Mass spectrometry of TFIIB purified from soluble cell lysate and chromatin fraction
The mechanism underlying the role of TFIIB in initiation of transcription is well established. TFIIB enters the preinitiation complex (PIC) after TFIID and facilitates recruitment of RNAPII onto the promoter1,3,4. TFIIB interacts with the promoter region, just downstream of the TFIID binding site31–33. TFIIB also contacts TFIID, TFIIF and RNAPII subunits in the PIC34–40. How the factor performs its newly identified functions in termination, promoter-terminator interaction, and promoter directionality, however, remained unclear. To explore the comprehensive role of TFIIB in the transcription cycle, we performed purification of HA-tagged TFIIB from yeast cells. Unlike past attempts, where TFIIB was purified from soluble nuclear or cell lysate, we attempted purification from chromatin as well as soluble fractions. Exponentially growing yeast cells were lysed and separated into soluble and chromatin fractions by differential centrifugation following the protocol adapted from 41 and 42 (Fig. 1A). Authenticity of the soluble and chromatin fractions was confirmed using marker proteins a-tubulin and histone H3 as described in 42. a-tubulin was detected only in the soluble fraction, while histone H3 was exclusively localized in the chromatin fraction of the cell lysate (Fig. 1B). Chromatin-bound proteins were then eluted using high ionic strength buffer (0.5 M ammonium sulfate) as described in42. Epitope-tagged TFIIB from both soluble and chromatin derived fractions was affinity purified on anti-HA-magnetic beads and subjected to tandem mass spectrometry (MS/MS). The soluble fraction contains the form of TFIIB that is not engaged in transcription, while chromatin eluate harbors the transcriptionally active form of the factor. Simultaneously, purification was performed from the untagged strain so that background noise signal can be identified.
We followed a quantitative proteomic approach to understand the functional significance of TFIIB interacting proteins identified by mass spectrometry. Parameters used in these analyses are protein threshold of 1% FDR and peptide threshold of 0.1% FDR, displayed as normalized total spectra. We considered only those proteins in our analysis for which the probability of spectral counts match was greater than 95%. Relative abundance of proteins present in affinity purified TFIIB preparations was determined by calculating spectral abundance factor of the protein with respect to that of TFIIB as described in 43,44. First, the number of spectral counts for each protein was divided by its molecular mass to get the spectral abundance factor (SAF). Next, the SAF background signal from the untagged strain was deducted. The relative abundance of a protein in a purified TFIIB sample was then quantified by dividing the SAF value of the protein with that of TFIIB to get the TFIIB-normalized spectral abundance factor (BNSAF).
The data presented here is the result of four independent replicates and is presented in Supplementary Data 1. We considered only those proteins in our analyses whose BNSAF value was greater than 0.1, and which were consistently detected. In total, we identified 642 proteins copurifying with TFIIB isolated from the chromatin fractions, and 94 proteins copurifying with TFIIB from the soluble fraction on affinity column. Of these, 618 proteins were unique to the chromatin fraction, while 70 were unique to the soluble fraction. 24 proteins were present in both fractions (Fig. 1C). We first identified top interactors of TFIIB exhibiting a BNSAF value of greater than 2.0. In TFIIB-soluble preparation, no protein with BNSAF value of even > 0.5 could be detected. In TFIIB-chromatin preparation, however, there were a number of proteins with BNSAF value of 2.0 or more (Fig. 1D). None of these proteins were associated, either directly or indirectly, with transcription or mRNA processing. A majority of these proteins were linked to heat shock response with molecular chaperone function. Whether TFIIB has a hitherto undiscovered role in helping cells overcome high temperature shock, or these proteins simply facilitate proper folding of TFIIB by their chaperone activity needs further investigation.
RNAPII, TFIIF and TFIIH are the only PIC components detected in the chromatin eluted TFIIB complex
Since TFIIB is a component of the PIC in a chromatin context (Fig. 2A), we next looked for the presence of general transcription factors and RNAPII in TFIIB preparations. There was no statistically significant signal for either RNAPII subunits or any of the general transcription factors in TFIIB purified from soluble fraction (Fig. 2B and 2C, dark blue bars). This is in accordance with previous purification efforts that also did not detect any RNAPII, GTFs or Mediator complex in TFIIB purified from soluble nuclear or cell lysate26,28−30.
There were, however, multiple RNAPII subunits in the TFIIB-chromatin preparation. Rpb1, Rpb2, Rpb3, Rpb8, Rpb11, Rpb4 and Rpb7 were the subunits that exhibited significant enrichment in TFIIB-chromatin (Fig. 2B, light blue bars). The primary structure of TFIIB has an N-terminal domain comprising of a Zn-ribbon, B-finger and linker regions, and a C-terminal domain with two almost identical cyclin fold repeats1. In the three-dimensional structure of the RNAPII-TFIIB complex, the B-finger protrudes into the active center, while the cyclin repeat interacts with the ‘protrusion’ and ‘wall’34–36, 39,45. Our proteomic data show that TFIIB makes strong contact with Rpb1 and Rpb2. Since both of these subunits form the heart of the catalytic center of the polymerase, our data agrees with three-dimensional structural studies. Furthermore, the high-resolution structure shows cyclin repeats making contact with the ‘protrusion’ and ‘wall’ of enzyme. This is also in conformity with the proteomic data which shows interaction of TFIIB with Rpb3 and Rpb11, both of which are part of a ‘protrusion’ located on the surface of the enzyme away from the catalytic center35,36,39. Rpb4 and Rpb7, which are not components of the 10-subunit core enzyme, were also significantly enriched in chromatin eluted TFIIB. Of the general transcription factors, TFIIF subunits Tfg1 and Tfg2 were consistently detected in the TFIIB-chromatin preparation (Fig. 2C, light blue bars). Four subunits of TFIIH; Tfb1, Tfb2, Rad3 and Ssl2 were also present, though to a lesser extent (Fig. 2C, light blue bars). TFIIA, TFIID and TFIIE subunits were not consistently detected and if present exhibited a BNSAF value of < 0.05. Mediator complex was also altogether absent from both TFIIB preparations.
3’ End Processing-termination Factors Are Exclusively Present In The Chromatin Eluted Tfiib Complex
Studies during the last couple of decades have implicated TFIIB in termination of transcription5,8,9. TFIIB also facilitates gene loop formation, possibly through interaction of the promoter-bound molecule with the termination factors occupying the 3’ end of the gene5,9,14. These studies gave rise to the speculation that TFIIB interacts with the 3’ end processing-termination factors during transcription. In principle, TFIIB-termination factor interaction can explain both its role in termination as well as gene loop formation. In yeast, 3’ end processing-termination is accomplished by three multiprotein complexes: CF1, CPF and Rat146,47 (Fig. 3A). We therefore looked for the presence of all three termination complexes in soluble and chromatin derived TFIIB preparations. TFIIB-chromatin exhibited significant enrichment for CF1 and Rat1 complex subunits. All four subunits of CF1A complex; Rna14, Rna15, Pcf11 and Clp1 consistently displayed BNSAF values ranging from 0.2–0.37 in TFIIB-chromatin preparation (Fig. 3B, light blue bars). Hrp1, which is the only subunit of the CF1B complex, was also detected in TFIIB-chromatin preparation (Fig. 3B). All three subunits of the Rat1 complex; Rat1, Rtt103 and Rai1, exhibited a statistically significant enrichment in in the chromatin derived TFIIB (Fig. 3D, light blue bars). Of CPF complex subunits, only Glc7, Cft1, Cft2, Pta1, Ssu72 and Swd2 were present in the affinity purified TFIIB-chromatin (Fig. 3C, light blue bars). Glc7 was also present in TFIIB-soluble preparation, and there was no statistically significant enrichment for the factor in the chromatin-derived preparation (Fig. 3C).
Interaction Of Splicing Factors With Tfiib
Since splicing is a cotranscriptional process and a splicing-competent intron has been shown to enhance transcription by facilitating recruitment of general transcription factors including TFIIB on the promoter48, we examined the presence of splicing factors in purified TFIIB. We consistently detected the presence of three splicing factors: Prp19, Prp43 and Sub2 in affinity purified TFIIB (Fig. 4A). Of these, only Sub2 exhibited a statistically significant enrichment in the chromatin-derived preparation (Fig. 4A, light blue bars). Prp43 was present in both soluble and chromatin associated TFIIB (Fig. 4A).
Lsm Complex And Arp2/3 Complex Associates With Tfiib In The Chromatin Environment
Synthesis of mRNAs in the nucleus is often coupled to their degradation in the cytoplasm49. This results in unstable mRNAs being transcribed more efficiently. Coupling of synthesis and decay is through RNA decay complexes or decaysomes. Lsm complex is one such decaysome complex pathway50–55. It consists of seven Lsm proteins; Lsm1, Lsm2, Lsm3, Lsm4, Lsm5, Lsm6 and Lsm7 forming a heteroheptameric complex along with Pat1, which is a decapping enzyme and 5’ to 3’ exoribonuclease Xrn150,56−58. The Lsm complex degrades mRNA in the cytoplasm and stimulates transcription of the same mRNA species in the nucleus59. The capability of the Lsm complex to stimulate both degradation as well as synthesis of the same mRNA species is dependent on its ability to shuttle between the cytoplasm and nucleus. In the nucleus, the Lsm complex directly stimulates transcription by binding to the promoter region about 30 bp upstream of the transcription start site and affects the initiation and elongation steps of transcription59. The ability of the Lsm complex to stimulate initiation of transcription may be dependent on its interaction with the PIC components. We therefore looked for the presence of the Lsm complex in affinity purified TFIIB preparations. Our results show that the TFIIB-chromatin preparation is enriched in Lsm1, Lsm2, Lsm4, Lsm6 and Xrn1 subunits of the complex (Fig. 4B, light blue bars). Dcp2 and Pat1 were also detected with low BNSAF value. The Lsm complex associates with TFIIB only in the chromatin context as soluble TFIIB did not exhibit enrichment for any component of the complex (Fig. 4B, dark blue bars). Whether the Lsm complex facilitates recruitment of TFIIB and other components of the PIC on the promoter or it helps in release of RNAPII from the promoter for elongation needs further investigation.
Actin related proteins (Arps), being components of chromatin modifying complexes, have been implicated in transcription both in yeast and higher eukaryotes60. Arp7 and Arp9 are components of the SWI/SNF complex; Arp4, Arp5 and Arp8 of the Ino80 complex, while Arp4 is associated with the NuA4 complex in budding yeast61. We therefore examined the presence of Arp4, Arp5, Arp7, Arp8 and Arp9 in our purified TFIIB preparations. We could not detect a significant presence of any of these Arps in either soluble or chromatin-TFIIB preparations. Instead, we detected the heteroheptameric Arp2/3 complex in purified TFIIB from chromatin (Fig. 4C). The Arp2/3 complex consists of seven subunits, two of which Arp2 and Arp3 are actin-related proteins, while the remaining six; Arc1, Arc15, Arc18, Arc19, Arc35 and Arc50 are non-actin related proteins62. The complex nucleates formation of branched actin filaments in yeast63. The complex associates with RNAPII in vivo and has been implicated in transcription by RNAPII both under in vitro and in vivo conditions in HeLa cells64. The presence of the Arp2/3 complex exclusively in chromatin eluted TFIIB strongly suggested that the complex may be playing a role like its mammalian counterpart in transcription in yeast as well. We therefore checked if Arp2 and Arp3 interact with RNAPII in the chromatin context as has been reported in HeLa cells. Both Arp subunits were HA-tagged at the C-terminus and their association with the Rpb1 subunit of RNAPII was examined by coimmunoprecipitation. Our results show that the chromatin linked Arp2 (Fig. 4D, lane 4), but not the soluble Arp2 (Fig. 4D, lane 2), exhibits interaction with RNAPII. A similar association of Arp3 with RNAPII was observed exclusively in the chromatin context (Fig. 4E, lane 4). These results strongly suggest that the Arp2/3 complex has a novel role in transcription in budding yeast as well.
Tfiib Interactions With Factors In The Chromatin Context Are Not Mediated By Dna Or Rna
Exclusive interaction of TFIIB-chromatin with RNAPII, CF1 complex, Rat1 complex, Lsm complex and Arp2/3 complex gave rise to the speculation that the interaction of the factor with these complexes may be indirect, being mediated by the template DNA or transcribing mRNA. To examine if the observed TFIIB interactions are direct or indirect, we digested the chromatin eluate with MNase before performing affinity chromatography. Affinity purification, mass spectrometry and statistical analyses was performed as described above. After MNase digestion, interaction of TFIIB with all four subunits of the CF1A termination complex and with subunits of the CPF complex except for Glc7 remained unaffected (Table 1C and Supplementary Data 2). In contrast, the Rat1 subunit of the Rat1 complex registered a statistically significant drop upon nuclease digestion (Table 1C). Of PIC components, interaction of TFIIB with Tfb1 and Tfg1 exhibited a decline, but the most drastic affect was with RNAPII subunits (Table 1A). Apart from Rpb3, Rpb4 and Rpb7, TFIIB interaction with other subunits including Rpb1 and Rpb2 was completely abrogated. Overall, however, a majority of interactions of chromatin associated TFIIB were maintained even after MNase digestion (Table 1A, 1B, 1C, 1D and 1E).