Figure 1.
Effects of meiosis, mmi1, and RNA processing mutants on transcripts.
Hierarchical clustering of microarray data for 50 genes that accumulate most in mmi1-ts3 is shown for a meiotic time-course and various mutants. pab2Δ data are from [8]. Other strains were grown vegetatively and conditional mutants were shifted to restrictive temperature. Colors represent the log2 ratio of transcript levels in the mutant strain/wild-type control such that red means the transcript accumulates in the mutant. The 31-gene “Mmi1 regulon” is colored red in the dendogram. The independently-derived classification of genes with respect to meiotic timing is taken from [57]: yellow = early meiotic, orange = mid meiotic, red = late meiotic, green = nitrogen starvation and grey = non-meiotic. * Ubi4 and SPBC19F8.04c are probably not genuine Mmi1-targets (see text). Two genes, mek1 and meu13, that may be real Mmi1-targets [3], [8] but are not in our defined Mmi1-regulon are attached at the bottom of the cluster.
Figure 2.
The sequence logo of the MEME predicted Mmi1 binding motif is shown on top left. The distribution of this motif (red bars) in the 29 Mmi1 regulated genes is drawn to scale.
Table 1.
The effects of Mmi1 on 7 meiotic genes.
Figure 3.
Splicing and stability of the rec8 transcript.
(A) Splicing of endogenous rec8 and crs1. Gene structures of rec8 and crs1 are drawn to scale. Box, exon; Black line, constitutively spliced intron; red line, regulated intron; grey line, 200 nt 3' UTR; red bar, Mmi1 motif. The three introns clustered at the 5' end of rec8 or the 3'-most 4th intron were assayed with 5' or 3' primer pairs, respectively. All four introns of crs1 were assay with one primer pair. The dpb3 control indicates equal loading and the -RT control (no reverse transcriptase) shows the absence of contaminating genomic DNA. (B) Splicing of rec8 variants. A rec8Δ (“WT”) or rec8Δ mmi1-ts3 (“mmi1-ts3”) strain were transformed with plasmids bearing the indicated rec8 constructs. Splicing of the 4th intron was assayed by RT-PCR. Top, wild type rec8+, middle, rec8 in which 4th intron has been moved 1140 bp 5' to be adjacent to the three unregulated introns (rec8 int4@638), and bottom rec8 in which 4th intron is replaced by the 4th intron from cdc2 (rec8 int4cdc2). Results from two transformants are shown. (C) qPCR (triplicate assays for two transformants) of rec8 mRNA levels in a rec8Δ (“WT”) transformed with plasmids expressing wild type rec8+, rec8 with the 4th intron deleted (int4Δ, mimicking spliced rec8), and rec8 in which the branch point sequence of 4th intron is mutated so that splicing is blocked (BPmut, mimicking unspliced rec8). For comparison, rec8 levels from the rec8+ plasmid are shown in a rec8Δ mmi1-ts3 strain. Error bar represents SEM.
Figure 4.
3' end processing assays of rec8.
A PCR-based RACE-PAT assay shows the cleavage sites and the amount of polyadenylated rec8 in WT vegetative, meiotic, and mutant strains. This assay does not reflect the length of the polyA tail because most PCR products collapse to the minimum length. Two major cleavage and polyadenylation sites were seen, PAS1 and PAS2 (arrow heads). Each band was sequenced to confirm the assignments. A read-through assay using primers flanking PAS1 and PAS2 (light arrows) yields product representing read-through transcripts (i.e., transcripts progressing past both PAS1 and PAS2). dpb3 is the internal control and -RT shows the absence of contaminating genomic DNA.
Figure 5.
Northern blot analysis of transcript size and polyA tail length.
(A) Total RNA was isolated from WT or mutant cells in vegetative growth or from 4 hours in meiosis. RNA was treated with RNase H in the absence (-) or presence (+) of oligo d(T) to cleave off the polyA tail. RNAs (rec8, and adh1 as a control) were then analyzed by strand specific Northern blot. Hybridization to srp7 and ethidium bromide staining of rRNAs are shown below each blot as loading and RNA quality controls. (B) Gene structure of three Mmi1 regulated genes, rec8, ssm4 and mei4, are drawn to scale. The red line above each gene indicates the mapped DSR sequence [3] and the red bars under the gene indicate the putative Mmi1 motif. RNA was isolated from the rrp6–9 mutant and subjected to RNase H treatment in the absence (−) or presence (+) of oligo d(T). (C) Analysis of two Rrp6-responsive genes that are independent of Mmi1. These are SPAC16E8.16, which encodes TFIIB, and SPCC1442.04c. A blank space in the TFIIB panel indicates removal of irrelevant lanes.
Figure 6.
Analysis of ribozyme constructs.
A hammerhead ribozyme was placed downstream of rec8. (A) Illustration of rec8-A65RZ and rec8-A65RZ ΔPAS constructs showing the 3' region of rec8 (boxes show exons, 4th intron is a line), the DNA-encoded polyA tail, the ribozyme, the DSR (red line), and PAS1 and PAS2. For rec8-A65RZ, 65 A residues (“A65”) were inserted 91 nt 3' of PAS2, immediately followed by the hammerhead ribozyme (called RZ and shown as a stem loop). Scissors show the ribozyme self-cleavage site. rec8-A65RZmut is identical except for a point mutation that disrupts ribozyme activity (red cross). Arrows show primers used to detect read-through transcripts. The PAS region (green line) is deleted in the rec8-A65RZ ΔPAS construct. (B) Read-through assay on rec8-A65RZ (+) and rec8-A65RZmut (-) to measure ribozyme activity and rec8 3' end cleavage at PAS1 and PAS2. Top panel: read-through assay with primers across the ribozyme sequence. Middle panel: total rec8 measured by primers within the rec8 ORF. Bottom panel: minus reverse transcriptase control. (C) RACE-PAT assay to determine the cleavage sites of polyadenylated transcripts. Left panel: assay on endogenous rec8+ in vegetative and meiotic cells. Transcripts ending at the two major polyadenylation sites are marked PAS1 and PAS2. Middle panel: assay on transcripts of rec8-A65RZ in rec8Δ strains bearing various mutations. Ribozyme-generated transcripts are marked RZ end. Right panel: assay on transcripts of rec8-A65RZ ΔPAS in rec8Δ strains bearing various mutations. Ribozyme-generated transcripts are marked RZ ΔPAS end. RZ end and RZ ΔPAS end bands were confirmed by sequencing. (D) Northern blot analysis of rec8-A65RZ ΔPAS in rec8Δ strains. Upper panel: level of rec8 transcripts from rec8-A65RZ ΔPAS. Middle panel: LEU2 (expressed from the plasmid) used as a normalization control. Lower panel: ethidium stained rRNAs. rec8-A65RZ ΔPAS to LEU2 ratios are the average of two experiments using independent transformants. (E) Analysis of rec8-AnRZ ΔPAS constructs with different lengths of encoded polyA sequence, A0, A65 and A110, in rec8Δ WT cells. (F) The ratio of each transcript level in the rec8Δ mmi1-ts3 mutant to the rec8Δ WT and in the rec8Δ rrp6–9 mutant to the rec8Δ WT. Relative transcript levels are shown on the Y-axis as determined by Q-PCR in triplicate for three transformants of each strain. Error bar represents SEM. (G) Splicing assay on the 4th intron of rec8-A65RZ ΔPAS in rec8Δ WT and rec8Δ mmi1-ts3 strains. Results of two independent transformants are shown. Loading control dpb3 and –RT control are shown.
Figure 7.
(A) Mmi1 binds the nascent transcript and inhibits splicing. (B) Mmi1 also inhibits cleavage on ∼5% of transcripts. These read-through transcripts are removed by Dis3. (C) Mmi1 promotes hyperadenylation. This hyperadenylation also depends on the polyA polymerase Pla1 [6] and probably Pab2. Hyperadenylated transcripts are rapidly degraded by Rrp6. Protein-protein interactions are shown by dotted lines based on the following evidence: Mmi1-Pfs2: genetic interactions (data not shown); Mmi1-Rna15: yeast two-hybrid and co-IP [6]; Mmi1-Pab2 and Mmi1-Pla1: yeast two-hybrid and co-IP [6]; Mmi1-Rrp6: unpublished observation [3]; Rrp6-Pab2: co-IP [9]; Rrp6-Pla1: synthetic rescue in budding yeast [58] and Pab2-Pla1: biochemical interaction in mammalian cells [59].