may enable this. Under repression, with increased interaction of TFIIIB and TFIIIC (Ciesla et al. 2018) and abolition of transcription-dependent DS enrichment of Spt16, Nhp6 may stabilize its interaction with TFIIIC on the gene body. Thus, the Spt16 interactions with the three components of the Pol III TC enable a transcription-dependent Spt16 dynamics on the Pol III-transcribed genes, orchestrated by the Pol III TC.

Transcription-dependent enrichment of Spt16 in the DS region

The physical interaction of Spt16 and Pol III and the transcriptional dependence of Spt16 enrichment in the DS region of tRNA genes suggest that Spt16 piggybacks on the Pol III to the DS nucleosome at the 3′ gene end. Spt16 was reported earlier to join the Pol II near the TSS and disengage with the elongation complex upstream of the polyadenylation site at the 3′-end of Pol II-transcribed ORFs (Mason and Struhl 2003; Mayer et al. 2010). Similar to this, it is highly likely that the association of TFIIIC–Pol III–Spt16 found on the gene body is disrupted at the terminator, as Spt16 finds and binds the DS nucleosome in the vicinity via its histone-binding domain (Zheng et al. 2014; Kemble et al. 2015; Mao et al. 2016). DS nucleosome serves as a sink for Spt16 at the end of each transcription cycle and repetition of this process builds up Spt16 levels at the 3′ gene end. In the case of the mutant Pol III, a slow elongation rate of transcription slows down this process in every cycle and hence lowers the Spt16 enrichment at the 3′ gene end. This is consistent with the different peak positions of Spt16 and Pol III and the loss of Spt16 with Pol III during starvation from the DS region.

Spt16 facilitates transcription of the non-tRNA Pol III-transcribed genes

The Pol III-transcribed non-tRNA genes differ from tRNA genes in several aspects like the length of the transcribed region and chromatin structure around them, which may influence their transcription. FACT is reported to restore the gene body nucleosomes evicted due to continuous histone exchange during transcription (Schwabish and Struhl 2004; Voth et al. 2014). A high level Asf1-dependent, replication-independent histone H3 exchange on the NFR flanking nucleosomes has been reported on the Pol II-transcribed genes (Rufiange et al. 2007). No histone exchange was seen on the SNR6 gene (Dewari and Bhargava 2014), where a positioned nucleosome on the gene body was earlier reported to support high levels of transcription (Shivaswamy et al. 2004; Shivaswamy and Bhargava 2006; Arimbasseri and Bhargava 2008). In comparison, an Asf1-dependent H3 exchange was found at the 3′-end and 3′-half of the gene body of other Pol III-transcribed non-tRNA genes (Dewari and Bhargava 2014). TFIIIC binding to the short transcribed regions on the tRNA genes may exclude the nucleosome from there and support the NFR maintenance by the chromatin remodelers for uninhibited transcription (Kumar and Bhargava 2013). In comparison, TFIIIC presence toward the 5′-half of the gene body may be the reason that on the non-tRNA genes, with a length longer than the tRNA genes, nucleosomes are seen toward the 3′-half of the genes. Therefore, additional mechanisms may be needed to overcome the nucleosomal inhibition on the non-tRNA genes. The transcription-dependent nucleosomes and histones depletion by Spt16 and Asf1 from the 3′-half of the RPR1 and SCR1 genes resolves this and facilitates the transcription by Pol III.

The differential and gene-specific transcription by Pol III

Spt16 depletion results in reduced Pol II occupancies on the Pol II-transcribed genes (Pathak et al. 2018). In mammalian cells, FACT depletion results in lower 7SL RNA and tRNATyr levels (measured by RT-qPCR) due to a reduced Pol III transcription (Birch et al. 2009). The sequence-based measurement methods for tRNAs suffer a setback in designing unique primers for the primary transcript levels from individual genes. Nonavailability of the suitable and unique primers due to the multiple isogenes found in any single family often results in nondetection of individual gene changes within the family. Nevertheless, using these methods, we could guage the gene-specific primary transcription of five out of the ten tested genes in our measurements. For better understanding of the results, we argued that while lower tRNA levels could be due to a lower stability/maturation process and not necessarily lower tanscription; higher tRNA levels may not result without an increase in transcription. Similarly, equal steady-state tRNA levels could mean either similar transcription and maturation or higher transcription but reduced stability/maturation process to balance. With the majority of steady-state tRNAs at similar levels in the wild-type and mutant cells, we could conclude that the spt16-197 mutation either does not affect or increases the tRNA gene transcription, in a gene-specific manner.

Shifting to higher temperature or other stress conditions was found to affect the end-processing of tRNAs, resulting in accumulation of aberrant pre-tRNA products with extended 3′ termini (Foretek et al. 2016). Though it is not known whether pre-tRNA processing is defective in the spt16-197 cells, the higher total tRNA levels found at 37°C in both cell types could be due to the gene-specific aberrant transcripts accumulation. In this context, the ∼30 bp further downstream shift of the DS nucleosome on the tT(CGU)K gene in the mutant cells may enable the synthesis of the 3′ extended transcript beyond the gene terminator. Such a readthrough of the terminator may be a general feature of yeast Pol III transcription (Rijal and Maraia 2016; Turowski et al. 2016; Leśniewska and Boguta 2017), which may be modulated by a DS nucleosome, depending on the context.

Nucleosomal organization is the main Spt16 role on the Pol III-transcribed genes

One of the reasons for the high transcription rate of the Pol III-transcribed genes could be the terminator-dependent facilitated recycling of Pol III (Dieci et al. 2002, 2013b; Dieci and Sentenac 2003; Arimbasseri et al. 2013; Leśniewska and Boguta 2017). Spt16 maintains the DS nucleosome position and hence transcription terminator accessibility, which eventually influences the rate of transcription. Keeping the DS nucleosome away makes the terminator more accessible (Mahapatra et al. 2011), while moving the DS nucleosome near the terminator would affect the rate of transcription by causing steric hindrance to the recycling Pol III.

On the tRNA gene bodies, the steric block of the gene body due to TFIIIC binding, high rate of transcription and concerted activities of several chromatin remodelers generate the NFR (Kumar and Bhargava 2013). An association with TFIIIC may allow Spt16 to coordinate with these activities during transcription. Hence, unlike its role on the non-tRNA Pol III- and Pol II-transcribed genes, nucleosome eviction is not the primary role of Spt16 on the tRNA gene body. Its transcription-associated enrichment is seen only in the DS region where it influences the DS nucleosome dynamics, reported to influence the gene terminator accessibility (Mahapatra et al. 2011; Kumar and Bhargava 2013).

The DS nucleosome dynamics is linked to transcription activity with precision conferred by different chaperones and remodelers like RSC, ISW (Kumar and Bhargava 2013) and FACT (this study) on the tRNA genes. With a multitude of such factors in play, where each one may even individually be modulated, every gene may be expressed in a manner specific to its own context. Indeed, different isogenes belonging to the same tRNA family are expressed differently (Kumar and Bhargava 2013; Bloom-Ackermann et al. 2014; Turowski et al. 2016). A shift in either direction (delocalization) of the DS nucleosome in the spt16-197 mutant could result in either increase or decrease of the transcription, giving gene-specific expression regulation of Pol III-transcribed genes, as demonstrated by examples of some of the genes in this study. Accordingly, different chromatin context of different genes may be the cause of the gene-specific effects of Spt16 on transcription.

Regulating nucleosome dynamics at the tDNA loci by Spt16 is a stress response

Recent evidence demonstrates the involvement of FACT in chromatin dynamics rather than the actual process of RNA synthesis (Gurova et al. 2018; Martin et al. 2018; McCullough et al. 2019). Our results also emphasize a chromatin structure rather than the transcription modulation as the primary role of Spt16 on the Pol III-transcribed genes. On these genes, the position of the US nucleosome does not change for short-term abolition of transcription with the environmental and growth condition changes. In contrast, position of the DS nucleosome found at various distances from the TTS changes under different conditions (Kumar and Bhargava 2013). Maintenance of the NFR and the DS nucleosome position by the chromatin remodelers and FACT (this study) underscores the importance of DS nucleosome in the regulation of the tRNA gene transcription (Mahapatra et al. 2011; Kumar and Bhargava 2013). The DS Spt16 enrichment in the normal condition and its loss from the 3′ gene ends under starvation (nutritional stress) may be constituting a stress response elicited via the DS nucleosome dynamics rather than the transcription by Pol III. Linking the transcription to the DS nucleosome dynamics via Spt16 involvement enables this response. Therefore, on the Pol III-transcribed genes, a role for Spt16 in the DS nucleosome mobility and the dependence of its DS enrichment on the transcription by Pol III may play a regulatory role serving as a target of environmental cues that influence the tDNA transcription.

FACT is recruited to replisome and helps maintain chromatin structure during replication fork progression (Foltman et al. 2013). It binds DNA Pol α (Wittmeyer et al. 1999), participates in the replication-coupled nucleosome assembly (Yang et al. 2016) and enhances transcription-coupled DNA repair (Wienholz et al. 2019). Consistent with this, increase of Spt16 occupancy (this study) on the tRNA genes under replication stress suggests a role for Spt16 in resolving transcription–replication conflict-associated R loops on the tRNA genes (El Hage et al. 2014; Herrera-Moyano et al. 2014). In an earlier study, the spt16-197 mutation was found to induce genome-wide cryptic transcription by Pol II (Cheung et al. 2008). The nutritional stress was found to transiently induce the same cryptic transcription from a set of the genes in the wild-type cells (Cheung et al. 2008), suggesting a connection between the nutrient starvation and Spt16. Under nutrient starvation, the DS nucleosome mobility toward the terminator was seen on many tRNA genes (Mahapatra et al. 2011; Kumar and Bhargava 2013). Earlier reported, widespread repression of transcription from most of the tRNA genes under the stress condition (Cieśla et al. 2007; Turowski et al. 2016) may also be due to the similar control of the terminator accessibility by the DS nucleosome. Its subtle dynamics on the Pol III-transcribed genes under nutritional, heat and replicative stress and on the ESR genes under starvation are consistent with a genome-wide, general stress-responsive role for Spt16.

The yeast Pol III-transcribed genes are generally not known to have upstream regulatory elements (Orioli et al. 2012). The transcription-associated, directional deposition of Spt16 and hence dynamics of the DS nucleosome may be well suited for the regulatory purposes. Therefore, targeting the DS nucleosome and associating with Pol III and transcription gives Spt16 a direct handle to sense the transcription status and translate it into the DS nucleosome dynamics on the tRNA genes. Thus, Spt16 at the tDNA loci functions in the transcription status signaling to the downstream events like chromatin remodeling under a stress condition like starvation.

MATERIALS AND METHODS

Yeast strains, media, and growth conditions

Yeast strains and primer sequences used in this study are given in the Supplemental Tables S2 and S3, respectively. Pol III elongation mutant rpc160-112 (Dieci et al. 1995) and its control cells, provided by Olivier Lefebvre had the mutated or wild-type RPC160 gene, tagged with 3XHA at its genomic locus. Yeast cells were grown in YEP (yeast extract and peptone) medium containing 2% dextrose at 30°C unless stated otherwise. For nutrient starvation condition, cells grown to OD600 nm of 0.8 in YEP media with dextrose were shifted to 0.15× YEP media lacking dextrose for 2 h (Arimbasseri and Bhargava 2008). For a time course study with the Pol III-specific small inhibitor ML60218 (Millipore), the SJY25 cells carrying myc-tagged SPT16 gene were grown to OD600 nm of 0.6 in YEPD media before exposing to the LD50 dose of the inhibitor. Unless otherwise stated, the temperature-sensitive Spt16 mutant strain spt16-197 was grown to OD600 nm 0.6 at 30°C and shifted to 37°C for 90 min before harvesting and further processing. PCR amplified toolbox cassettes (Janke et al. 2004) were used for the deletion or epitope tagging of the desired gene. All taggings were confirmed by the genomic DNA sequencings and western blottings. Tagging of genes per se in any of the cells did not cause any adverse phenotypes like thermo-sensitivity or slow growth.

Antibodies

Anti-H2B (Active motif, 39237), anti-H3 (Abcam, ab1791), anti-actin (Abcam, ab8224), anti-Myc (Millipore, 05-724), anti-HA (Millipore, 05-904), anti-Flag (Millipore, MAB3118), anti-Pgk1 (Novex, 459250) antibodies were purchased and a polyclonal anti-Spt16 (True et al. 2016) antibody was a kind gift from Tim Formosa.

Chromatin immunoprecipitation (ChIP), ChIP-seq, and qPCR

ChIP experiments were performed at least in triplicates as described previously with slight modifications (Aparicio et al. 2005; Arimbasseri and Bhargava 2008). Briefly, cells were grown to mid-log phase and cross-linked with 1% formaldehyde for 30 min (15 min for histone ChIP). Cells were lysed and chromatin was fragmented to a length of 150–350 bp using sonication (MNase digestion in the case of histone ChIP). Samples were prepared by using Protein A/G Sepharose beads with specific antibodies for ChIP or nonspecific antibody IgG for mock immunoprecipitations. For Spt16 genome-wide mapping, samples were prepared from Spt16–TAP tagged cells using IgG Sepharose beads. A parallel mock immunoprecipitation was also performed using Protein A/G Sepharose beads without adding any antibody. The beads were washed and DNA was extracted.

The ChIP DNA was used either for high-throughput sequencing or analyzed by quantitative real time PCR (qPCR) using SYBR Green (Roche) chemistry for locus-specific protein occupancies. The % Input method and the “Relative fold enrichment method” were used to calculate the relative enrichments over the background for Pol III, Spt16 and H3 using TELVIR, ORF-Free region 1 or ORF-free region 2 as internal control regions, respectively (Supplemental Table S3). In the fold enrichment method, Ct values obtained from the ChIP and Mock samples were first normalized with Ct values of Input and then by Ct values obtained from the control region. Average and scatter from minimum three independent estimations are plotted.

CoImmunoprecipitation (co-IP) and western blotting

Yeast cells doubly tagged (two different tags for two different proteins in the same cell) were used for coimmunoprecipitations (Vernekar and Bhargava 2015) from the DNase I-treated whole cell extracts. Anti-Flag and anti-HA immunoprecipitations (IPs) were performed from the extracts of the strains harboring Myc-tagged Rpc160 or Flag-tagged Tfc1/Brf1 and HA-tagged Spt16. The immunoprecipitated proteins bound to the Flag M2 agarose or HA or Myc antibody-bound Protein A/G Sepharose beads were boiled in SDS-containing sample buffer and resolved by SDS-PAGE. The immunoprecipitates (IPs) were western transferred to PVDF membrane and probed sequentially with different antibodies to ascertain the co-IP. For measuring the cellular levels of a protein, total proteins were TCA precipitated from the whole cell extracts and the protein precipitate was dissolved in the 1× Laemmli buffer for resolving by SDS-PAGE. Equal level proteins were loaded per lane. The western blots were probed with antibodies specific for the native protein, if available or against the epitope tag attached to the protein of interest. Pgk1 levels were used for normalizations. Protein A/G beads without antibody were used in mock immunoprecipitations. Whole cell extract (Input) was used as a positive control. Multiple western blots for co-IPs of TFIIIB, TFIIIC and Pol III with Spt16 were quantified using ImageJ software (Fuji). The ratios of the co-IP and IP bands in western blots for a pair were calculated for both forward and reverse co-IPs. Average ratios and scatter for a minimum of three biological replicates have been plotted.

For the co-IP experiments in vitro, the purified RNA Pol III (lab stock) was immobilized on the Flag M2 agarose beads using its Flag-tagged Rpc128 subunit and incubated overnight at 4°C with the DNase I-treated crude extract of the SJY25 cells (Jimeno-González et al. 2006) carrying Spt16-18xMyc. The beads were washed, and proteins were eluted with Flag peptide, resolved by SDS-PAGE, western transferred and probed for co-IP with the anti-Myc antibody.

RNA isolation and quantification

For RNA extractions cells were grown at 30°C till OD600 nm was 0.6 and then continued at 30°C or shifted to 37°C for 90 min before harvesting. Total cellular RNA was isolated using the acidic hot phenol method as previously described (Arimbasseri and Bhargava 2008). Typically, 600–1000 ng RNA was used for anaysis on a 2.8% agarose gel, run in 0.5× TBE buffer. Bands were visualized under UV exposure after ethidium bromide staining. The Pol III-transcribed total RNA was measured by the Real-Time PCR method using 5 µg of RNA for cDNA synthesis (Kumar and Bhargava 2013). For all measurements made, averages from at least three biological replicates and scatter have been plotted. For northern blotting (Karkusiewicz et al. 2011; Vernekar and Bhargava 2015), 15 µg of RNA was used per sample. Pol II-transcribed U4 snRNA gene was used as positive control in the northern blots, which were sequentially probed with the 5′ end-labeled gene-specific oligos and visualized in a Bio-Rad PhosphorImager. RNA levels were also estimated by the tRNA-HySeq method, essentially as described previously (Arimbasseri et al. 2015) and detailed previously (Bhalla et al. 2019b).

ChIP-seq and MNase ChIP-seq data analysis

The ChIP-seq data for Spt16 and Pol III were processed and analyzed as described earlier (Kumar and Bhargava 2013) with minor modifications. The reads uniquely aligned to sacCer3 yeast assembly with mapping quality scores >30 were processed using the “dpeak” function of DANPOS 2.2.2 package (Chen et al. 2013). The resulting files were normalized for sequencing depth (per million reads) and background (mock IP with Protein A/G or CL4B Sepharose beads, respectively) was subtracted. Normalization with Input instead of the mock returned the same occupancy profile showing the highest Spt16 levels at the 5′ end of ORFs and in the DS region on the tRNA genes (Supplemental Fig. S1). Further analysis details are given under the Supplemental File S1. MNase ChIP-seq H3 data for the wild-type or spt16-197 mutant (True et al. 2016) were analyzed as described above using the “dpos” function of DANPOS (Chen et al. 2013). The “Profile” function of DANPOS was further used to compute the mean signal intensity at each position relative to TSSs of Pol II- and Pol III-transcribed genes. Although we did not find any change in the average Spt16 profile; tRNA genes in the close vicinity (300 bp up- and down- stream of TSS and TTS of tDNA) of Pol II ORF were filtered out to avoid the influence of neighboring Pol II genes on the average profiles. All the high-throughput data retrieved from different sources are listed in the Supplemental Table S4.

ChIP-chip data analysis

Earlier published Spt16 occupancy ChIP-chip data on ORFs normalized with the mock IP and input in wild-type cells (Mayer et al. 2010) were downloaded from NCBI GEO. The open source “Bio toolbox” from Timothy J. Parnell was used for mapping the binding sites around TSS of Pol II ORFs and Spt16 occupancy around the TSS was plotted.

Statistical analysis

Each measurement was repeated several times and the average with scatter for a minimum of three biological replicates was taken for plotting the graphs. Statistical significance of a difference between the two values was calculated using unpaired Student's t-test. Wherever marked, the dot denotes a nonsignificant change (P > 0.05) while changes of increasing significance are marked with increasing number of asterisks. One-way ANOVA was used to analyze the differences among the group means and P-values for Figure 7A,B were calculated.

DATA DEPOSITION

The Spt16 ChIP-seq data sets from this study have been deposited in the NCBI GEO database (http://www.ncbi.nlm.nih.gov/geo/) under accession number GSE83086.

SUPPLEMENTAL MATERIAL

Supplemental material is available for this article.