The Mediator complex plays a central, highly conserved role in eukaryotic transcription by stimulating the cooperative assembly of a pre-initiation complex (PIC) and recruitment of RNA Polymerase II (Pol II) for gene activation (Allen and Taatjes, 2015; Jeronimo and Robert, 2017; Kornberg, 2005; Malik and Roeder, 2010; Soutourina, 2018). In budding yeast (Saccharomyces cerevisiae), Mediator is composed of 25 subunits divided into four domains based on structural and functional criteria: the tail, middle, head, and kinase modules (Guglielmi et al., 2004; Plaschka et al., 2016; van de Peppel et al., 2005). Mediator recruitment has generally been ascribed to sequence-specific activators engaging the tail module triad of Med2-Med3-Med15 (Lee et al., 1999; Myers et al., 1999; Zhang et al., 2004; Ansari and Morse, 2013); consistent with this model, the tail module can under some circumstances be recruited independently of the remainder of Mediator (Ansari et al., 2009; He et al., 2008; Paul et al., 2015; Petrenko et al., 2016; Zhang et al., 2004; Anandhakumar et al., 2016). Normally, however, the tail is recruited as a complex with the head, middle, and kinase modules at upstream activating sequences (UASs) (Jeronimo and Robert, 2014; Wong et al., 2014; Grünberg et al., 2016; Kuras et al., 2003). Once Mediator is recruited, the kinase module dissociates from the complex and Mediator bridges the enhancer to connect with Pol II and general transcription factors (GTFs) at the proximal promoter of active genes (Jeronimo et al., 2016; Petrenko et al., 2016). Mediator association with the PIC at proximal promoters is quickly disrupted following phosphorylation of the Pol II C-terminal domain (CTD) by Kin28, marking the end of initiation (Jeronimo and Robert, 2014; Wong et al., 2014).

This orchestrated recruitment of Mediator is broadly required for transcription, evidenced by studies that showed a loss of Mediator function results in greatly decreased mRNA levels and concomitant diminished association of Pol II with essentially all Pol II transcribed genes (Ansari et al., 2009; Holstege et al., 1998; Jeronimo and Robert, 2014; Paul et al., 2015; Plaschka et al., 2015; Thompson and Young, 1995). In spite of the purported role of the tail module triad in Mediator recruitment by transcriptional activators, med3Δ med15Δ yeast are viable, show defects in mRNA levels for only 5–10% of active genes during normal growth, and exhibit continued Mediator and Pol II occupancy by ChIP-seq (Ansari et al., 2012a; Jeronimo et al., 2016; Paul et al., 2015; Petrenko et al., 2016). Additional studies report that the tail module is not required for Mediator’s interaction with the PIC (Jeronimo et al., 2016; Petrenko et al., 2016; Plaschka et al., 2015). That transcription is unaffected at many genes in med3Δ med15Δ yeast supports the notion that Mediator is recruited through an unknown mechanism that is independent of the tail module, possibly through components of the PIC which are known to engage Mediator subunits at proximal promoters (Ansari and Morse, 2013; Esnault et al., 2008; Soutourina et al., 2011; Eychenne et al., 2016; Larivière et al., 2006). Indirect recruitment of Mediator via interactions with the general transcription machinery has also been proposed on the basis of in vitro experiments (Johnson and Carey, 2003). However, because Med2 is still present in med3∆ med15∆ mutants, the possibility that recruitment occurs via this remaining tail module triad subunit cannot formally be excluded.

Several studies show that a loss of Mediator leads to a decrease in PIC components and diminished Pol II presence at active genes, highlighting the unilateral importance of Mediator in stimulating the assembly of a PIC (Eyboulet et al., 2015; He et al., 2008; Holstege et al., 1998; Paul et al., 2015; Thompson and Young, 1995; Kuras and Struhl, 1999; Li et al., 1999). Yet, how the PIC itself influences Mediator recruitment has received little attention (Eychenne et al., 2016; Robinson et al., 2016). The close spatial relationship of Mediator with the PIC at proximal promoters has been shown in single particle cryo-electron microscopy and cross-linking mass spectrometry studies of yeast in which Mediator is bound with the PIC (Nozawa et al., 2017; Plaschka et al., 2015; Robinson et al., 2016; Tsai et al., 2017). The TFIID-TBP complex and Pol II in particular occupy proximal promoters and have the potential to engage Mediator directly. In vivo, however, Mediator’s interaction with the PIC at proximal promoters is likely brief in light of recent evidence demonstrating that Mediator association at proximal promoters is transient, making interactions with the PIC difficult to assess in living cells (Fan and Struhl, 2009; Jeronimo and Robert, 2014; Wong et al., 2014).

To investigate PIC-dependent Mediator recruitment in vivo we have performed ChIP-seq against Mediator subunits in Saccharomyces cerevisiae after employing the anchor away method of nuclear depletion to export individual components of the PIC to the cytoplasm (Haruki et al., 2008). We confirm Mediator stabilization at proximal promoters in the kin28-AA strain. In the same background, we show that in med2Δ med3Δ med15Δ yeast, which completely lack the tail module triad, association of Mediator and Pol II is greatly reduced but nonetheless still detected at the majority of targets. We also investigate Mediator association by ChIP-seq after depletion of PIC components TBP, Rpb3, and Taf1, both alone and together with Kin28, to assess the contribution of the PIC to Mediator occupancy. Taken together, our findings provide new insights into mechanisms of Mediator recruitment during gene activation in yeast.

The major findings reported in this work are that 1) the Med2-Med3-Med15 tail module triad plays a major role in recruitment of Mediator and Pol II, but is dispensable for viability; 2) Mediator association with active gene promoters is observed even in the complete absence of the tail module triad; 3) PIC components, in particular Taf1 and Pol II (Rpb3) contribute to Mediator association with gene promoters, especially at TFIID-dominated genes, and 4) TBP plays a critical role in the transit of Mediator from UAS to promoter sites.

Previous studies reported altered transcript levels for only 5–10% of all genes in mutants lacking one or two tail module triad subunits, suggesting that this moiety might serve as a requisite target for transcriptional activators only for a limited fraction of the genome (Ansari et al., 2012a; van de Peppel et al., 2005). Pol II occupancy was only moderately reduced in med3∆ med15∆ yeast, particularly at TFIID-dominated genes, and Mediator occupancy at gene promoters under conditions of Kin28 inactivation was reduced by less than two-fold in this same mutant, further suggesting a limited role for the tail module triad in recruitment of Mediator and Pol II (Jeronimo et al., 2016; Paul et al., 2015). In the present work, however, we find a more substantial reduction of occupancy by Mediator and Pol II in med2∆ med3∆ med15∆ yeast, indicating a widespread role for the tail module triad in recruitment of the Mediator complex and functional redundancy in the tail module subunits, reminiscent of previous observations in comparing med3∆ med15∆ yeast to single deletion mutants (Ansari et al., 2012a). This redundancy could reflect ‘fuzzy’ interactions between activators and subunits of the tail module triad, such that loss of any individual tail module triad subunit would result in loss but not elimination of recruitment and consequent activation (Tuttle et al., 2018).

Recruitment of Mediator via the tail module triad is likely to occur via activators bound to UAS sites, as in the canonical model for gene activation (Figure 8A). Consistent with this model is the finding that Mediator is detected upstream of active genes by ChEC-seq, at distances consistent with occupancy of UAS sites, in both the presence and absence of functional Kin28 (Grünberg et al., 2016; Grünberg and Zentner, 2017). This argument is further supported by our findings that Mediator occupancy is observed by ChIP at upstream regions of genes when both TBP and Kin28 are depleted (Figure 5). Altogether, then, these various observations support a model in which Mediator is principally recruited first to UAS sites by activators contacting subunits of the tail module triad before transiting to the promoter and facilitating PIC formation.

Figure 8

Download asset Open asset

Although the dominant mechanism of Mediator recruitment is likely to occur via tail-module dependent interactions with UAS-bound activators, our findings that Mediator and Pol II are recruited to gene promoters even in the complete absence of the tail module triad indicate the existence of an alternative, tail-module-independent mechanism for Mediator recruitment (Figure 8B). In this second pathway, PIC components could be recruited before or together with Mediator, presumably via activator-mediated interactions, and Mediator association would be facilitated by contacts with PIC components and in turn stabilize binding of those components (Figure 8B). In both pathways, following formation of the Mediator-PIC complex, TFIIH is recruited and its Kin28 subunit phosphorylates the carboxyl-terminal domain of the Rpb1 subunit of Pol II; this allows escape of Pol II into the elongation complex, and results in rapid loss of Mediator association (Jeronimo and Robert, 2014; Wong et al., 2014). Direct recruitment of Mediator to promoter regions via interactions with PIC components may be favored at TFIID-dominated genes relative to SAGA-dominated genes; this would be consistent with the greater effect on Mediator occupancy seen at the former compared to the latter class upon depletion of PIC components (Figure 5).

Our results indicate that PIC components are important for transit of Mediator from UAS to promoter, and for association with promoters following that transit. PIC components are not necessary for recruitment of Mediator to UAS sites, as their depletion does not lead to reduction in Mediator association (Figure 3). Depletion of Taf1 results in greatly diminished Mediator occupancy at TFIID-dominated promoters, and little effect at SAGA-dominated promoters, whereas depletion of TBP causes a shift in Mediator occupancy to UAS regions at both promoter categories, along with diminished occupancy at TFIID-dominated genes (Figure 5). Thus, Taf1, and presumably TFIID, is important for stabilizing Mediator occupancy at promoters of TFIID-dominated genes following its transit from the initial site of recruitment, whereas TBP is critical for the transit from UAS to promoter of both SAGA- and TFIID-dominated genes (Figure 8, right). Depletion of Rpb3 together with Kin28 causes a decrease in Mediator occupancy at promoters, but also results in a new, upstream peak of Mediator occupancy that is much lower in intensity than that seen upon depletion of TBP and Kin28 (Figure 5C–D). We interpret this as a combined effect of Pol II stabilizing Mediator at promoter regions, together with Pol II depletion causing a partial decrease in TBP occupancy, which causes Mediator occupancy to shift upstream in a fraction of the cells being sampled (Figure 3—figure supplement 3D) (Joo et al., 2017).

Related to our findings that PIC components facilitate Mediator association with active gene promoters, in vitro results have indicated cooperative interactions among Mediator, TFIID, and Pol II (Johnson and Carey, 2003; Johnson et al., 2002; Baek et al., 2002; Takahashi et al., 2011). In vivo evidence for such cooperative interactions, however, has been sparse. An ‘activator bypass’ experiment showed that direct recruitment of TFIIB, via tethering to the DNA-binding domain of an activator, resulted in recruitment of a functional PIC and the Mediator complex in yeast, pointing to ‘backwards’ recruitment of Mediator via PIC interactions (Lacombe et al., 2013). More recently, depletion of Taf1 was shown to result in a widespread reduction in association of head module subunit Med8 with active genes when mapped by ChEC-seq (Grünberg et al., 2016; Grünberg and Zentner, 2017). Neither of these in vivo studies, however, examined PIC-Mediator cooperativity at promoter regions, and the distinct effects of depleting different PIC components on Mediator association (Figure 5) could not have been predicted on the basis of previous work.

Several interactions between Mediator and TFIID components, and Mediator and Pol II, have been documented that could contribute to Mediator association with promoters. Structural studies have shown the head module subcomplexes Med8-Med18-Med20 and Med11-Med17-Med22 bind with TBP (Cai et al., 2010; Larivière et al., 2006). Disruption of the Med8-Med18-Med20 subcomplex in med18Δ or med20Δ yeast does not cause loss of viability, suggesting that multiple Mediator-TBP interactions are possible at the proximal promoter for initiation (Ranish et al., 1999). In addition, interaction between the tail subunits Med15 or Med16 with TFIID subunits Taf14 or Taf1 have been observed in yeast two-hybrid and mass spectrometry experiments (Gavin et al., 2006; Gavin et al., 2002; Lim et al., 2007). Interactions between Mediator and Pol II have also been demonstrated in both structural and genetic studies (Davis et al., 2002; Esnault et al., 2008; Plaschka et al., 2015; Soutourina et al., 2011; Cai et al., 2010; Robinson et al., 2015).

Interactions between Mediator and Pol II or TFIID appear more important for stabilizing Mediator occupancy at TFIID-dominated genes than for SAGA-dominated genes, based on the stronger effect on Mediator occupancy in med2∆ med3∆ med15∆ yeast at the latter class than the former (Figure 1), and the stronger effect seen at TFIID-dominated genes upon depletion of Taf1 or Rpb3 (Figure 5). The distinction of these two pathways in terms of dependence on PIC components is also suggested by recent work showing that disruption of a Mediator-TFIIB connection via med10-ts yeast differently reduces association of Mediator, TFIIB, and GTFs (Eychenne et al., 2016). TFIID-dominated genes exhibit a greater reduction in Mediator rather than GTFs, while SAGA-dominated genes exhibit a greater reduction in GTFs rather than Mediator (Eychenne et al., 2016). Furthermore, two recent studies report that Mediator in general, and not just tail module subunits, contributes more strongly to Pol II recruitment at SAGA-dominated than TFIID-dominated genes (Bruzzone et al., 2018; Petrenko et al., 2017).

What is the basis for the differential dependence of SAGA-dominated and TFIID-dominated genes on Mediator? The average distance between UAS and TSS is greater for TATA-containing than TATA-less genes (Kristiansson et al., 2009; Erb and van Nimwegen, 2011). UASs show diminished ability to activate transcription from a reporter gene as the distance between UAS and TATA box increases (Dobi and Winston, 2007), and Mediator mutations were found to affect the dependence of UAS activity on distance (Reavey et al., 2015). Thus, the greater UAS-TSS distance of TATA-containing gene promoters may cause a greater dependency on Mediator for activation.

In summary, the work here provides new insights into the role of the Mediator tail module triad and PIC components in Mediator recruitment and dynamics. Future work will be aimed at a fuller understanding of the mechanism of tail module-independent recruitment of Mediator, and of the differential dependence of SAGA-dominated and TFIID-dominated genes on PIC components for stable Mediator association at gene promoters.

Data from ChIP-seq and RNA-seq experiments have been deposited at the NCBI Short Read Archive under project number PRJNA413080.

1. 1. Knoll ER
   2. Zhu ZI
   3. Sarkar D
   4. Landsman D
   5. Morse RH

   (2018) NCBI BioProject

   ID PRJNA413080. Data from ChIP-seq and RNA-seq experiments.

   https://www.ncbi.nlm.nih.gov/bioproject/PRJNA413080/

1. 1. Allen BL
   2. Taatjes DJ

   (2015) The Mediator complex: a central integrator of transcription

   Nature Reviews Molecular Cell Biology 16:155–166.

   https://doi.org/10.1038/nrm3951

   • PubMed
   • Google Scholar
2. 1. Anandhakumar J
   2. Moustafa YW
   3. Chowdhary S
   4. Kainth AS
   5. Gross DS

   (2016) Evidence for multiple mediator complexes in yeast independently recruited by activated heat shock factor

   Molecular and Cellular Biology 36:1943–1960.

   https://doi.org/10.1128/MCB.00005-16

   • PubMed
   • Google Scholar
3. 1. Ansari SA
   2. He Q
   3. Morse RH

   (2009) Mediator complex association with constitutively transcribed genes in yeast

   PNAS 106:16734–16739.

   https://doi.org/10.1073/pnas.0905103106

   • PubMed
   • Google Scholar
4. 1. Ansari SA
   2. Ganapathi M
   3. Benschop JJ
   4. Holstege FC
   5. Wade JT
   6. Morse RH

   (2012a) Distinct role of Mediator tail module in regulation of SAGA-dependent, TATA-containing genes in yeast

   The EMBO Journal 31:44–57.

   https://doi.org/10.1038/emboj.2011.362

   • PubMed
   • Google Scholar
5. 1. Ansari SA
   2. Morse RH

   (2012b) Selective role of Mediator tail module in the transcription of highly regulated genes in yeast

   Transcription 3:110–114.

   https://doi.org/10.4161/trns.19840

   • PubMed
   • Google Scholar
6. 1. Ansari SA
   2. Morse RH

   (2013) Mechanisms of Mediator complex action in transcriptional activation

   Cellular and Molecular Life Sciences 70:2743–2756.

   https://doi.org/10.1007/s00018-013-1265-9

   • PubMed
   • Google Scholar
7. 1. Ansari SA
   2. Paul E
   3. Sommer S
   4. Lieleg C
   5. He Q
   6. Daly AZ
   7. Rode KA
   8. Barber WT
   9. Ellis LC
   10. LaPorta E
   11. Orzechowski AM
   12. Taylor E
   13. Reeb T
   14. Wong J
   15. Korber P
   16. Morse RH

   (2014) Mediator, TATA-binding protein, and RNA polymerase II contribute to low histone occupancy at active gene promoters in yeast

   Journal of Biological Chemistry 289:14981–14995.

   https://doi.org/10.1074/jbc.M113.529354

   • PubMed
   • Google Scholar
8. 1. Baek HJ
   2. Malik S
   3. Qin J
   4. Roeder RG

   (2002) Requirement of TRAP/mediator for both activator-independent and activator-dependent transcription in conjunction with TFIID-associated TAF(II)s

   Molecular and Cellular Biology 22:2842–2852.

   https://doi.org/10.1128/MCB.22.8.2842-2852.2002

   • PubMed
   • Google Scholar
9. 1. Baptista T
   2. Grünberg S
   3. Minoungou N
   4. Koster MJE
   5. Timmers HTM
   6. Hahn S
   7. Devys D
   8. Tora L

   (2017) SAGA Is a general cofactor for RNA Polymerase II Transcription

   Molecular Cell 68:130–143.

   https://doi.org/10.1016/j.molcel.2017.08.016

   • PubMed
   • Google Scholar
10. 1. Bhaumik SR

    (2011) Distinct regulatory mechanisms of eukaryotic transcriptional activation by SAGA and TFIID

    Biochimica Et Biophysica Acta (BBA) - Gene Regulatory Mechanisms 1809:97–108.

    https://doi.org/10.1016/j.bbagrm.2010.08.009

    • PubMed
    • Google Scholar
11. 1. Bourbon HM
    2. Aguilera A
    3. Ansari AZ
    4. Asturias FJ
    5. Berk AJ
    6. Bjorklund S
    7. Blackwell TK
    8. Borggrefe T
    9. Carey M
    10. Carlson M
    11. Conaway JW
    12. Conaway RC
    13. Emmons SW
    14. Fondell JD
    15. Freedman LP
    16. Fukasawa T
    17. Gustafsson CM
    18. Han M
    19. He X
    20. Herman PK
    21. Hinnebusch AG
    22. Holmberg S
    23. Holstege FC
    24. Jaehning JA
    25. Kim YJ
    26. Kuras L
    27. Leutz A
    28. Lis JT
    29. Meisterernest M
    30. Naar AM
    31. Nasmyth K
    32. Parvin JD
    33. Ptashne M
    34. Reinberg D
    35. Ronne H
    36. Sadowski I
    37. Sakurai H
    38. Sipiczki M
    39. Sternberg PW
    40. Stillman DJ
    41. Strich R
    42. Struhl K
    43. Svejstrup JQ
    44. Tuck S
    45. Winston F
    46. Roeder RG
    47. Kornberg RD

    (2004) A unified nomenclature for protein subunits of mediator complexes linking transcriptional regulators to RNA polymerase II

    Molecular Cell 14:553–557.

    https://doi.org/10.1016/j.molcel.2004.05.011

    • PubMed
    • Google Scholar
12. 1. Bruzzone MJ
    2. Grünberg S
    3. Kubik S
    4. Zentner GE
    5. Shore D

    (2018) Distinct patterns of histone acetyltransferase and Mediator deployment at yeast protein-coding genes

    Genes & Development 32:1252–1265.

    https://doi.org/10.1101/gad.312173.118

    • PubMed
    • Google Scholar
13. Book

    1. Burke D
    2. Dawson D
    3. Stearns T

    (2000)

    Methods in Yeast Genetics: A Cold Spring Harbor Laboratory Course Manual

    New York: Cold Spring Harbor Press.

    • Google Scholar
14. 1. Cai G
    2. Imasaki T
    3. Yamada K
    4. Cardelli F
    5. Takagi Y
    6. Asturias FJ

    (2010) Mediator head module structure and functional interactions

    Nature Structural & Molecular Biology 17:273–279.

    https://doi.org/10.1038/nsmb.1757

    • PubMed
    • Google Scholar
15. 1. Davis JA
    2. Takagi Y
    3. Kornberg RD
    4. Asturias FA

    (2002) Structure of the yeast RNA polymerase II holoenzyme: Mediator conformation and polymerase interaction

    Molecular Cell 10:409–415.

    https://doi.org/10.1016/S1097-2765(02)00598-1

    • PubMed
    • Google Scholar
16. 1. DeGennaro CM
    2. Alver BH
    3. Marguerat S
    4. Stepanova E
    5. Davis CP
    6. Bähler J
    7. Park PJ
    8. Winston F

    (2013) Spt6 regulates intragenic and antisense transcription, nucleosome positioning, and histone modifications genome-wide in fission yeast

    Molecular and Cellular Biology 33:4779–4792.

    https://doi.org/10.1128/MCB.01068-13

    • PubMed
    • Google Scholar
17. 1. Dobi KC
    2. Winston F

    (2007) Analysis of transcriptional activation at a distance in saccharomyces cerevisiae

    Molecular and Cellular Biology 27:5575–5586.

    https://doi.org/10.1128/MCB.00459-07

    • PubMed
    • Google Scholar
18. 1. Erb I
    2. van Nimwegen E

    (2011) Transcription factor binding site positioning in yeast: proximal promoter motifs characterize TATA-less promoters

    PLOS ONE 6:e24279.

    https://doi.org/10.1371/journal.pone.0024279

    • PubMed
    • Google Scholar
19. 1. Esnault C
    2. Ghavi-Helm Y
    3. Brun S
    4. Soutourina J
    5. Van Berkum N
    6. Boschiero C
    7. Holstege F
    8. Werner M

    (2008) Mediator-dependent recruitment of TFIIH modules in preinitiation complex

    Molecular Cell 31:337–346.

    https://doi.org/10.1016/j.molcel.2008.06.021

    • PubMed
    • Google Scholar
20. 1. Eyboulet F
    2. Wydau-Dematteis S
    3. Eychenne T
    4. Alibert O
    5. Neil H
    6. Boschiero C
    7. Nevers MC
    8. Volland H
    9. Cornu D
    10. Redeker V
    11. Werner M
    12. Soutourina J

    (2015) Mediator independently orchestrates multiple steps of preinitiation complex assembly in vivo

    Nucleic Acids Research 43:9214–9231.

    https://doi.org/10.1093/nar/gkv782

    • PubMed
    • Google Scholar
21. 1. Eychenne T
    2. Novikova E
    3. Barrault MB
    4. Alibert O
    5. Boschiero C
    6. Peixeiro N
    7. Cornu D
    8. Redeker V
    9. Kuras L
    10. Nicolas P
    11. Werner M
    12. Soutourina J

    (2016) Functional interplay between Mediator and TFIIB in preinitiation complex assembly in relation to promoter architecture

    Genes & Development 30:2119–2132.

    https://doi.org/10.1101/gad.285775.116

    • PubMed
    • Google Scholar
22. 1. Fan X
    2. Chou DM
    3. Struhl K

    (2006) Activator-specific recruitment of Mediator in vivo

    Nature Structural & Molecular Biology 13:117–120.

    https://doi.org/10.1038/nsmb1049

    • PubMed
    • Google Scholar
23. 1. Fan X
    2. Struhl K

    (2009) Where does mediator bind in vivo?

    PLOS One 4:e5029.

    https://doi.org/10.1371/journal.pone.0005029

    • PubMed
    • Google Scholar
24. 1. Gavin AC
    2. Bösche M
    3. Krause R
    4. Grandi P
    5. Marzioch M
    6. Bauer A
    7. Schultz J
    8. Rick JM
    9. Michon AM
    10. Cruciat CM
    11. Remor M
    12. Höfert C
    13. Schelder M
    14. Brajenovic M
    15. Ruffner H
    16. Merino A
    17. Klein K
    18. Hudak M
    19. Dickson D
    20. Rudi T
    21. Gnau V
    22. Bauch A
    23. Bastuck S
    24. Huhse B
    25. Leutwein C
    26. Heurtier MA
    27. Copley RR
    28. Edelmann A
    29. Querfurth E
    30. Rybin V
    31. Drewes G
    32. Raida M
    33. Bouwmeester T
    34. Bork P
    35. Seraphin B
    36. Kuster B
    37. Neubauer G
    38. Superti-Furga G

    (2002) Functional organization of the yeast proteome by systematic analysis of protein complexes

    Nature 415:141–147.

    https://doi.org/10.1038/415141a

    • PubMed
    • Google Scholar
25. 1. Gavin AC
    2. Aloy P
    3. Grandi P
    4. Krause R
    5. Boesche M
    6. Marzioch M
    7. Rau C
    8. Jensen LJ
    9. Bastuck S
    10. Dümpelfeld B
    11. Edelmann A
    12. Heurtier MA
    13. Hoffman V
    14. Hoefert C
    15. Klein K
    16. Hudak M
    17. Michon AM
    18. Schelder M
    19. Schirle M
    20. Remor M
    21. Rudi T
    22. Hooper S
    23. Bauer A
    24. Bouwmeester T
    25. Casari G
    26. Drewes G
    27. Neubauer G
    28. Rick JM
    29. Kuster B
    30. Bork P
    31. Russell RB
    32. Superti-Furga G

    (2006) Proteome survey reveals modularity of the yeast cell machinery

    Nature 440:631–636.

    https://doi.org/10.1038/nature04532

    • PubMed
    • Google Scholar
26. 1. Gentleman RC
    2. Carey VJ
    3. Bates DM
    4. Bolstad B
    5. Dettling M
    6. Dudoit S
    7. Ellis B
    8. Gautier L
    9. Ge Y
    10. Gentry J
    11. Hornik K
    12. Hothorn T
    13. Huber W
    14. Iacus S
    15. Irizarry R
    16. Leisch F
    17. Li C
    18. Maechler M
    19. Rossini AJ
    20. Sawitzki G
    21. Smith C
    22. Smyth G
    23. Tierney L
    24. Yang JY
    25. Zhang J

    (2004) Bioconductor: open software development for computational biology and bioinformatics

    Genome Biology 5:R80.

    https://doi.org/10.1186/gb-2004-5-10-r80

    • PubMed
    • Google Scholar
27. 1. Grimaldi Y
    2. Ferrari P
    3. Strubin M

    (2014) Independent RNA polymerase II preinitiation complex dynamics and nucleosome turnover at promoter sites in vivo

    Genome Research 24:117–124.

    https://doi.org/10.1101/gr.157792.113

    • PubMed
    • Google Scholar
28. 1. Grünberg S
    2. Henikoff S
    3. Hahn S
    4. Zentner GE

    (2016) Mediator binding to UASs is broadly uncoupled from transcription and cooperative with TFIID recruitment to promoters

    The EMBO Journal 35:2435–2446.

    https://doi.org/10.15252/embj.201695020

    • PubMed
    • Google Scholar
29. 1. Grünberg S
    2. Zentner GE

    (2017) Genome-wide characterization of Mediator recruitment, function, and regulation

    Transcription 8:169–174.

    https://doi.org/10.1080/21541264.2017.1291082

    • PubMed
    • Google Scholar
30. 1. Guglielmi B
    2. van Berkum NL
    3. Klapholz B
    4. Bijma T
    5. Boube M
    6. Boschiero C
    7. Bourbon HM
    8. Holstege FC
    9. Werner M

    (2004) A high resolution protein interaction map of the yeast Mediator complex

    Nucleic Acids Research 32:5379–5391.

    https://doi.org/10.1093/nar/gkh878

    • PubMed
    • Google Scholar
31. 1. Haruki H
    2. Nishikawa J
    3. Laemmli UK

    (2008) The anchor-away technique: rapid, conditional establishment of yeast mutant phenotypes

    Molecular cell 31:925–932.

    https://doi.org/10.1016/j.molcel.2008.07.020

    • PubMed
    • Google Scholar
32. 1. He Q
    2. Battistella L
    3. Morse RH

    (2008) Mediator requirement downstream of chromatin remodeling during transcriptional activation of CHA1 in yeast

    Journal of Biological Chemistry 283:5276–5286.

    https://doi.org/10.1074/jbc.M708266200

    • PubMed
    • Google Scholar
33. 1. Hill J
    2. Donald KA
    3. Griffiths DE
    4. Donald G

    (1991) DMSO-enhanced whole cell yeast transformation

    Nucleic Acids Research 19:5791.

    https://doi.org/10.1093/nar/19.20.5791

    • PubMed
    • Google Scholar
34. 1. Holstege FC
    2. Jennings EG
    3. Wyrick JJ
    4. Lee TI
    5. Hengartner CJ
    6. Green MR
    7. Golub TR
    8. Lander ES
    9. Young RA

    (1998) Dissecting the regulatory circuitry of a eukaryotic genome

    Cell 95:717–728.

    https://doi.org/10.1016/S0092-8674(00)81641-4

    • PubMed
    • Google Scholar
35. 1. Huisinga KL
    2. Pugh BF

    (2004) A genome-wide housekeeping role for TFIID and a highly regulated stress-related role for SAGA in Saccharomyces cerevisiae

    Molecular Cell 13:573–585.

    https://doi.org/10.1016/S1097-2765(04)00087-5

    • PubMed
    • Google Scholar
36. 1. Jeronimo C
    2. Langelier MF
    3. Bataille AR
    4. Pascal JM
    5. Pugh BF
    6. Robert F

    (2016) Tail and kinase modules differently regulate core mediator recruitment and function In Vivo

    Molecular Cell 64:455–466.

    https://doi.org/10.1016/j.molcel.2016.09.002

    • PubMed
    • Google Scholar
37. 1. Jeronimo C
    2. Robert F

    (2014) Kin28 regulates the transient association of Mediator with core promoters

    Nature structural & molecular biology 21:449–455.

    https://doi.org/10.1038/nsmb.2810

    • PubMed
    • Google Scholar
38. 1. Jeronimo C
    2. Robert F

    (2017) The Mediator Complex: At the Nexus of RNA Polymerase II Transcription

    Trends in Cell Biology 27:765–783.

    https://doi.org/10.1016/j.tcb.2017.07.001

    • PubMed
    • Google Scholar
39. 1. Johnson KM
    2. Wang J
    3. Smallwood A
    4. Arayata C
    5. Carey M

    (2002) TFIID and human mediator coactivator complexes assemble cooperatively on promoter DNA

    Genes & Development 16:1852–1863.

    https://doi.org/10.1101/gad.995702

    • PubMed
    • Google Scholar
40. 1. Johnson KM
    2. Carey M

    (2003) Assembly of a mediator/TFIID/TFIIA complex bypasses the need for an activator

    Current Biology 13:772–777.

    https://doi.org/10.1016/S0960-9822(03)00283-5

    • PubMed
    • Google Scholar
41. 1. Joo YJ
    2. Ficarro SB
    3. Soares LM
    4. Chun Y
    5. Marto JA
    6. Buratowski S

    (2017) Downstream promoter interactions of TFIID TAFs facilitate transcription reinitiation

    Genes & Development 31:2162–2174.

    https://doi.org/10.1101/gad.306324.117

    • PubMed
    • Google Scholar
42. 1. Kim D
    2. Pertea G
    3. Trapnell C
    4. Pimentel H
    5. Kelley R
    6. Salzberg SL

    (2013) TopHat2: accurate alignment of transcriptomes in the presence of insertions, deletions and gene fusions

    Genome Biology 14:R36.

    https://doi.org/10.1186/gb-2013-14-4-r36

    • PubMed
    • Google Scholar
43. 1. Kornberg RD

    (2005) Mediator and the mechanism of transcriptional activation

    Trends in Biochemical Sciences 30:235–239.

    https://doi.org/10.1016/j.tibs.2005.03.011

    • PubMed
    • Google Scholar
44. 1. Kristiansson E
    2. Thorsen M
    3. Tamás MJ
    4. Nerman O

    (2009) Evolutionary forces act on promoter length: identification of enriched cis-regulatory elements

    Molecular Biology and Evolution 26:1299–1307.

    https://doi.org/10.1093/molbev/msp040

    • PubMed
    • Google Scholar
45. 1. Kuras L
    2. Borggrefe T
    3. Kornberg RD

    (2003) Association of the Mediator complex with enhancers of active genes

    PNAS 100:13887–13891.

    https://doi.org/10.1073/pnas.2036346100

    • Google Scholar
46. 1. Kuras L
    2. Struhl K

    (1999) Binding of TBP to promoters in vivo is stimulated by activators and requires Pol II holoenzyme

    Nature 399:609–613.

    https://doi.org/10.1038/21239

    • PubMed
    • Google Scholar
47. 1. Lacombe T
    2. Poh SL
    3. Barbey R
    4. Kuras L

    (2013) Mediator is an intrinsic component of the basal RNA polymerase II machinery in vivo

    Nucleic Acids Research 41:9651–9662.

    https://doi.org/10.1093/nar/gkt701

    • PubMed
    • Google Scholar
48. 1. Larivière L
    2. Geiger S
    3. Hoeppner S
    4. Röther S
    5. Strässer K
    6. Cramer P

    (2006) Structure and TBP binding of the Mediator head subcomplex Med8-Med18-Med20

    Nature Structural & Molecular Biology 13:895–901.

    https://doi.org/10.1038/nsmb1143

    • PubMed
    • Google Scholar
49. 1. Lee YC
    2. Park JM
    3. Min S
    4. Han SJ
    5. Kim YJ

    (1999) An activator binding module of yeast RNA polymerase II holoenzyme

    Molecular and Cellular Biology 19:2967–2976.

    https://doi.org/10.1128/MCB.19.4.2967

    • PubMed
    • Google Scholar
50. 1. Lee J
    2. Choi ES
    3. Seo HD
    4. Kang K
    5. Gilmore JM
    6. Florens L
    7. Washburn MP
    8. Choe J
    9. Workman JL
    10. Lee D

    (2017) Chromatin remodeller Fun30^Fft3 induces nucleosome disassembly to facilitate RNA polymerase II elongation

    Nature Communications 8:14527.

    https://doi.org/10.1038/ncomms14527

    • PubMed
    • Google Scholar
51. 1. Li XY
    2. Virbasius A
    3. Zhu X
    4. Green MR

    (1999) Enhancement of TBP binding by activators and general transcription factors

    Nature 399:605–609.

    https://doi.org/10.1038/21232

    • PubMed
    • Google Scholar
52. 1. Lim MK
    2. Tang V
    3. Le Saux A
    4. Schüller J
    5. Bongards C
    6. Lehming N

    (2007) Gal11p dosage-compensates transcriptional activator deletions via Taf14p

    Journal of Molecular Biology 374:9–23.

    https://doi.org/10.1016/j.jmb.2007.09.013

    • PubMed
    • Google Scholar
53. 1. Longtine MS
    2. McKenzie A
    3. Demarini DJ
    4. Shah NG
    5. Wach A
    6. Brachat A
    7. Philippsen P
    8. Pringle JR

    (1998) Additional modules for versatile and economical PCR-based gene deletion and modification in Saccharomyces cerevisiae

    Yeast 14:953–961.

    https://doi.org/10.1002/(SICI)1097-0061(199807)14:10<953::AID-YEA293>3.0.CO;2-U

    • PubMed
    • Google Scholar
54. 1. Malik S
    2. Roeder RG

    (2010) The metazoan Mediator co-activator complex as an integrative hub for transcriptional regulation

    Nature Reviews Genetics 11:761–772.

    https://doi.org/10.1038/nrg2901

    • PubMed
    • Google Scholar
55. 1. Myers LC
    2. Gustafsson CM
    3. Hayashibara KC
    4. Brown PO
    5. Kornberg RD

    (1999) Mediator protein mutations that selectively abolish activated transcription

    PNAS 96:67–72.

    https://doi.org/10.1073/pnas.96.1.67

    • PubMed
    • Google Scholar
56. 1. Nagalakshmi U
    2. Wang Z
    3. Waern K
    4. Shou C
    5. Raha D
    6. Gerstein M
    7. Snyder M

    (2008) The transcriptional landscape of the yeast genome defined by RNA sequencing

    Science 320:1344–1349.

    https://doi.org/10.1126/science.1158441

    • PubMed
    • Google Scholar
57. 1. Nozawa K
    2. Schneider TR
    3. Cramer P

    (2017) Core Mediator structure at 3.4 Å extends model of transcription initiation complex

    Nature 545:248–251.

    https://doi.org/10.1038/nature22328

    • PubMed
    • Google Scholar
58. 1. Orlando DA
    2. Chen MW
    3. Brown VE
    4. Solanki S
    5. Choi YJ
    6. Olson ER
    7. Fritz CC
    8. Bradner JE
    9. Guenther MG

    (2014) Quantitative ChIP-Seq normalization reveals global modulation of the epigenome

    Cell Reports 9:1163–1170.

    https://doi.org/10.1016/j.celrep.2014.10.018

    • PubMed
    • Google Scholar
59. 1. Paul E
    2. Zhu ZI
    3. Landsman D
    4. Morse RH

    (2015) Genome-wide association of mediator and RNA polymerase II in wild-type and mediator mutant yeast

    Molecular and Cellular Biology 35:331–342.

    https://doi.org/10.1128/MCB.00991-14

    • PubMed
    • Google Scholar
60. 1. Petrenko N
    2. Jin Y
    3. Wong KH
    4. Struhl K

    (2016) Mediator Undergoes a Compositional Change during Transcriptional Activation

    Molecular Cell 64:443–454.

    https://doi.org/10.1016/j.molcel.2016.09.015

    • PubMed
    • Google Scholar
61. 1. Petrenko N
    2. Jin Y
    3. Wong KH
    4. Struhl K

    (2017) Evidence that Mediator is essential for Pol II transcription, but is not a required component of the preinitiation complex in vivo

    eLife 6:e28447.

    https://doi.org/10.7554/eLife.28447

    • PubMed
    • Google Scholar
62. 1. Plaschka C
    2. Larivière L
    3. Wenzeck L
    4. Seizl M
    5. Hemann M
    6. Tegunov D
    7. Petrotchenko EV
    8. Borchers CH
    9. Baumeister W
    10. Herzog F
    11. Villa E
    12. Cramer P

    (2015) Architecture of the RNA polymerase II-Mediator core initiation complex

    Nature 518:376–380.

    https://doi.org/10.1038/nature14229

    • PubMed
    • Google Scholar
63. 1. Plaschka C
    2. Nozawa K
    3. Cramer P

    (2016) Mediator Architecture and RNA Polymerase II Interaction

    Journal of Molecular Biology 428:2569–2574.

    https://doi.org/10.1016/j.jmb.2016.01.028

    • PubMed
    • Google Scholar
64. 1. Ranish JA
    2. Yudkovsky N
    3. Hahn S

    (1999) Intermediates in formation and activity of the RNA polymerase II preinitiation complex: holoenzyme recruitment and a postrecruitment role for the TATA box and TFIIB

    Genes & Development 13:49–63.

    https://doi.org/10.1101/gad.13.1.49

    • PubMed
    • Google Scholar
65. 1. Reavey CT
    2. Hickman MJ
    3. Dobi KC
    4. Botstein D
    5. Winston F

    (2015) Analysis of polygenic mutants suggests a role for mediator in regulating transcriptional activation distance in saccharomyces cerevisiae

    Genetics 201:599–612.

    https://doi.org/10.1534/genetics.115.181164

    • PubMed
    • Google Scholar
66. 1. Robinson JT
    2. Thorvaldsdóttir H
    3. Winckler W
    4. Guttman M
    5. Lander ES
    6. Getz G
    7. Mesirov JP

    (2011) Integrative genomics viewer

    Nature Biotechnology 29:24–26.

    https://doi.org/10.1038/nbt.1754

    • PubMed
    • Google Scholar
67. 1. Robinson PJ
    2. Trnka MJ
    3. Pellarin R
    4. Greenberg CH
    5. Bushnell DA
    6. Davis R
    7. Burlingame AL
    8. Sali A
    9. Kornberg RD

    (2015) Molecular architecture of the yeast mediator complex

    eLife 4:e08719.

    https://doi.org/10.7554/eLife.08719

    • PubMed
    • Google Scholar
68. 1. Robinson PJ
    2. Trnka MJ
    3. Bushnell DA
    4. Davis RE
    5. Mattei PJ
    6. Burlingame AL
    7. Kornberg RD

    (2016) Structure of a Complete Mediator-RNA Polymerase II Pre-Initiation Complex

    Cell 166:1411–1422.

    https://doi.org/10.1016/j.cell.2016.08.050

    • PubMed
    • Google Scholar
69. 1. Schmitt ME
    2. Brown TA
    3. Trumpower BL

    (1990) A rapid and simple method for preparation of RNA from Saccharomyces cerevisiae

    Nucleic Acids Research 18:3091–3092.

    https://doi.org/10.1093/nar/18.10.3091

    • PubMed
    • Google Scholar
70. 1. Seoighe C
    2. Wolfe KH

    (1999) Updated map of duplicated regions in the yeast genome

    Gene 238:253–261.

    https://doi.org/10.1016/S0378-1119(99)00319-4

    • PubMed
    • Google Scholar
71. 1. Sharma VM
    2. Li B
    3. Reese JC

    (2003) SWI/SNF-dependent chromatin remodeling of RNR3 requires TAF(II)s and the general transcription machinery

    Genes & Development 17:502–515.

    https://doi.org/10.1101/gad.1039503

    • PubMed
    • Google Scholar
72. 1. Shetty A
    2. Kallgren SP
    3. Demel C
    4. Maier KC
    5. Spatt D
    6. Alver BH
    7. Cramer P
    8. Park PJ
    9. Winston F

    (2017) Spt5 plays vital roles in the control of sense and antisense transcription elongation

    Molecular Cell 66:77–88.

    https://doi.org/10.1016/j.molcel.2017.02.023

    • PubMed
    • Google Scholar
73. 1. Soutourina J
    2. Wydau S
    3. Ambroise Y
    4. Boschiero C
    5. Werner M

    (2011) Direct interaction of RNA polymerase II and mediator required for transcription in vivo

    Science 331:1451–1454.

    https://doi.org/10.1126/science.1200188

    • PubMed
    • Google Scholar
74. 1. Soutourina J

    (2018) Transcription regulation by the mediator complex

    Nature Reviews Molecular Cell Biology 19:262–274.

    https://doi.org/10.1038/nrm.2017.115

    • PubMed
    • Google Scholar
75. 1. Sun M
    2. Schwalb B
    3. Schulz D
    4. Pirkl N
    5. Etzold S
    6. Larivière L
    7. Maier KC
    8. Seizl M
    9. Tresch A
    10. Cramer P

    (2012) Comparative dynamic transcriptome analysis (cDTA) reveals mutual feedback between mRNA synthesis and degradation

    Genome Research 22:1350–1359.

    https://doi.org/10.1101/gr.130161.111

    • PubMed
    • Google Scholar
76. 1. Takahashi H
    2. Parmely TJ
    3. Sato S
    4. Tomomori-Sato C
    5. Banks CA
    6. Kong SE
    7. Szutorisz H
    8. Swanson SK
    9. Martin-Brown S
    10. Washburn MP
    11. Florens L
    12. Seidel CW
    13. Lin C
    14. Smith ER
    15. Shilatifard A
    16. Conaway RC
    17. Conaway JW

    (2011) Human mediator subunit MED26 functions as a docking site for transcription elongation factors

    Cell 146:92–104.

    https://doi.org/10.1016/j.cell.2011.06.005

    • PubMed
    • Google Scholar
77. 1. Thompson CM
    2. Young RA

    (1995) General requirement for RNA polymerase II holoenzymes in vivo

    PNAS 92:4587–4590.

    https://doi.org/10.1073/pnas.92.10.4587

    • PubMed
    • Google Scholar
78. 1. Tramantano M
    2. Sun L
    3. Au C
    4. Labuz D
    5. Liu Z
    6. Chou M
    7. Shen C
    8. Luk E

    (2016) Constitutive turnover of histone H2A.Z at yeast promoters requires the preinitiation complex

    eLife 5:e14243.

    https://doi.org/10.7554/eLife.14243

    • PubMed
    • Google Scholar
79. 1. Tsai KL
    2. Yu X
    3. Gopalan S
    4. Chao TC
    5. Zhang Y
    6. Florens L
    7. Washburn MP
    8. Murakami K
    9. Conaway RC
    10. Conaway JW
    11. Asturias FJ

    (2017) Mediator structure and rearrangements required for holoenzyme formation

    Nature 544:196–201.

    https://doi.org/10.1038/nature21393

    • PubMed
    • Google Scholar
80. 1. Tuttle LM
    2. Pacheco D
    3. Warfield L
    4. Luo J
    5. Ranish J
    6. Hahn S
    7. Klevit RE

    (2018) Gcn4-Mediator Specificity Is Mediated by a Large and Dynamic Fuzzy Protein-Protein Complex

    Cell Reports 22:3251–3264.

    https://doi.org/10.1016/j.celrep.2018.02.097

    • PubMed
    • Google Scholar
81. 1. van de Peppel J
    2. Kettelarij N
    3. van Bakel H
    4. Kockelkorn TT
    5. van Leenen D
    6. Holstege FC

    (2005) Mediator expression profiling epistasis reveals a signal transduction pathway with antagonistic submodules and highly specific downstream targets

    Molecular Cell 19:511–522.

    https://doi.org/10.1016/j.molcel.2005.06.033

    • PubMed
    • Google Scholar
82. 1. Warfield L
    2. Ramachandran S
    3. Baptista T
    4. Devys D
    5. Tora L
    6. Hahn S

    (2017) Transcription of nearly all yeast RNA Polymerase II-Transcribed genes is dependent on transcription factor TFIID

    Molecular Cell 68:118–129.

    https://doi.org/10.1016/j.molcel.2017.08.014

    • PubMed
    • Google Scholar
83. 1. Wong KH
    2. Jin Y
    3. Struhl K

    (2014) TFIIH phosphorylation of the Pol II CTD stimulates mediator dissociation from the preinitiation complex and promoter escape

    Molecular Cell 54:601–612.

    https://doi.org/10.1016/j.molcel.2014.03.024

    • PubMed
    • Google Scholar
84. 1. Zang C
    2. Schones DE
    3. Zeng C
    4. Cui K
    5. Zhao K
    6. Peng W

    (2009) A clustering approach for identification of enriched domains from histone modification ChIP-Seq data

    Bioinformatics 25:1952–1958.

    https://doi.org/10.1093/bioinformatics/btp340

    • PubMed
    • Google Scholar
85. 1. Zhang F
    2. Sumibcay L
    3. Hinnebusch AG
    4. Swanson MJ

    (2004) A triad of subunits from the Gal11/tail domain of Srb mediator is an in vivo target of transcriptional activator Gcn4p

    Molecular and Cellular Biology 24:6871–6886.

    https://doi.org/10.1128/MCB.24.15.6871-6886.2004

    • PubMed
    • Google Scholar

[Editors’ note: a previous version of this study was rejected after peer review, but the authors submitted for reconsideration. The first decision letter after peer review is shown below.]

Thank you for submitting your work entitled "Role of the pre-initiation complex in Mediator recruitment and dynamics" for consideration by eLife. Your article has been reviewed by three peer reviewers and the evaluation has been overseen by a Reviewing Editor, Alan Hinnebusch, and a Senior Editor. The reviewers have opted to remain anonymous.

Our decision has been reached after consultation between the reviewers. Based on these discussions and the individual reviews below, we regret to inform you that your work will is not acceptable for publication in eLife.

The reviewers found intriguing the data obtained from anchor-away of TBP in the kin28-AA background, which seemed to reverse the normal transition of Mediator from UAS to promoter, whereas anchor-away of Rpb3 did not have this effect and resulted instead in loss of Mediator from the promoter. As you have concluded, these results suggest that the transition from UAS to promoter specifically requires TBP, such that Mediator appears stranded at the UAS in TBP's absence; whereas on depletion of Pol II, the TBP-mediated transfer presumably still occurs but Mediator is subsequently lost from the promoter and Mediator occupancy at both UAS and promoter is very low. Your results further suggested that Taf1 might be required along with Pol II to stabilize Mediator at the promoter; although it was felt that additional experiments are required to rule out the possibility that Taf1 is required instead for initial recruitment of Mediator to the UAS. There was also an appreciation of new findings regarding Mediator recruitment that are germane to the functional distinction between TFIID- versus SAGA-dependent genes.

Although the role of the PIC in stabilizing Mediator at promoters is strongly predicted by previous studies, uncovering distinct functions for TBP versus Pol II and Rpb3 in the UAS to promoter transition would represent novel findings. The difficulty is that other interpretations of the existing data are possible, and it was felt that significant additional experiments would be required to confirm the existing findings. One critically important question is whether the different outcomes elicted by depletion of TBP, Rpb3, or Taf1 merely reflect different quantitative degrees of PIC reduction rather than specific effects of depleting particular PIC components. It is important, therefore, to analyze the consequences of depleting each factor on the occupancies of the others, and ideally a fourth PIC component like TFIIB, to determine whether a coherent model can still be proposed in which TBP and Pol II, and Taf1 make distinct contributions to the UAS to promoter transition by Mediator.

There are also substantial differences in the data regarding the behavior of Med15 and Med8 on depletion of TBP versus Rpb3/Taf1, and there is a concern that the Med8 results are not accurately representing the behavior of Mediator head. Thus, additional head subunit(s) would have to be analyzed to confirm the Med8 findings and firmly establish whether or not different Mediator modules are affected differently by the three depletions. Finally, there were also several instances in which results were not presented comprehensively and, thus, doubts emerged about how representative the data being presented might be of the entire data set collected.

In view of the considerable amount of new experimental work that seems necessary, the extended time-frame this would likely entail, and the largely uncertain consequences of the new results for your final model, it was decided that the manuscript should be rejected rather than returned to you for revisions. The following list summarizes the most important criticisms, some of which were already mentioned above.

1) Repeat the Taf1 anchor away experiment in a WT KIN28 strain to determine whether Med15 occupancy at UASs is increased (as in Tbp3 depletion) or decreased, because if the latter was observed, this would implicate Taf1 in the initial recruitment of Mediator to promoters in addition to, or instead of, the transition of Mediator from UAS to PIC.

2) To confirm the differences seen between the Tail Med15 and Head Med18 occupancies on depletion of TBP in the kin28-AA strain in Figure 4C, examine the genome-wide behavior of additional Head subunits.

3) Compare the effects of depleting TBP, Rpb3, or Taf1 on PIC assembly to rule out the trivial possibility that the different outcomes on depletion of each factor represent different extents of reducing PIC assembly rather than indicating distinct functions in Mediator transfer or retention at the promoter. This would also address whether depletion of one factor by anchor away is simply more efficient than another.

4) Results for all four Mediator subunits, for all groups of genes, must be shown for a comprehensive analysis of the effects of depleting TBP vs. Rpb3 vs Taf1.

5) Provide a biological replicate for the results on Med2 in Figure 1A, as well as an input or untagged control.

6) Present time-courses of anchor away to justify the rapamycin treatments chosen for each protein.

Reviewer #1:

Summary of findings:

This study probes the role of PIC assembly, particularly the contributions of TBP, Taf1, and Pol II itself in the transition of Mediator from UAS to promoters in the course of the initiation of transcription in budding yeast. They show that a triple deletion of Mediator tail subunits is viable and allowed appreciable transcription with significant down-regulation of SAGA-dependent genes. In a kin28-AA strain, in which Mediator occupancy at promoters is enhanced, the triple tail deletion reduced Mediator Head occupancies at promoters and essentially abolished it at UAS elements. This finding largely confirms results/conclusions made previously from analysis of a double deletion of Med3/Med15 that the tail module strongly contributes to Mediator recruitment to the majority of genes, including those in which Mediator occupancy can be observed only on depletion of Kin28 and attendant shift of Mediator from UAS to promoter owing to loss of the CDK module, and its stabilization at the promoter owing to loss of Kin28 kinase activity.

They then show that depletion of TBP or Rpb3 actually increases WT Mediator tail subunit occupancy at UAS regions, while not affecting head subunit occupancy, for genes that show strong UAS occupancy of Mediator in WT cells, and also allows detection of Mediator at genes, e.g. RPGs, where it is not normally detected. This provides support for the model proposed by Struhl and Robert that, at genes for which Mediator association with UAS regions is observed, a single Mediator complex bridges the UAS and the promoter, with tail module subunits contacting the UAS and head and middle subunits engaging the promoter region through contacts with Pol II and GTFs; and suggests that interruption of the latter contacts by impairing PIC assembly (by depleting TBP or Rpb1) appears to strand Mediator at the UAS. These results also show that Mediator is actually present at genes (RPGs) where it was previously undetectable, as its transit from UAS to promoter is being blocked by impairing PIC assembly.

They go on to show that depleting TBP in cells where Kin28 is also depleted suppresses the effect of depleting Kin28 alone in shifting Mediator tail occupancy from UAS to PIC, providing evidence that this transition, which has been attributed to loss of the CDK module with attendant enhancement of Mediator interaction with the PIC at promoters, strongly depends on the presence of TBP in the PIC. However, depletion of Rpb3 in the kin28-AA background has a rather different effect-there is some shift from promoter to UAS, but the amount of tail signal (as well as head signal) overall is strongly reduced. This could be interpreted to indicate that Pol II is required for recruitment of Mediator per se, even to UAS elements; however, this possibility is disfavored by the fact that depletion of Rpb3 alone increases UAS occupancy of tail and scaffold subunits (Figure 2). Thus, they interpret the decrease in Mediator occupancy on Rbp3 depletion, seen only in the presence of Kin28 depletion, to result from a failure to retain Mediator at promoters following the shift from UAS to promoters that is evoked by depletion of the CDK module, thus implicating Pol II itself in stabilizing the PIC at the promoter. The shift of some mediator occupancy from promoter to UAS observed on depleting Rpb1 together with Kin28 is being interpreted (I believe) as an indirect result of reduced TBP occupancy on depletion of Rpb3 and attendant reduction in the TBP-dependent shift from UAS to promoter at certain genes; however the reduced TBP occupancy on depletion of Rpb3 should probably be confirmed experimentally if not already demonstrated previously.

Depleting Taf1 together with Kin28 has yet another consequence, of greatly reducing Mediator occupancy at the promoters, without evoking any shift back to the UAS. Thus, it seems that Taf1 could also be required for stabilizing Mediator at promoters and its depletion would have no effect on the shift from UAS to promoter and only on Mediator retention at promoters following the shift evoked by depletion of the CDK module. However, it seems possible that these results could be explained differently if Taf1 contributes to initial recruitment, which I think they are actually suggesting in panel B of the model in Figure 6. I believe that they should be asked to deplete Taf1 alone to determine whether Mediator occupancy at UASs is increased (as in Rpb3 depletion) or decreased, and if the latter is observed, this would implicate Taf1 in the initial recruitment of Mediator to promoters.

General critique:

While some of their results can be viewed as being largely confirmatory of previous findings from the Struhl and Robert labs, other findings increase our understanding of how Mediator is recruited to genes, and the transition of Mediator from the UAS to the promoter. Examination of the triple Mediator tail subunit deletion, in which the tail is completely eliminated provides stronger evidence that obtained previously with just two subunits deleted, that the tail module is important for Mediator recruitment at most active genes; but also reveal a tail-independent pathway that allows appreciable Mediator recruitment and transcription to continue, which they are able to argue involves Mediator recruitment by Pol II. Their findings that TBP is required for the transition of Mediator from UAS to promoter, even when the CDK module is depleted, whereas Pol II is important for retention of Mediator at promoters, and might contribute only indirectly to enhancing the UAS-promoter transition by enhancing TBP occupancy at the promoter, are significant. The role of Taf1 is clearly distinct from that of TBP, but it remains unclear whether Taf1 is enhancing Mediator recruitment to the UAS (like TBP) or stabilizing Mediator at the promoter after the UAS-promoter transition (like Pol II).

To resolve this last issue, the authors should conduct the additional experiment of depleting Taf1 alone to clarify its role in Mediator recruitment vs stabilization at promoters.

Reviewer #2:

The manuscript by Knoll and colleagues describes the effect of PIC components depletion on Mediator occupancy in budding yeast. They show that Mediator association with core promoters depends on the integrity of the PIC. Although this had never been formally demonstrated, this finding was expected based on the bulk of data available on Mediator-PolII and Mediator-GTFs interactions in vitro, including recent structural data. They also show that tampering with the PIC leads to increased detection of Mediator (although not all subunits tested) at UAS. This last finding was less expected and the fact that not all subunits behave the same is intriguing. The experiments are well executed and controlled. As is stands, however, the manuscript only represents a small increment of knowledge relative to current literature. Further exploring the effect of PIC depletion on Mediator-UAS interactions could potentially increase the impact of the work.

Specific comments:

First section entitled "Mediator is recruited to gene promoters in the absence of the tail module triad" is simply confirming data from Struhl and Robert labs. The only difference here is the use of a triple deletion of the triad (instead of a double deletion that leads to the loss of the third component). The data is very nice but constitutes a very small increment over published data.

The second section entitled "Mediator association at UAS regions is stabilized by loss of TBP or Pol II" shows that Mediator (Med2, Med15 and Med14) detection at UAS is increased upon disruption of the PIC by Rpb3 or TBP anchor-away. This observation is interesting and the authors interpret this data in the context of the "transition model" proposed by Struhl and Robert labs and suggest that interfering with the PIC leads to the stabilisation of Mediator at UAS. Unexpectedly, however, no change is observed with the Head subunit Med18. If the whole Mediator was stabilised at UAS as proposed by the authors, then all subunits should be affected. The Med18 data may suggest more complex interpretations. For instance, it may be that after its transient interactions with the PIC, Head/Middle dissociates, leaving the Tail stably associated with the UAS. In subsequent rounds, the Head/Middle would re-attach on the UAS-bound Tail and so on. This section of the manuscript has probably the most potential for novelty, but more experiments are needed. I would also suggest reorganizing Figure 2 (and Figure 4). The data for all subunits tested should be represented the same way. Why are Med18, Med2 and Med14 not fully represented as for Med15? Also, why are Med2 and Med14 mapped on SAGA and TFIID genes while the other subunits are mapped on "UAS genes".

The third and fourth sections "Mediator association with promoters depends on PIC components" and "Depletion of TBP, and to lesser extent Pol II, stabilizes Mediator occupancy at UAS regions" examine the effect of TBP, TAF1 and Rpb3 depletion on Mediator occupancy at promoters and UAS in kin28AA cells. I found the description of this data somewhat confusing and some of the conclusions not to be supported by the data. First, both sections really describe the same data so I do not see why separating them. It would be simpler to go through the data systematically and (as suggested for Figure 2) represent all datasets the same way (Figure 4). Currently, the third section describes TAF1 and Rpb3 and says nothing about TBP. It is not clear why. Second, it seems to me that all three depletions lead to the same phenotype, although with different severity. The TBP depletion is described as having fundamentally different effects on Mediator compared to Rpb3 and TAF1 depletions but one could argue that TBP depletion simply has a more pronounced effect on Mediator as a consequence of a more profound effect on the PIC. The authors should show the effect of their depletions on a few PIC components. One would expect TBP depletion to abolish the PIC altogether, whereas TAF1 and Rpb3 depletions may leave partial PIC behind. Alternatively, the differences in magnitude may be due to variable efficiencies (or timing) of depletion for TBP, Rpb3 and TAF1, a possibility not acknowledged. On a related note, the authors should show the depletion of their AA strains over time (easily done by ChIP-qPCR) in order to justify the time point used and relativize the effect observed for each of them. Third, I do not see what data allows the authors to conclude about timing of events as claimed in their discussion: "[…] TBP being most important for transit of Mediator from UAS to promoter and Rpb3 and Taf1 being more important for stabilizing occupancy at the promoter following transit".

Reviewer #3:

This paper describes a very nice set of genome-wide experiments that demarcate how and where Mediator is recruited to promoters. In one pathway the UAS/Activator complex recruits Mediator via the tail module to the UAS region, and then ultimately it sets down over the PIC. In the second pathway, which is UAS-independent and tail-independent, Mediator is directly recruited to the PIC, which most likely involves TFIID. The data and manuscript were presented clearly, and in general the conclusions were supported by the data. Overall the manuscript represents an important contribution to our understanding of Mediator, and will be valued by the scientific community.

Specific comments

1) It seems that single replicates were performed for some "main figure" experiments, including Figure 1A (Med2 TAP). Please confirm Figure 1A with a second biological replicate for Med2, since this is a main figure and drives the rationale for the rest of the manuscript. Also, is there an untagged/input control for Figure 1A? This would be helpful to get a sense of the recruitment of the tailless mutant over background.

2) There needs to be a correlation coefficient for Figure 1C to get a quantitative measure of difference in the gene expression in mutant relative to WT. "dispensable for most gene expression" would be more convincing with showing the numerical value for the graph.

3) Figure 2B, why isn't occupancy of Med 18 changed with AAR?

4) Anchor away was performed for 1 hr. Please provide an explanation for why this time frame, as opposed to less time like 20-30 min. Obviously, there is a trade-off between completeness of the sequestering and the potential for indirect effects to creep in. Was this condition chosen for a reason or was it arbitrary? Is there GFP localization experiment to verify the anchor away?

5) The results in Figure 4 are striking, and show that when Kin28 is depleted Mediator occupancy at TFIID promoters is highly dependent on Pol II and Taf1. I'm a bit puzzled as to why there needs to be a dependence on Kin28. Or is that a product of the experimental design, in that Taf1 or pol II AAR was not conducted in a WT KIN28 strain? Why are opposite effects seen for Med 15 and Med 18 in tbp+kin28 AAR?

[Editors’ note: what now follows is the decision letter after the authors submitted for further consideration.]

Thank you for submitting your article "Role of the pre-initiation complex in Mediator recruitment and dynamics" for consideration by eLife. Your article has been reviewed by three peer reviewers and the evaluation has been overseen by a Reviewing Editor, Alan Hinnebusch, and Kevin Struhl as the Senior Editor. The reviewers have opted to remain anonymous.

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

While two of the reviewers were satisfied with the additional work and revisions you made to the previous (rejected) version of the manuscript, and considered the results now to be worthy of publication after some minor revisions, reviewer #2 has raised serious concerns about the apparent lack of reproducibility between certain replicates, and a lack of clarity about whether results from two replicates have been used generally to draw key conclusions throughout the work. The recommended approach would be to draw conclusions after combining results from at least two highly correlated replicate datasets. Hence, it is possible that you would be obliged to conduct additional replicate experiments and re-evaluate your findings and conclusions after combining results from only highly correlated datasets. Reviewer #2 has also raised concerns that the AA tag on Taf1 is affecting its function in the absence of Rapamycin, and this complication has to be taken into account in the interpretation of results obtained in the presence of Rapamycin for this mutant, as the tag is setting a baseline of constitutive impairment against which the effect of the drug must be compared. Reviewer #3 has also raised some minor concerns that would also need to be addressed.

Reviewer #1:

The authors have done a considerable amount of additional work to address the major criticisms of the previous version. In particular, they carried out the experiment I had requested of depleting Taf1 in a strain with WT Kin28, and can now rule out a role for Taf1 in Mediator recruitment.

The work is significant for several reasons. Their findings indicating that TBP is required for transit of Mediator from UAS to promoter, whereas Taf1 of TFIID is dispensable for that function but appears to be as important as Pol II itself in retaining Mediator at promoters following the transfer from UAS are both important findings that could not have been predicted from previous work. It is also interesting that the importance of Taf1 and Pol II in retaining Mediator at promoters differs significantly in magnitude between TFIID and SAGA-dominated promoters. The additional work done to create the triple tail subunit deletion and quantify the effects of a completely tail-less mediator on recruitment of Mediator and Pol II occupancy is also valuable, as the effects are more severe than observed for the double tail subunit deletion. They also show unequivocally that the tail is not essential for Mediator recruitment, and that other interactions of Mediator with Pol II, TFIID, or TBP suffice for less efficient recruitment in the absence of the tail. On the other hand, the TBP anchor away experiments show that PIC formation is dispensable for Mediator recruitment in the presence of an intact Mediator tail, as Pol II recruitment is nearly abolished on TBP AA, with a shift in Mediator occupancy to UAS regions indicating that Mediator is being recruited but is stranded at the UAS in the absence of TBP's function in Mediator transit. These latter experiments also reveal that Mediator is recruited to the UASs of many genes where it cannot normally be detected.

I have no additional criticisms.

Reviewer #2:

The manuscript by Knoll et al. describes how yeast Mediator occupancy at UAS and promoters is affected by PIC components. Using combinations of anchor-away experiments coupled to ChIP-seq, they propose that TBP is required for the transit of Mediator from UAS to promoters while Pol II and Taf1 are required for stabilising Mediator at promoters. They also provide data showing that the Tail triad is not required for the association of Mediator with promoters.

This work is building on prior work from the Struhl and Robert lab and several aspects are largely confirmatory. The most novel aspect concerns how different PIC components contribute to the association of Mediator with promoters. Unfortunately, the data appears to be of variable quality and the conclusions are based on interpreting cherry-picked replicates while ignoring others that would have led to very different conclusions.

My understanding is that some experiments were done as only one replicate while others were done in duplicates. The main problem is that the analysis appears to have been done on one of the replicates (which is shown in the main Figures) while the second replicate is (at least in some occasions and it is not clear whether this is systematic) shown as a figure supplement and not used to build the story. Normally, experiments are done at least in duplicate and some correlation metrics are used to insure they are alike. Then replicates are combined and analysed further. Here, not only the correlation between replicates has not been assessed but one is used and the other relayed to a sup figure. Yet, at least in some cases, the replicates are so different that considering the alternate would have led to a completely different story. Here is a striking example: Figure 3A shows that Med15 occupancy is massively increased upon anchor-away of TBP. As the authors emphasised, this is radically different from Med18, the occupancy of which is not affected by the anchoring of TBP (as shown in Figure 3B). Yet, the second replicate of these experiments, shown in Figure 3—figure supplement 3, shows no effect of TBP anchoring on Med15 (rather that a massive increase) and a decrease for Med18 (rather than no change). Similar inconsistencies are also present for the rpb3 and taf1 anchoring on the same sets of Figures. The massive increase of Med15 occupancy upon TBP anchoring being the founding data for the story, I am finding it difficult to consider the rest of the story. In addition to Figure 3, the main results of this paper are those in Figure 5. In this case, I was not able to figure out whether the data presented in Figure 5 was for 1 rep out 1, 1 rep our 2 (in which case rep2 is nowhere to be found) or a combination of 2 reps. Given the variability between reps observed in Figure 3, and not knowing about the replicate status of Figure 5, I am not confident in interpreting data from Figure 5.

Another major concern with the manuscript has to do with the Taf1 anchor-away data. As acknowledged by the authors, the taf1-FRB strain appears to show a phenotype even in the absence of rapamycin. This means that the tagging affected the function of the protein. Because rapamycin did not lead to any addition effect, I do not think that the data can be interpreted as a consequence of conditional depletion. Unless the authors find a way to conditionally affect the function of Taf1 (perhaps auxin degron or tagging FRB in N-terminal would work better), I believe the Taf1 data has to be removed.

In sum, I recommend the authors go back to the basics and make sure their experiments are reproducible. For that, they need to show that their replicates are highly correlated. At current state, I suspect that most of the effects observed are batch effects rather than biological effects. For instance, simply inspecting the red tracks in Figure 3—figure supplement 5A clearly shows that rep2 did not work nearly as well as rep1. Not seeing the labels, one could not even tell how to pair them. Also, looking at the blue tracks from the same Figures suggests that the samples cluster by batches rather than by conditions.

Reviewer #3:

In this revised manuscript, the authors show that mediator recruitment to upstream activated regions (UAS) via transcription activators may be dependent or independent of the tail module. Importantly their triple KO when compared to WT has reduced occupancy of mediator complex (via Med17 ChIP) and PolII (via Rbp1 ChIP) at most genes, but occupancy of both is present in the absence of the Med2, Med3, and Med15. This reinforces the role of the Tail in binding to UAS and that the UAS enhances transcription, but importantly highlighting that even without Med2, there is minor mediator occupancy at UAS under conditions of basal transcription. This point is not merely confirmatory but key to understanding the overall role of mediator in basal and stress-induced transcription not only in yeast, but even in mammals akin to the role of enhancers enhancing transcription through the tail when needed. They show increased occupancy of Med15 (tail) with depletion of Tbp or Rpb3, but this is not noted for Med18 (head) or Med14 (backbone). This may suggest that PIC disruption may lead to Mediator not moving from the UAS region to the promoters (being stalled). Further, they show that Med15 and Med18 occupancy is decreased more at TFIID-dominated genes than SAGA-dominated genes upon Taf1 and Rpb3 depletion in the context of Kin28 anchor away. TBP depletion causes an upstream shift of mediator occupancy at SAGA-dominated genes (corresponding to UAS) and usually a loss of mediator occupancy at TFIID-dominated genes. Rpb3 loss leads to decreased occupancy, particularly at TFIID-dominated genes and an upstream shift of Med15 occupancy at SAGA-dominated genes with concomitant Kin28 depletion.

This paper clearly shows key mediator occupancy differences at SAGA- and TFIID-dominated genes in connection with PIC stabilization that is in line with many other papers. Importantly, they connect the mediator tail module to SAGA-dominated genes via their triple Tail KO. They demonstrate that the movement of the mediator from UAS to promoter is dependent on TBP and that the mediator is stabilized at promotors especially in TFIID-dominated genes by Taf1. While the model of dependent and independent recruitment of mediator by its tail module is plausible, the differentiation of mediator dynamics at SAGA- vs. TFIID-dominated genes is a very exciting aspect of this paper and nicely highlights mediator movement after initial recruitment in the context of PIC components.

Multiple aspects of this paper connect other previously isolated papers such as SAGA & TFIID-dominated genes, UAS and TSS distance in relationship to mediator, and PIC-component interactions with mediator modules; in so doing, they put minor aspects of key mechanisms at the UAS-promoter looping into context of each other to present a clear and more encompassing model. We greatly appreciate them starting to perturb the assumed complex relationship between PIC components and mediator and think their data will enable others to further explore these relationships. Moreover, the authors have sufficiently addressed our concerns from the previous review. We find this paper not to be merely confirmatory but addressing key issues of mediator biology.

[Editors' note: further revisions were requested prior to acceptance, as described below.]

Thank you for resubmitting your work entitled "Role of the pre-initiation complex in Mediator recruitment and dynamics" for further consideration at eLife. Your revised article has been favorably evaluated by Kevin Struhl (Senior Editor), a Reviewing Editor, and one reviewer.

The manuscript has been improved with the addition of new datasets, but there are several remaining issues raised by reviewer #2 that need to be addressed before acceptance. Please make appropriate revisions of text in response to all of the criticisms/comments in the review below and provide a point-by-point explanation of your responses along with the revised manuscript.

Reviewer #2:

The revised manuscript is improved, notably by the addition of datasets completing some that were done only once and by the replacement of non-consistent replicates. However, there are a few issues that are still of concern and should be addressed.

First, some of the anchor-away strains behave strangely. While this is unlikely to drive erroneous conclusions, it should at least be mentioned. For instance, Kin28-TBP-AA Med15-Myc is very sick according to Figure 3—figure supplement 1. More worrisome, tagging various PIC components with FRB greatly affects Mediator ChIP in the absence of rapamycin. This is evident by looking at several figures, notably in Figure 3 (pay attention to the y-axis for all the no rapa traces (black) or look at the intensities of the no rapa heatmaps). This is also directly shown in Fig 3—figure supplement 2B. The authors use this figure supplement to claim that Taf1-AA leads to increased Mediator ChIP signal in absence of rapamycin but they fail to acknowledge that this is also the case for TBP-AA (the average signal is 2 to 3 fold up compared to non-tag). It is not clear whether these differences are batch differences (due to technical variation between experiments performed on different days) or biological differences (caused by partial loss of function of tagged proteins).

Second, the Med18 ChIP data displayed in Figure 3 and showing that Med18 does not change in TBP-AAR or Rpb3-AAR does not fit the author's model very well. In subsection “Mediator association at UAS regions is stabilized by loss of TBP or Pol II” the authors say "The increase in ChIP signal observed for Med15, but not for Med18, would be consistent with the most immediate contacts with UAS-bound activators occurring through tail module subunits (such as Med15), while ChIP signal for Med18 at UAS regions would depend on protein-protein cross-links, thus lowering the efficiency of ChIP for this subunit". While I agree that this explains the lower ChIP efficiency of Med18 at UAS, it does not explain why, unlike other subunits tested, Med18 signal does not increase at UAS in TBP-AAR and Rpb3-AAR. The observed Med18 ChIP signal is clearly over the UAS so if Mediator is stuck there, this Med18 signal should increase together with other subunits. The fact that it does not (assuming the data is real), suggests a more complicated model. This result should at least be described more critically.

Third, the conclusion that TBP is required for the transition while Rpb3 and Taf1 are stabilising the PIC-associated Mediator is in my opinion a hazardous extrapolation. As explained by the authors in subsection “TBP is required for transit of Mediator from UAS to promoter”, depletion of PIC components has impact on other PIC components. This makes the attribution of specific roles to specific PIC components in Mediator dynamics difficult. Hence, with these limitations in mind, it seems like an over interpretation to say "TBP being critical for transit of Mediator from UAS to promoter, while Pol II and Taf1 stabilize Mediator association at proximal promoters". A more conservative interpretation would be that the PIC is necessary to allow Mediator transiting to the promoter (as shown when PIC is severely compromised; TBP-AA) and that partially formed PICs (without trying to attribute specific roles to subunits or proteins), while allowing the transit, lead to unstable Mediator-PIC association.

These last two points had been raised in previous review rounds but are still not clearly dealt with in the most recent version of the manuscript.

Finally, it is not clear why the authors were expecting a decrease of Mediator at UAS upon depletion of TBP, Rpb3, or Taf1. Isn't an increase what the "dynamic model" proposed by the Struhl and Robert labs predicted?

[Editors’ note: the author responses to the first round of peer review follow.]

[…] In view of the considerable amount of new experimental work that seems necessary, the extended time-frame this would likely entail, and the largely uncertain consequences of the new results for your final model, it was decided that the manuscript should be rejected rather than returned to you for revisions. The following list summarizes the most important criticisms, some of which were already mentioned above.

1) Repeat the Taf1 anchor away experiment in a WT KIN28 strain to determine whether Med15 occupancy at UASs is increased (as in Tbp3 depletion) or decreased, because if the latter was observed, this would implicate Taf1 in the initial recruitment of Mediator to promoters in addition to, or instead of, the transition of Mediator from UAS to PIC.

We have included Taf1 anchor away experiments as requested; these are included in the revised Figure 3 (formerly Figure 2). These experiments clearly show that there is no decrease in Mediator occupancy at UAS regions upon depletion of Taf1.

2) To confirm the differences seen between the Tail Med15 and Head Med18 occupancies on depletion of TBP in the kin28-AA strain in Figure 4C, examine the genome-wide behavior of additional Head subunits.

We have included new experiments (Figure 5—figure supplement 4) examining the occupancy of two additional Mediator subunits, Med17 from the head module and Med2 from the tail module, after simultaneous depletion of Kin28 and TBP. The results corroborate those obtained with Med15 and Med18.

3) Compare the effects of depleting TBP, Rpb3, or Taf1 on PIC assembly to rule out the trivial possibility that the different outcomes on depletion of each factor represent different extents of reducing PIC assembly rather than indicating distinct functions in Mediator transfer or retention at the promoter. This would also address whether depletion of one factor by anchor away is simply more efficient than another.

We have included data on the effect of depleting PIC components on occupancy of other PIC subunits, and also cited relevant literature on this point (we had cited a bioRxiv paper on this previously, which has now been published in Genes & Development). Specifically, we document depletion of TBP, Rpb3, and Taf1 (which is less complete) in the corresponding anchor away strains in Figure 3—figure supplement 4 and 5. We also show the effects on occupancy of Pol II following depletion of TBP (Figure 3—figure supplement 4D), TBP and Kin28 (Figure 7A), and Taf1 and Kin28 (Figure 7A); on occupancy of Taf1 after depletion of TBP or TBP and Kin28 (Figure 7B); on occupancy of Taf1 after depletion of Rpb3 (Figure 7C); and on occupancy of TBP after depletion of Rpb3, and occupancy of Rpb1 after depletion of TBP (Figure 3—figure supplement 4). The implications of these results on interpreting Mediator occupancy changes after PIC component depletion are considered towards the end of the Results section.

4) Results for all four Mediator subunits, for all groups of genes, must be shown for a comprehensive analysis of the effects of depleting TBP vs. Rpb3 vs Taf1.

The results for the effects on occupancy of four Mediator subunits following depletion of TBP or Rpb3 are shown in Figure 3 (formerly Figure 2) in the revised manuscript, and results for two subunits are shown for Taf1 (all that we have done), as little change was seen for Taf1, and the major point for that PIC component is that recruitment of Mediator is not decreased upon its depletion. These results are now presented in the same way (at “UAS genes”) in all cases.

5) Provide a biological replicate for the results on Med2 in Figure 1A, as well as an input or untagged control.

We conducted a biological replicate for the results in Figure 1A of the original submission, and discovered that the signal we observed for Med2 occupancy was not significantly different to that seen with a control, untagged strain. We therefore have removed this part of Figure 1 from the manuscript.

6) Present time-courses of anchor away to justify the rapamycin treatments chosen for each protein.

We have presented a time course for anchor away of TBP, and also provide justification of our choice of one hour depletion by comparison with other published work using this method to deplete components of the transcription machinery.

In addition, we have included new data showing that Pol II occupancy is reduced genome-wide to about 25% of wild type levels in the med2∆ med3∆ med15∆ mutant, demonstrating that the tail module plays an important role in Pol II occupancy at essentially all genes.

Reviewer #1:

[…] While some of their results can be viewed as being largely confirmatory of previous findings from the Struhl and Robert labs, other findings increase our understanding of how Mediator is recruited to genes, and the transition of Mediator from the UAS to the promoter. Examination of the triple Mediator tail subunit deletion, in which the tail is completely eliminated provides stronger evidence that obtained previously with just two subunits deleted, that the tail module is important for Mediator recruitment at most active genes; but also reveal a tail-independent pathway that allows appreciable Mediator recruitment and transcription to continue, which they are able to argue involves Mediator recruitment by Pol II. Their findings that TBP is required for the transition of Mediator from UAS to promoter, even when the CDK module is depleted, whereas Pol II is important for retention of Mediator at promoters, and might contribute only indirectly to enhancing the UAS-promoter transition by enhancing TBP occupancy at the promoter, are signficant. The role of Taf1 is clearly distinct from that of TBP, but it remains unclear whether Taf1 is enhancing Mediator recruitment to the UAS (like TBP) or stabilizing Mediator at the promoter after the UAS-promoter transition (like Pol II).

To resolve this last issue, the authors should conduct the additional experiment of depleting Taf1 alone to clarify its role in Mediator recruitment vs stabilization at promoters.

We thank the reviewer for a thoughtful and thorough critique, and are pleased that s/he finds our results significant. The reviewer states that the experiments using the med2∆ med3∆ med15∆ are “largely confirmatory” of previous work from the Struhl and Robert labs, but also states that this result “provides stronger evidence than obtained previously”. We agree with both of these comments, mostly; although the effect of the triple deletion is not unexpected, the results establish definitively that the tail module triad is non-essential, contributes considerably to Mediator recruitment at most active genes, and demonstrates that alternative modes of Mediator recruitment and gene activation, independent of the tail module triad, exist. Prior to this work, the possibility existed that the tail module triad was essential, and that residual function of Med2 was sufficient to allow viability in the double mutant; this is now conclusively ruled out. We have also added a new experiment showing that Pol II occupancy is reduced to ~25% of wild type levels in the med2∆ med3∆ med15∆ mutant across the genome.

With regard to the last point, that it is unclear whether Taf1 is important for Mediator recruitment to the UAS or for stabilizing Mediator at the promoter after transit from the UAS, we have included in the revised manuscript experiments that provide evidence that Mediator continues to occupy UAS regions after Taf1 depletion, in the presence of active Kin28. These data support the idea that Taf1 is important for stabilization of Mediator at promoters after transit from the UAS.

Reviewer #2:

[…] First section entitled "Mediator is recruited to gene promoters in the absence of the tail module triad" is simply confirming data from Struhl and Robert labs. The only difference here is the use of a triple deletion of the triad (instead of a double deletion that leads to the loss of the third component). The data is very nice but constitutes a very small increment over published data.

We are pleased that the reviewer considers our data to be solid. We disagree that the first section, regarding the knockout of the complete tail module triad, is “simply confirming data from Struhl and Robert labs.” Our finding that Mediator occupancy at UAS regions in med2∆ med3∆ med15∆ yeast is nearly completely eliminated can fairly be regarded as confirmatory, as a similar result was reported for the double mutant (med3∆ med15∆) and the clear expectation would be that loss of Mediator occupancy at UAS regions would be seen in the triple mutant also. However, it could not be predicted with confidence that the triple mutant would be viable, nor what the effect would be on Mediator occupancy at promoters. Our results provide the first definitive evidence that the Mediator tail module triad is not essential, and that Mediator can be recruited to gene promoters in its absence. We also show in the revised manuscript that association of both Mediator and Pol II is substantially decreased in the triple mutant, considerably more than the less than two-‐fold effect reported previously (Jeronimo et al., 2016), showing that the tail module triad is broadly important for recruitment of these components.

We agree with the reviewer that based on the in vitro results of Carey and co-workers, it seemed likely that Mediator association with gene promoters would depend on PIC integrity. But in vitro results do not always accurately predict in vivo interactions, and this prediction had not previously been tested experimentally.

The second section entitled "Mediator association at UAS regions is stabilized by loss of TBP or Pol II" shows that Mediator (Med2, Med15 and Med14) detection at UAS is increased upon disruption of the PIC by Rpb3 or TBP anchor-away. This observation is interesting and the authors interpret this data in the context of the "transition model" proposed by Struhl and Robert labs and suggest that interfering with the PIC leads to the stabilisation of Mediator at UAS. Unexpectedly, however, no change is observed with the Head subunit Med18. If the whole Mediator was stabilised at UAS as proposed by the authors, then all subunits should be affected. The Med18 data may suggest more complex interpretations. For instance, it may be that after its transient interactions with the PIC, Head/Middle dissociates, leaving the Tail stably associated with the UAS. In subsequent rounds, the Head/Middle would re-attach on the UAS-bound Tail and so on. This section of the manuscript has probably the most potential for novelty, but more experiments are needed. I would also suggest reorganizing Figure 2 (and Figure 4). The data for all subunits tested should be represented the same way. Why are Med18, Med2 and Med14 not fully represented as for Med15? Also, why are Med2 and Med14 mapped on SAGA and TFIID genes while the other subunits are mapped on "UAS genes".

We have remade Figure 2 (Figure 3 in the revised manuscript) so that it shows heat maps and line graphs at “UAS genes” for all four Mediator subunits. Because depletion of Taf1 showed little effect on Med15 and Med18, we did not examine the effect of Taf1 depletion on Med2 and Med14. As to the possibility that the tail module might interact with UASs in the absence of the rest of Mediator, we do mention this in the text but do not favor this possibility. First, the increased occupancy seen for Med14 upon depletion of Rpb3 or TBP (Figure 3D) is consistent with results for Med15 and Med2, and suggests that Mediator as a whole shows increased occupancy under these conditions. We believe this interpretation is most consistent with results from the Struhl and Robert labs, and in particular with the model shown in Figure 7 of Jeronimo et al., 2016. We show in Figure 3—figure supplement 5B that, as in the Jeronimo et al. paper, we observe ChIP signal at some genes after Kin28 depletion at both UAS and promoter regions for Med15, but only at promoter regions for Med17 and Med18. When considered together with the re-ChIP experiments in Petrenko et al., 2016, Figure 1, these results seem most consistent with Med18 exhibiting weaker ChIP signal at UAS regions than does Med15.

The third and fourth sections "Mediator association with promoters depends on PIC components" and "Depletion of TBP, and to lesser extent Pol II, stabilizes Mediator occupancy at UAS regions" examine the effect of TBP, TAF1 and Rpb3 depletion on Mediator occupancy at promoters and UAS in kin28AA cells. I found the description of this data somewhat confusing and some of the conclusions not to be supported by the data. First, both sections really describe the same data so I do not see why separating them. It would be simpler to go through the data systematically and (as suggested for Figure 2) represent all datasets the same way (Figure 4). Currently, the third section describes TAF1 and Rpb3 and says nothing about TBP. It is not clear why. Second, it seems to me that all three depletions lead to the same phenotype, although with different severity. The TBP depletion is described as having fundamentally different effects on Mediator compared to Rpb3 and TAF1 depletions but one could argue that TBP depletion simply has a more pronounced effect on Mediator as a consequence of a more profound effect on the PIC. The authors should show the effect of their depletions on a few PIC components. One would expect TBP depletion to abolish the PIC altogether, whereas TAF1 and Rpb3 depletions may leave partial PIC behind. Alternatively, the differences in magnitude may be due to variable efficiencies (or timing) of depletion for TBP, Rpb3 and TAF1, a possibility not acknowledged. On a related note, the authors should show the depletion of their AA strains over time (easily done by ChIP-qPCR) in order to justify the time point used and relativize the effect observed for each of them. Third, I do not see what data allows the authors to conclude about timing of events as claimed in their discussion: "[…] TBP being most important for transit of Mediator from UAS to promoter and Rpb3 and Taf1 being more important for stabilizing occupancy at the promoter following transit".

We agree that depletion of TBP is likely to have a more profound effect on the PIC than depletion of Rpb3 or Taf1, but we are puzzled by the statement that all three depletions lead to the same phenotype, but with different severity. The data in Figures 4C and D (Figure 5 in the revised manuscript) show that both Med15 and Med18 peaks shift upstream upon depletion of TBP at SAGA-dominated genes and genes having upstream Mediator occupancy in wild type yeast (“UAS genes”); Med15 can be seen to shift upstream at TFIID-dominated genes as well. Depletion of Rpb3 or Taf1 leads to a partial upstream shift or no upstream shift, respectively, but a strong decrease in signal. If TBP depletion simply had a more pronounced effect than depletion of Taf1, for instance, it should result in a stronger decrease and no upstream shift, or Taf1 should give rise to a partial upstream shift. These results indicate that TBP is necessary for Mediator occupancy to shift from UAS to promoter, but cannot determine whether it is sufficient, as Rpb3 occupancy is abolished by loss of TBP (Figure 3—figure supplement 4D and Figure 7A). In contrast, Mediator occupancy does not shift upstream upon depletion of Taf1, and only partially upon depletion of Rpb3, suggesting that these PIC components are not necessary for Mediator to transit from UAS to promoter region. Hence their behavior is qualitatively different. I believe the other two reviewers agree with this point.

We have included new data and citations to show the effect of individual PIC components on each other both with and without Kin28 depletion (Figure 3—figure supplement 4 and Figure 7). As the reviewer suggests, depletion of TBP eliminates Pol II association, while depletion of Rpb3 only causes partial loss of TBP, consistent also with recently published work (Joo et al., 2017). Perhaps surprisingly, Taf1 occupancy was only modestly reduced upon depletion of TBP or Rpb3 (Figure 7). We have added text to the Results that considers the results seen on Mediator occupancy in light of cooperative assembly of the PIC.

With regard to our rationale for using a one hour treatment with rapamycin, we chose this as a reasonable balance that would allow near complete depletion but not allow much time for indirect effects, as reviewer #3 suggests. Our ChIP data for anchoring away TBP indicated that depletion was essentially complete after 30 minutes, and we have now included this experiment as supplemental data. Other studies using anchor away to deplete components of the transcription machinery have used similar times of rapamycin treatment (30 min for Taf proteins (Warfield et al., 2017), 1 hr for Spt7 (Baptista et al., 2017) and for Mediator subunits, TBP, and Rpb1 (Wong and Struhl, 2014; Petrenko, 2016, 2017); 90 min for Mediator subunits (Jeronimo et al. 2016); 120 min for Mediator subunits (Anandhakumar et al., 2016)). Interestingly, the latter study found only about 80% depletion of Mediator subunits after 1 hr of rapamycin treatment, whereas the Struhl laboratory found depletion to background levels in 1 hr (Wong and Struhl, 2014), although the two studies used different criteria to monitor depletion. It is possible that the GFP tags also employed in the former study affected kinetics of depletion, but this has not been rigorously investigated.

Finally, with regard to organization, we prefer to keep the two sections mentioned separate. The first reports the evidence that Rpb3 and Taf1 are required for stable association of Mediator at promoters, while the second reports on the shift to the UAS region seen upon depletion of TBP. These distinct outcomes are best emphasized, in our opinion, by being discussed in separate sections.

Reviewer #3:

[…] 1) It seems that single replicates were performed for some "main figure" experiments, including Figure 1A (Med2 TAP). Please confirm Figure 1A with a second biological replicate for Med2, since this is a main figure and drives the rationale for the rest of the manuscript. Also, is there an untagged/input control for Figure 1A? This would be helpful to get a sense of the recruitment of the tailless mutant over background.

We repeated the experiment of Figure 1A and included an untagged control. The signal seen in this experiment for Med2 at UAS regions was small (as in the original figure), but importantly not different from that seen with the untagged control. We have therefore not included this figure in the revised manuscript. However, we do not (and did not) consider this to be essential as a rationale for examining the complete deletion of the tail module triad. Prior to the work reported here, it remained possible that Med2 could contribute to Mediator recruitment to UAS regions and/or promoters in med3∆ med15∆ yeast; we have now formally excluded this possibility, while also defining rigorously the extent to which the tail module triad contributes to Mediator and Pol II recruitment genome-wide.

With regard to other experiments, in all cases where exact replicate experiments were not performed, we have performed corroborating experiments by doing ChIP against multiple Mediator subunits under the same conditions (e.g. Figure 3), or have relied on published results that corroborate our experiments (e.g. decreased occupancy of TBP upon depletion of Pol II). We have also added new experiments in which occupancy of two additional Mediator subunits, Med2 and Med17, is examined after depletion of Kin28 and TBP. The results are consistent with those observed for Med15 and Med18, with the tail module subunit, Med2, exhibiting greater signal at the UAS following depletion, while the head module subunit, Med17, shows strongly decreased signal relative to depletion of Kin28 alone, but very little upstream signal.

2) There needs to be a correlation coefficient for Figure 1C to get a quantitative measure of difference in the gene expression in mutant relative to WT. "dispensable for most gene expression" would be more convincing with showing the numerical value for the graph.

We have added correlation coefficients to the graphs in question and calculated Fisher’s z-score for the difference between correlation coefficients of the two graphs, yielding a p-value < 10^-10. However, we have revised our conclusion regarding the role of the tail module in gene expression based on the new data showing that Pol II association is substantially (about 4‐fold) reduced in the tail module deletion mutant: evidently the tail module is broadly important for Pol II recruitment, but the effect of this reduced recruitment is likely offset by altered RNA decay, as has been observed for other perturbations in PIC components.

3) Figure 2B, why isn't occupancy of Med 18 changed with AAR?

We do not have a definitive answer for this question, but speculate that it may reflect Med18 being less proximate to UAS regions and correspondingly exhibiting a weaker ChIP signal there, as proposed by Jeronimo et al., 2016; see also our response to reviewer #2. Alternatively, Mediator may exist in a configuration that decreases accessibility to Med18 under the conditions of TBP depletion. Reviewer #2 suggests it could reflect loss of the Mediator head module (or perhaps only of specific subunits), but we do not favor this explanation at present as there is little evidence to support it.

4) Anchor away was performed for 1 hr. Please provide an explanation for why this time frame, as opposed to less time like 20-30 min. Obviously, there is a trade-off between completeness of the sequestering and the potential for indirect effects to creep in. Was this condition chosen for a reason or was it arbitrary? Is there GFP localization experiment to verify the anchor away?

We chose 1 hr as a reasonable balance that would allow near complete depletion but not allow much time for indirect effects, as the reviewer suggests. Our ChIP data for anchoring away TBP indicated that depletion was essentially complete after 30 minutes, and we have now included this experiment as supplemental data. Other studies using anchor away to deplete components of the transcription machinery have used similar times of rapamycin treatment (30 min for Taf proteins (Warfield et al., 2017), 1 hr for Spt7 (Baptista et al., 2017) and for Mediator subunits, TBP, and Rpb1 (Wong and Struhl, 2014;Petrenko 2016, 2017); 90 min for Mediator subunits (Jeronimo et al., 2016); 120 min for Mediator subunits (Anandhakumar et al., 2016)). Interestingly, the latter study found only about 80% depletion of Mediator subunits after 1 hr of rapamycin treatment, whereas the Struhl laboratory found depletion to background levels in 1 hr (Wong and Struhl, 2014). It is possible that the GFP tags also employed in the former study affected kinetics of depletion, but this has not been rigorously investigated.

5) The results in Figure 4 are striking, and show that when Kin28 is depleted Mediator occupancy at TFIID promoters is highly dependent on Pol II and Taf1. I'm a bit puzzled as to why there needs to be a dependence on Kin28. Or is that a product of the experimental design, in that Taf1 or pol II AAR was not conducted in a WT KIN28 strain? Why are opposite effects seen for Med 15 and Med 18 in tbp+kin28 AAR?

Depletion of kin28 in the experiment of Figure 4 was used to allow detection of Mediator at promoters of active genes, as shown by the Robert and Struhl labs (Jeronimo and Robert, 2014; Wong and Struhl, 2014). We have edited the text to clarify this point. We did include an experiment in which Pol II was depleted (“pol II AAR”), shown in Figure 2 (now Figure 3), and have added data for Taf1 depletion in the presence of wild type KIN28 to Figure 3.

I believe that what the reviewer is referring to as “opposite effects seen for Med15 and Med18 in tpb + kin28 AAR” is the apparent increased signal at SAGA-dominated genes for Med15 and decrease for Med18. We believe that this reflects Mediator being stranded at the UAS region after TBP depletion; because the contact with Med15 is much closer, ChIP efficiency is high and the signal is strong, whereas the Med18 subunit is more proximate to the promoter in kin28-‐AAR conditions than to the UAS in kin28-tbp-AAR conditions, resulting in decreased ChIP efficiency and lower signal. This interpretation is consistent with the model proposed by the Struhl and Robert labs, and specifically with the detailed structural considerations discussed in Jeronimo et al., 2016. We have added text to clarify this interpretation. See also Figure 3—figure supplement 5B, and our response to reviewer #2.

[Editors' note: the author responses to the re-review follow.]

Reviewer #2:

The manuscript by Knoll et al. describes how yeast Mediator occupancy at UAS and promoters is affected by PIC components. Using combinations of anchor-away experiments coupled to ChIP-seq, they propose that TBP is required for the transit of Mediator from UAS to promoters while Pol II and Taf1 are required for stabilising Mediator at promoters. They also provide data showing that the Tail triad is not required for the association of Mediator with promoters.

This work is building on prior work from the Struhl and Robert lab and several aspects are largely confirmatory. The most novel aspect concerns how different PIC components contribute to the association of Mediator with promoters. Unfortunately, the data appears to be of variable quality and the conclusions are based on interpreting cherry-picked replicates while ignoring others that would have led to very different conclusions.

We felt this reviewer was unduly harsh in accusing us of “interpreting cherry-picked replicates while ignoring others that would have led to very different conclusions”. This statement is based on the replicate data shown in Figure 3—figure supplement 3 in the critiqued manuscript, which included data (high signal for both Med15 and Med18 in the TBP anchor away strain in the absence of rapamycin) that was inconsistent with data in the main figure. In the text of the previously submitted version we argued that the preponderance of our results were most consistent with the interpretation we provided; it is not in our opinion accurate to state that we ignored the inconsistent results. Nonetheless the reviewer’s concern regarding reproducibility is valid and was a concern to us as well; that is why we chose to include replicate data for this experiment as a supplementary figure. We have now repeated this experiment—ChIP-seq of Med15 and of Med18 in the tbp-AA yeast strain in the absence and presence of rapamycin—and the results are consistent with those we had presented in Figure 3, including results for ChIP against Med2 and Med14 (see Author response image 1). We have revised Figure 3 by presenting data for Med15 and Med18 ChIP averaged from two independent biological replicate experiments (except for the Med18 ChIP in the taf1-AA strain, for which we have only one replicate).

Author response image 1

Download asset Open asset

Reviewer #2 was also concerned about whether data in Figure 5 was based on replicate experiments. The figure as presented in the critiqued manuscript included combined data for two experiments that were consistent with two additional replicate experiments for both Med15 and Med18 ChIP in kin28-tbp-AA yeast (four replicates in total), and data for single experiments that were consistent with replicate experiments for Med15 and Med18 ChIP in kin28-AA yeast. The reason for combining data only for kin28-tbp-AA yeast is that we obtained fewer reads for those experiments, and combining data yielded better signal to noise. For Med18 in kin28-taf1-AA and kin28-rpb3-AA strains, we showed data from single experiments that were qualitatively consistent with replicate experiments. However, the extent to which Med18 signal was affected in the kin28-taf1-AA mutant differed quantitatively in these replicate experiments. We have now performed an additional replicate of this experiment. All three replicates are qualitatively consistent (see Author response image 2): depletion of Kin28 together with Taf1 decreases signal at TFIID-dominated genes, and does not substantially shift Mediator peaks at UAS genes as observed for kin28-tbp-AA yeast (see Author response image 2). The effect of depletion of Kin28 and Taf1 on Med18 occupancy at SAGA-dominated and UAS genes (which are mostly SAGA-dominated) was variable in our three replicate experiments with regard to peak magnitude relative to depletion of Kin28 alone, although it was always a less pronounced decrease than seen for TFIID-dominated genes. We have modified Figure 5 to use the replicate experiment that shows intermediate changes in signal intensity, and make a note of this variability in the figure legend of the revised manuscript. Finally, only single replicates had been performed for Med15 ChIP using kin28-rpb3-AA and kin28-taf1-AA yeast. We have repeated these experiments and obtained consistent results in kin28-rpb3-AA yeast; however, we have seen stronger signals for Med15 in kin28-taf1-AA yeast in the presence of rapamycin than in the experiment, which was used for Figure 5 in the critiqued version (see Author response image 2). Again, we see an upstream shift upon depletion only with TBP and not with Taf1 or Rpb3, and strong decrease in signal at TFIID-dominated genes (see revised Figure 5). We have replaced the data for Med15 ChIP in kin28-taf1-AA in Figure 5 with data from one of the new replicates (the “run40” replicate shown in Author response image 2), to reflect the stronger signal we see for kin28-taf1-AAR relative to kin28-AAR in the new replicates. We have not used combined data for Figure 5 (except as noted above), as the signal to noise and number of reads has varied among replicates, and the combined data is dominated by the “cleanest” data. Information on replicate experiments provided as Supplementary file 3 has been updated to reflect these additional replicate experiments.

Author response image 2

Download asset Open asset

The reviewer also states that experiments are normally done in duplicate (at least) and some correlation metrics used to insure they are alike. We have found that use of correlation coefficients between ChIP-seq datasets is not informative, as variance in regions outside of the peaks that we focus on (the large majority of the genome) obscures the similarities or differences among the regions of interest in terms of quantitative metrics. Rather, we have used metagene analysis, as in Figures 3 and 5 and the Author response image 1 and 2 to assess similarity between replicate experiments.

Another major concern with the manuscript has to do with the Taf1 anchor-away data. As acknowledged by the authors, the taf1-FRB strain appears to show a phenotype even in the absence of rapamycin. This means that the tagging affected the function of the protein. Because rapamycin did not lead to any addition effect, I do not think that the data can be interpreted as a consequence of conditional depletion. Unless the authors find a way to conditionally affect the function of Taf1 (perhaps auxin degron or tagging FRB in N-terminal would work better), I believe the Taf1 data has to be removed.

The reviewer states that “addition of rapamycin did not lead to any additional effect”, but this is not correct: although adding rapamycin did not affect the ChIP signal for Med15 or Med18 (Figure 3), it did cause lethality (Figure 3—figure supplement 1) and reduced Taf1 occupancy (Figure 3—figure supplement 3 in the revised manuscript); furthermore, anchoring away Taf1 together with Kin28 suppresses the effect of anchoring away Kin28 at TFIID-dominated genes (Figure 5). We therefore believe that the data presented support the main conclusion of experiments on the taf1-AA strain, which is that Mediator association with UAS regions does not decrease upon depletion of Taf1, and therefore that Taf1 is not required for Mediator recruitment to UAS regions. Furthermore, we have observed in three independent experiments that depletion of Taf1 together with Kin28 results in preferential decrease in Mediator occupancy at TFIID-dominated genes. However, the effect of this double depletion on Med15 occupancy at SAGA-dominated genes has been variable in our experiments, as mentioned above, and so we have tempered our conclusions regarding this specific effect.

In sum, I recommend the authors go back to the basics and make sure their experiments are reproducible. For that, they need to show that their replicates are highly correlated. At current state, I suspect that most of the effects observed are batch effects rather than biological effects. For instance, simply inspecting the red tracks in Figure 3—figure supplement 5A clearly shows that rep2 did not work nearly as well as rep1. Not seeing the labels, one could not even tell how to pair them. Also, looking at the blue tracks from the same Figures suggests that the samples cluster by batches rather than by conditions.

We have revised this Figure (now Figure 3—figure supplement 4) so that tracks are normalized across samples for each strain (e.g., all four tbp-AA strain scans are normalized) rather than only between samples within a single experiment (e.g. tbp-AA-1 and tbp-AAR-1 samples were normalized relative to one another but separately from tbp-AA-2 and tbp-AAR-2 samples). This shows more clearly the reduced signal of TBP upon addition of rapamycin.

[Editors' note: further revisions were requested prior to acceptance, as described below.]

Reviewer #2:

The revised manuscript is improved, notably by the addition of datasets completing some that were done only once and by the replacement of non-consistent replicates. However, there are a few issues that are still of concern and should be addressed.

First, some of the anchor-away strains behave strangely. While this is unlikely to drive erroneous conclusions, it should at least be mentioned. For instance, Kin28-TBP-AA Med15-Myc is very sick according to Figure 3—figure supplement 1. More worrisome, tagging various PIC components with FRB greatly affects Mediator ChIP in the absence of rapamycin. This is evident by looking at several figures, notably in Figure 3 (pay attention to the y-axis for all the no rapa traces (black) or look at the intensities of the no rapa heatmaps). This is also directly shown in Figure 3—figure supplement 2B. The authors use this figure supplement to claim that Taf1-AA leads to increased Mediator ChIP signal in absence of rapamycin but they fail to acknowledge that this is also the case for TBP-AA (the average signal is 2 to 3 fold up compared to non-tag). It is not clear whether these differences are batch differences (due to technical variation between experiments performed on different days) or biological differences (caused by partial loss of function of tagged proteins).

We have edited the text to note the effect of the tbp-FRB allele on Mediator occupancy and the slow growth phenotype of the kin28-tbp-AA Med15-myc strain (Results section). The elevated Med15 occupancy observed in taf1-AA yeast in the absence of rapamycin was seen in two out of two replicate experiments, while lower levels of Med15 occupancy, similar to the control lacking any FRB tag, was seen in two out of two experiments with rpb3-AA yeast and two out of three experiments in tbp-AA yeast. One experiment showed elevated Med15 occupancy in tbp-AA yeast in the absence of rapamycin; that is the experiment that led to the request for an additional replicate, which showed low occupancy in the absence of rapamycin in tbp-AA yeast that increased upon rapamycin treatment. We have also edited Figure 3—figure supplement 2 by replacing the data for the “no FRB tag” control with data from a newer replicate (from October this year). The original data had higher levels of artifactual signal over ORF regions (as we and others often see), which made it more difficult to visually compare signal strength over the UAS regions.

Second, the Med18 ChIP data displayed in Figure 3 and showing that Med18 does not change in TBP-AAR or Rpb3-AAR does not fit the author's model very well. In subsection “Mediator association at UAS regions is stabilized by loss of TBP or Pol II” the authors say "The increase in ChIP signal observed for Med15, but not for Med18, would be consistent with the most immediate contacts with UAS-bound activators occurring through tail module subunits (such as Med15), while ChIP signal for Med18 at UAS regions would depend on protein-protein cross-links, thus lowering the efficiency of ChIP for this subunit". While I agree that this explains the lower ChIP efficiency of Med18 at UAS, it does not explain why, unlike other subunits tested, Med18 signal does not increase at UAS in TBP-AAR and Rpb3-AAR. The observed Med18 ChIP signal is clearly over the UAS so if Mediator is stuck there, this Med18 signal should increase together with other subunits. The fact that it does not (assuming the data is real), suggests a more complicated model. This result should at least be described more critically.

We agree with this criticism, and have modified the text accordingly. We did not make our point clearly, which is that under some circumstances, such as depletion of Kin28 (Figure 3—figure supplement 4B, and also Jeronimo et al., 2016), Med15 can be detected at UAS regions while Med18 gives rise to virtually no ChIP signal at those sites; in the same sample, signals from both Med15 and Med18 are easily discerned at the proximal promoter. This suggests that Mediator may adopt a configuration in which Med15 signal is robust and Med18 undetectable, or nearly so; such a configuration could result in increased Med15 signal with no increase in Med18 signal. The way we presented this before was poorly constructed at best, and we thank the reviewer for noticing this.

Third, the conclusion that TBP is required for the transition while Rpb3 and Taf1 are stabilising the PIC-associated Mediator is in my opinion a hazardous extrapolation. As explained by the authors in subsection “TBP is required for transit of Mediator from UAS to promoter”, depletion of PIC components has impact on other PIC components. This makes the attribution of specific roles to specific PIC components in Mediator dynamics difficult. Hence, with these limitations in mind, it seems like an over interpretation to say "TBP being critical for transit of Mediator from UAS to promoter, while Pol II and Taf1 stabilize Mediator association at proximal promoters". A more conservative interpretation would be that the PIC is necessary to allow Mediator transiting to the promoter (as shown when PIC is severely compromised; TBP-AA) and that partially formed PICs (without trying to attribute specific roles to subunits or proteins), while allowing the transit, lead to unstable Mediator-PIC association.

These last two points had been raised in previous review rounds but are still not clearly dealt with in the most recent version of the manuscript.

We believe that our statement that TBP is critical for transit of Mediator from UAS to promoter is logically accurate. It does not imply that other factors, in this case other PIC components that depend on TBP for stable association with promoters do not contribute. As we state in subsection “TBP is required for transit of Mediator from UAS to promoter”, we can infer that TBP is necessary for Mediator transit to promoter regions under conditions of Kin28 depletion, but not whether it is sufficient. In contrast, the evidence presented indicates that under conditions of Kin28 depletion, neither Taf1 nor Pol II are critical for this transit. However, the reviewer’s summary statement does provide a succinct summary of these points, and we have edited the text to borrow his/her phrasing (hopefully with permission).

Finally, it is not clear why the authors were expecting a decrease of Mediator at UAS upon depletion of TBP, Rpb3, or Taf1. Isn't an increase what the "dynamic model" proposed by the Struhl and Robert labs predicted?

We had initially anticipated that PIC depletion might reduce Mediator occupancy at both promoters and UAS regions. This could occur, for example, if Mediator contacted both UAS and promoter via DNA looping and reduced PIC occupancy destabilized Mediator occupancy at both regions. This was evidently not the right way to think about it, but it still seems to me not a totally far-fetched possibility in the absence of data. We have edited the text to clarify this point. Clearly what we anticipated was not as important in any event as what we observed.

Author details

1. Elisabeth R Knoll

   Department of Biomedical Sciences, School of Public Health, University at Albany, Albany, United States

   Contribution

   Conceptualization, Formal analysis, Validation, Investigation, Visualization, Writing—original draft, Writing—review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-1083-6472

2. Z Iris Zhu

   Computational Biology Branch, National Center for Biotechnology Information, National Library of Medicine, Bethesda, United States

   Contribution

   Data curation, Formal analysis, Visualization, Writing—original draft, Writing—review and editing

   Competing interests

   No competing interests declared

3. Debasish Sarkar

   Wadsworth Center, New York State Department of Health, Albany, United States

   Contribution

   Validation, Investigation

   Competing interests

   No competing interests declared

4. David Landsman

   Computational Biology Branch, National Center for Biotechnology Information, National Library of Medicine, Bethesda, United States

   Contribution

   Formal analysis, Supervision, Funding acquisition, Writing—review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-9819-6675

5. Randall H Morse

   1. Department of Biomedical Sciences, School of Public Health, University at Albany, Albany, United States
   2. Wadsworth Center, New York State Department of Health, Albany, United States

   Contribution

   Conceptualization, Formal analysis, Supervision, Funding acquisition, Validation, Investigation, Writing—original draft, Writing—review and editing

   For correspondence

   randall.morse@health.ny.gov

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-0000-8718

• 2,888

  Page views

• 454

  Downloads

• 22

  Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.