The spt5 C-terminal region recruits yeast 3' RNA cleavage factor I

Abstract

During transcription elongation, RNA polymerase II (Pol II) binds the general elongation factor Spt5. Spt5 contains a repetitive C-terminal region (CTR) that is required for cotranscriptional recruitment of the Paf1 complex (D. L. Lindstrom et al., Mol. Cell. Biol. 23:1368-1378, 2003; Z. Zhang, J. Fu, and D. S. Gilmour, Genes Dev. 19:1572-1580, 2005). Here we report a new role of the Spt5 CTR in the recruitment of 3' RNA-processing factors. Chromatin immunoprecipitation (ChIP) revealed that the Spt5 CTR is required for normal recruitment of pre-mRNA cleavage factor I (CFI) to the 3' ends of Saccharomyces cerevisiae genes. RNA contributes to CFI recruitment, as RNase treatment prior to ChIP further decreases CFI ChIP signals. Genome-wide ChIP profiling detected occupancy peaks of CFI subunits around 100 nucleotides downstream of the polyadenylation (pA) sites of genes. CFI recruitment to this defined region may result from simultaneous binding to the Spt5 CTR, to nascent RNA containing the pA sequence, and to the elongating Pol II isoform that is phosphorylated at serine 2 (S2) residues in its C-terminal domain (CTD). Consistent with this model, the CTR interacts with CFI in vitro but is not required for pA site recognition and transcription termination in vivo.

Fig 1

Spt5 CTR deletion leads to a growth defect in the presence of 6-AU and to a slight increase in Spt5 protein levels. (A) The Spt5 ΔCTR mutant shows a strong growth defect in the presence of 50 μg/ml of 6-AU compared to the wild type. No effects are observed at 30°C or 37°C (data not shown) or in the presence of 15 μg/ml of mycophenolic acid (MPA) on solid medium. (B) The Spt5 ΔCTR mutant grown in liquid YPD medium shows a slight growth defect at 30°C compared to the wild type. The standard deviations for three independent measurements are indicated for each data point. (C) Spt5 protein levels in the ΔCTR mutant are upregulated 1.7-fold compared to those of the wild type. Quantitative Western blotting was performed with antibodies against Spt5 and α-tubulin. Different amounts of protein were loaded, and quantification of the intensities was performed. Normalization of the Spt5/α-tubulin ratio obtained for Spt5 ΔCTR cells against the corresponding ratio of wild-type cells revealed that Spt5 protein levels were slightly elevated (1.7-fold) in cells lacking the CTR of Spt5.

Fig 2

ChIP analysis reveals that CFI occupancy is reduced in Spt5 ΔCTR cells. This is true for CFIA subunits (A), except for Pcf11, as well as for CFIB/Hrp1 (B). (C) Whereas Paf1 occupancy, as determined by ChIP, is strongly reduced in Spt5 ΔCTR cells, Pol II levels are not affected. (D) Pap1 occupancy is not changed in Spt5 ΔCTR cells. The fold enrichments at the ADH1 gene over a nontranscribed region that is located near the centromere of chromosome V are given for the TSS, the ORF region, the pA site, and the region 3′ of the pA site. ChIP-determined occupancies are indicated for wild-type and Spt5 ΔCTR cells as black and gray bars, respectively. The standard deviations are for at least two independent ChIP experiments.

Fig 3

The Spt5 CTR interacts with CFI subunits in vitro. GST pulldown experiments were performed with a GST-Spt5 CTR fusion protein (GST-CTR), with GST alone (GST), and without protein (−), which served as a negative control. Western blotting was performed for the last washing fractions (W), the combined elution fractions (E) of the samples, and 1% of the Rna14-TAP (first panel), Rna15-TAP (second panel), and Hrp1-TAP (third panel) yeast cell lysates (Input) with antibodies against the TAP tag of the corresponding CFI subunit and the GST tag (for details, see Materials and Methods). The signals obtained for GST and GST-CTR were similar and are exemplarily shown for the pulldown experiments with Rna14-TAP (fourth panel).

Fig 4

RNase-ChIP assays reveal that RNA contributes to CFI recruitment in Spt5 ΔCTR cells. RNA-dependent binding is given for the ORF region, the pA site, and a region 3′ of the pA site of the ADH1 gene. The ChIP occupancy signal of Rna15 without RNase treatment was set to 100%. The relative ChIP signals of Rna15 for wild-type and Spt5 ΔCTR cells after RNase treatment are indicated. The corresponding percentages are given above the bars. Standard deviations were calculated from four independent experiments.

Fig 5

Genome-wide ChIP-chip occupancy profiling of CFI subunits in yeast. (A) Gene-averaged profiles for the long-gene-length class (2,350 ± 750 nt, 299 genes; see Materials and Methods) for Pcf11 (52), Rna14, and Rna15. Profiles of other length classes are generally similar (data not shown). Dashed black lines indicate the TSS and pA site. (B) Gene-averaged profiles as for panel A for the transcription elongation factor Spt5 (52) and the S2-phosphorylated (S2P) CTD form of Pol II (52). Occupancies and signal intensities are given for Spt5 and the S2-phosphorylated form of Pol II on the left and right y axes, respectively.

Fig 6

Spt5 CTR deletion provokes neither transcriptional readthrough of Pol II nor alternative pA site usage. (A) Schematic representation of the yeast PMA1 locus. The ORF region and the two pA sites according to Ahn et al. (3) are indicated by a box and vertical arrows, respectively. The forward primer (fw) and two reverse primers (1-rev and 2-rev) that were used for Pol II readthrough detection are depicted as horizontal arrows. (B) Agarose gel electrophoresis of the five PCR products as described for panel A for the wild type (wt), Spt5 ΔCTR cells (ΔCTR), and the rna14-1 temperature-sensitive strain grown at permissive (24°C) and restrictive (37°C) temperatures. The rna14-1 mutant led to a readthrough transcript at 37°C (1.5 kb; black arrow) and served as a positive control. No differences in the lengths of the PCR products could be detected between wild-type and Spt5 ΔCTR cells. The unspecific PCR product also obtained with the second reverse primer (2-rev) is marked by an asterisk. The heights of the marker lanes in base pairs (bp) are shown on lane 1. (C) Gene-averaged occupancy profiles as in Fig. 5 but for the medium-gene-length class (1,238 ± 300 nt, 339 genes; see Materials and Methods) of Pol II (Rpb3) in wild-type and Spt5 ΔCTR cells. (D) The nucleotide sequence of the 3′ region of yeast ACT1 is shown. Key sequence elements are labeled. 3′-RACE revealed three major RNA cleavage and pA sites (red-filled boxes), site 1 (1,506 nt downstream of start codon ATG; seven and three sequenced clones for the wild-type and Spt5 ΔCTR mutant, respectively), site 2 (1,534 nt downstream of start codon ATG; four and nine sequenced clones, respectively), and site 3 (1,538 nt downstream of start codon ATG; six and five sequenced clones, respectively), and three minor pA sites (red frames), site 1 (1,509 nt downstream of the start codon ATG; one and two sequenced clones for the wild type and Spt5 ΔCTR mutant, respectively) and sites 2 and 3 (1,551 and 1,568 nt downstream of the start codon ATG; two and one sequenced clones, respectively) that were used equally in wild-type and Spt5 ΔCTR cells. Additionally, 3′-RACE led to the mapping of three rare pA sites that were used exclusively in Spt5 ΔCTR cells (yellow boxes; one sequenced clone for each site). For further details, see Materials and Methods.

Fig 7

Model of CFI recruitment in yeast. The complete yeast Pol II elongation complex with bound Spt4/5 is viewed from the back (51). Pol II and Spt4/5 are shown as molecular surfaces with key domains highlighted in color and labeled. Exiting RNA, the C-terminal KOW domains, the CTR of Spt5, and the Pol II CTD extend from Pol II around the Rpb4/7 subcomplex, establishing a main interface for CFI recruitment. Rna14 may directly contact the Spt5 CTR, whereas the RNA is bound by the C-terminal RRM domain of Rna15 and by two internal RRM domains of Hrp1. The Pol II CTD is bound by the N-terminal CID domain of Pcf11. CFI subunits are drawn to scale. Important protein domains are illustrated as extensions from the protein cores.