A screen for Npl3 export factors identifies SPT15, which encodes TATA-binding protein
The export of Npl3 is intimately tied to that of mRNA; therefore, we postulated that studying Npl3 export would yield information about general mRNA export. To identify factors important for Npl3 and mRNA nuclear export, a mutant npl3-27 was employed. Npl3-27 contains a single point mutation (E409K), and cells bearing this mutation are viable at all temperatures (Lee et al. 1996
). However, the cellular distribution of Npl3-27 is altered with respect to wild-type Npl3. At steady state, wild-type Npl3 localizes predominantly to the nucleus (Fig. 1A, panels a-c) whereas the Npl3-27 mutant localizes to both the nucleus and cytoplasm at 37°C (Fig. 1A, panels d-f) as detected by indirect immunofluorescence using polyclonal 
-Npl3 antibodies (Bossie et al. 1992; Krebber et al. 1999). Npl3-27 appears to have a slowed rate of import because overexpression of the Npl3 import receptor gene, MTR10, results in restoration of Npl3-27 nuclear localization (Krebber et al. 1999). The E409K mutation maps to the region that is important for Npl3 nuclear localization, further supporting this explanation (Flach et al. 1994). It has been shown previously that mutations that cause mRNA export defects also cause nuclear accumulation of Npl3-27, indicating that Npl3-27 export is subject to the same requirements as mRNA export (Krebber et al. 1999). As seen by in situ hybridization using an oligo (dT)50 probe, wild-type (Fig. 1B, panels a-c) and npl3-27 cells (Fig. 1B, panels d-f) display the same distribution of mRNA throughout the nucleus and cytoplasm (Amberg et al. 1992; Krebber et al. 1999). Therefore, Npl3-27 is functional for mRNA export despite its altered cellular localization. Furthermore, as has been shown with similar mutant forms of Npl3, Npl3-27 is capable of binding mRNA in vivo at 25°C and 37°C as determined by UV-crosslinking (data not shown; Gilbert et al. 2001). The cellular level of Npl3-27 is equivalent to that of wild type at both 25°C and 37°C (Krebber et al. 1999).
Taking advantage of the altered localization but wild-type activity of Npl3-27, we performed a genetic screen to identify genes required for Npl3-27 export. A temperature-sensitive (ts
) library from the npl3-27 strain was created by EMS mutagenesis and was screened for nuclear accumulation of Npl3-27. An intergenic mutation (F183I) in an RRM domain of NPL3 was obtained that also causes an mRNA export defect. Moreover, a mutation in the mRNA export factor MTR2 was also found (data not shown). Identification of these two mutations verified that this screen is capable of identifying genes that affect Npl3 and mRNA export. This screen also identified a mutation in SPT15, which encodes the transcription initiation factor TBP. Sequencing of this mutant revealed that the mutation is identical to a previously reported ts spt15 mutant termed spt15-ts1 (Cormack and Struhl 1992). This mutant displays a nuclear accumulation of Npl3-27 at the nonpermissive temperature of 37°C indicating a block of Npl3-27 export (Fig. 1A, panels g-i). Levels of Npl3-27 are not affected (data not shown).
NPL3 and SPT15 genetically interact to promote mRNA export
Unexpectedly, in addition to nuclear accumulation of Npl3-27, spt15-ts1 npl3-27 cells display an mRNA export defect. Poly(A)+ RNA accumulates in the nucleus of spt15-ts1 npl3-27 cells shifted to 37°C for 1 h (Fig. 1B, panels g-i). Neither the single spt15-ts1 mutation (Fig. 1B, panels j-l) nor the npl3-27 mutation (Fig.1B, panels d-f) in isolation cause nuclear accumulation of mRNA. Consistent with a defect in transcription at 37°C, spt15-ts1 cells display a lower intensity poly(A)+ RNA signal distributed throughout the cells relative to wild type (Fig. 1B, panels j-l). Double mutant spt15-ts1 npl3-27 cells (Fig. 1B, panels g-i) display a nuclear RNA signal that is of considerably higher intensity than that of spt15-ts1 cells. One explanation is that blocked export and subsequent concentration of mRNA in the nucleus causes a brighter signal. Another possibility is that npl3-27 decreases the rate of mRNA degradation in the nucleus, resulting in an increase of the overall poly(A)+ RNA signal and an apparent accumulation in the nucleus.
To verify that spt15-ts1 npl3-27 cells exhibit a true decrease in mRNA export and not a defect in mRNA stability, we harvested total RNA from wild-type, spt15-ts1, spt15-ts1 npl3-27, and npl3-27 cells and determined the amount of steady state poly(A)+ RNA in strains grown at 25°C or shifted to 37°C for 1 h. The amount of poly(A)+ RNA was determined by hybridization to a radiolabeled poly dT probe. Levels of 18S rRNA were verified to be constant among samples (data not shown). Single spt15-ts1 mutant cells (Fig. 2A, lanes 5,6) and double spt15-ts1 npl3-27 cells (Fig. 2A, lanes 3,4) showed an identical decrease in poly(A)+ RNA after the shift to 37°C. The poly(A)+ RNA level in single npl3-27 mutant cells (Fig. 2A, lanes 7,8) matched that of wild type (Fig 2A, lanes 1,2), indicating that npl3-27 has no effect on total mRNA levels. To rule out the effects of poly(A) tail length, Northern analysis was performed on the ACT1 transcript, and these results mirrored those of the poly(A)+ RNA slot blot. Wild-type ACT1 levels (Fig. 2B, lanes 1,2) matched those of npl3-27 cells (Fig. 2B, lanes 7,8), and spt15-ts1 npl3-27 ACT1 levels (Fig. 2B, lanes 3,4) matched those of spt15-ts1 cells (Fig. 2B, lanes 5,6) after a shift to 37°C for 1 h. Differences in reduction of transcription levels seen in the spt15-ts1 and spt15-ts1 npl3-27 cells between the two assays are likely attributable to the higher stability of the ACT1 transcript relative to total mRNA. Therefore, nuclear mRNA accumulation in spt15-ts1 npl3-27 cells results from decreased mRNA export and not altered mRNA stabilization.
These results indicate that NPL3 and SPT15 genetically interact, ensuring proper mRNA export. Furthermore, npl3-27 exerts a dominant effect on spt15-ts1 cells and not wild-type cells in that expression of exogenous npl3-27 over endogenous wild-type NPL3 causes an mRNA export defect in spt15-ts1 cells. We tested other ts alleles of spt15 (spt15-328 and spt15-341; Arndt et al. 1995) in combination with npl3-27 and found that these alleles also displayed nuclear accumulation of Npl3-27 and mRNA. To determine the specificity of this genetic interaction, we examined strains mutated for RNA Pol II that decrease transcription (rpb1-1, rpb1
101, rpb1103), but these strains do not display Npl3-27 or mRNA export defects (data not shown). Therefore, there is a specific genetic interaction between SPT15 and NPL3, which suggests interplay between the mRNA transcription and nuclear export machinery.
Npl3 and RNA Pol II exist in a complex
The genetic interaction between NPL3 and SPT15 suggested that Npl3 and TBP or some component of the transcription machinery may interact physically. Therefore, we performed -Npl3 immunoprecipitation experiments to isolate endogenous Npl3 and bound proteins from yeast cell extracts. We were unable to detect TBP in a complex with Npl3 (data not shown). However, we found that a small amount of RNA Pol II coimmunoprecipitates with -Npl3 antibodies (Fig. 3, lane 2). Furthermore, we found that this association is not RNA-dependent as Pol II is still found associated with Npl3 after incubation with RNase A (Fig. 3, cf. lanes 2 and 3). These results indicate that Npl3 is associated with the transcription machinery and may be recruited to transcribing genes by RNA Pol II.
Npl3 is associated with the chromatin of genes in a transcription-dependent manner
To determine whether Npl3 is localized in the vicinity of transcribed genes, we tested whether Npl3 is recruited to mRNAs cotranscriptionally by the method of chromatin immunoprecipitation (Orlando et al. 1997). We chose to analyze the promoter and coding region of the constitutively and highly transcribed PMA1 gene as well as a nontranscribed region devoid of open reading frames (ORFs) on a different chromosome. Chromatin of an average size of 200 bp was prepared from wild-type cells, and as a positive control, TBP was immunoprecipitated using polyclonal -TBP antibodies. Similarly, Npl3 was immunoprecipitated using polyclonal -Npl3 antibodies. The amount of DNA associated with each protein was determined by performing quantitative PCR using primer sets spanning the indicated regions (Fig. 4A). For illustrative purposes, PCR products from a single dilution of input and a single dilution of each immunoprecipitate that are known to be in the linear range of PCR are shown. In accordance with previous studies, -TBP antibodies preferentially immunoprecipitate promoter DNA (Fig. 4B, center panel, lane 1) in comparison to DNA spanning the coding sequence (Fig. 4B, center panel, lanes 2,3) and intergenic region (Fig. 4B, center panel, lane 4; Komarnitsky et al. 2000; Kuras and Struhl 1999).
Npl3 immunoprecipitates modest levels of the PMA1 promoter region (Fig. 4B, right panel, lane 1) and high levels of ORF DNA (Fig. 4B, right panel, lanes 2,3) in comparison with the intergenic region (Fig. 4B, right panel, lane 4). For each primer set, percentages of input DNA present in the -Npl3 immunoprecipitate were graphed with quantitation error for a single experiment (Fig. 4C). Values were normalized by dividing the percentage obtained for each primer set by the percentage of intergenic region (Fig. 4C, bottom). Approximately 1.5-fold more promoter DNA (Fig. 4C, bar 1) and 12- to 17-fold more coding sequence (Fig. 4C, bars 2,3) immunoprecipitate with -Npl3 antibodies compared to the intergenic region (Fig. 4C, bar 4), indicating that Npl3 is associated with chromatin of a transcribing gene. Similar results were obtained using a monoclonal antibody to Npl3 (1E4) and -myc antibodies to a myc epitope-tagged version of Npl3. Background signal ( obtained with -myc antibodies in a strain in which Npl3 was not tagged with myc (data not shown). To examine the specificity of Npl3 chromatin association with Pol II transcribed genes, we analyzed two Pol III transcribed tRNA genes (Fig. 4D). In comparison to input (Fig. 4D, lanes 1-3), Npl3 does not associate with regions containing the genes encoding tRNAGUC (Fig. 4D, lane 4) and tRNACUU (Fig. 4D, lane 5) or the intergenic region (Fig. 4D, lane 6). These results show that Npl3 associates specifically with Pol II transcribed genes.
To determine if the association of Npl3 with genes is dependent on transcription, we analyzed the highly transcribed galactose inducible GAL10 gene using primer sets spanning the promoter and coding sequence (Fig. 5A). Chromatin was prepared from cells grown in noninducing glucose-containing media as well as cells grown in inducing galactose-containing media. In cells grown in glucose, GAL10 expression is repressed, and -TBP antibodies immunoprecipitate background levels of DNA (Fig. 5B, center panel, lanes 1-5). With galactose induction, -TBP antibodies immunoprecipitate a significant level of upstream activating sequence (UAS) DNA (Fig. 5B, center panel, lane 1) and DNA directly downstream of the TATA-box and 5' coding sequence (Fig. 5B, center panel, lane 2) but not DNA well into the coding region of GAL10 (Fig. 5B, center panel, lanes 3,4) or the intergenic region (Fig. 5B, center panel, lane 5).
-Npl3 antibodies immunoprecipitate background levels of DNA from cells grown in glucose. Under transcription-inducing conditions, -Npl3 antibodies immunoprecipitate a significant amount of GAL10 UAS DNA (Fig. 5B, right panel, lane 1) and a much greater amount of DNA corresponding to coding sequence (Fig. 5B, right panel, lanes 2-4). For each primer set, percentages of input DNA present in the -Npl3 immunoprecipitate after galactose induction (Fig. 5C, gray bars) and noninducing conditions (Fig. 5C, white bars), were graphed with quantitation error for a single experiment (Fig. 5C). Values were normalized by dividing the percentage obtained for each primer set by the percentage of intergenic region. The ratio of these normalized values of galactose to glucose are indicated (Fig. 5C, bottom). Approximately a 7-fold increase of UAS DNA (Fig. 5C, bar 1) and 30- to 50-fold increase of ORF DNA (Fig. 5C, bars 2-4) was immunoprecipitated with -Npl3 antibodies after galactose induction compared to noninducing conditions, indicating that Npl3 is associated with GAL10 in a transcription dependent manner.
Yra1 associates with chromatin at a later step of transcription
To determine whether cotranscriptional recruitment of mRNA export factors is a general phenomenon, we examined whether Yra1 associates with genes by chromatin immunoprecipitation. Chromatin was prepared from cells containing a myc-epitope tagged version of Yra1 (K. Straesser and E.C. Hurt, unpubl.), and Yra1-myc was immunoprecipitated with -myc antibodies. Quantitative PCR was performed for regions of PMA1 as described in Figure 5A. In comparison to the intergenic reference primer (Fig. 6A, bar 4), Yra1-myc does not associate with PMA1 promoter DNA (Fig. 6A, bar 1) but does so modestly with PMA1 5' coding sequence (Fig. 6A, bar 2) and strongly with 3' coding sequence (Fig. 6A, bar 3). Values were normalized to the intergenic region, indicating an eightfold increase of 5' coding sequence and a 40-fold increase of 3' coding sequence (Fig. 6A, bottom). Therefore, Yra1 associates with the coding sequence of a transcribing gene. Furthermore, Yra1 association is biased toward the 3' of PMA1 in contrast to Npl3, which associates strongly at the 5' end of PMA1 (Fig. 4).
To assess transcriptional dependence of Yra1 association with genes, we examined the GAL10 inducible gene (Fig. 5A) in chromatin prepared from cells grown in noninducing raffinose-containing media or inducing galactose-containing media. Under noninducing conditions (Fig. 6B, white bars), Yra1-myc did not immunoprecipitate a significant amount of GAL10 sequence (Fig. 6B, bars 1-4) compared to the intergenic reference primer (Fig. 6B, bar 5). However, under inducing conditions (Fig. 6B, gray bars), Yra1-myc preferentially associated with the middle and 3' regions of GAL10 coding sequence (Fig. 6B, bars 3,4) and not UAS or 5' sequences (Fig. 6B, bars 1,2). Normalized ratios were obtained by dividing percent input values obtained in galactose by values obtained in raffinose and show that there is a 10-fold induction of association with GAL10 middle and 3' coding sequence (Fig. 6B, bottom). These results indicate that Yra1 associates with genes in a transcription-dependent manner. Furthermore, Yra1 binds preferentially to the 3' ends of both PMA1 and GAL10 in contrast to Npl3, which binds strongly to the 5' ends and throughout the coding sequence of both genes.
Answers to all your Biology Questions
