Following Temperature Stress, Export of Heat Shock mRNA Occurs Efficiently in Cells with Mutations in Genes Normally Important for mRNA Export^▿

Yeast strains, genetic methods, growth conditions, and antibodies.

Yeast strains were grown in yeast extract-peptone-dextrose (YPD)-rich medium or in synthetic complete medium lacking leucine, tryptophan, or both. Yeast transformation was performed using a standard lithium acetate method (25). The yeast strains and plasmids used in these studies are listed in Table 1.

The YRA1 shuffle strain (CSY1037) was obtained by replacement of one copy of the YRA1 gene in a wild-type diploid with a kanamycin resistance gene. The strain was transformed with a URA3 CEN plasmid containing YRA1. The desired strain was isolated following sporulation and selection for haploids that are kanamycin resistant and able to grow on media lacking uracil.

The YRA1 mutant strains were obtained by transforming plasmids encoding mutants of YRA1 both with and without HA epitope tags into the YRA1 shuffle strain. The former were obtained from Francoise Stutz (44) and the latter from Ed Hurt (35). After transformation, the cells were plated on synthetic complete plates lacking tryptophan. After 2 days, a single colony was streaked onto a 5-fluoro-orotic acid (5-FOA) plate and incubated an additional 2 days. Colonies from those plates lack the wild-type copy of YRA1 and contain a mutant form of YRA1 as the only copy.

yra1 npl3-1 double mutants were made by crossing the YRA1 shuffle strain CSY1037 with the npl3-1 strain PSY777. Colonies were replica plated, and those that did not grow at 37°C and on 5-FOA plates were detected as the double-mutant strains containing both the yra1 disruption and the npl3-1 allele. yra1 mutant plasmids were transformed into the double-mutant strain. Selection for cells able to grow on 5-FOA plates permitted the isolation of strains lacking the wild-type copy of YRA1.

The integration of a green fluorescent protein (GFP) tag at the end of the coding region of the SSA4 gene was performed by linearizing an integrating plasmid encoding Ssa4p-GFP with SalI and transforming it into wild-type cells, using the lithium acetate method.

Monoclonal antibodies to Npl3p and Nab2p were obtained from M. Swanson, University of Florida, Gainesville, FL.

Growth assay.

Cells were grown overnight in 5 ml of YPD. Cell suspensions were diluted to an optical density at 600 nm (OD₆₀₀) of 0.05, and 1:5 serial dilutions were spotted onto YPD plates. The plates were incubated for 4 days at both 30°C and 34°C.

SSA4 in situ hybridization assay.

In situ hybridization was performed to localize SSA4 mRNA, as described previously (2). Yeast strains containing plasmid-based SSA4 on a 2μ high-copy plasmid were grown overnight to a maximum OD₆₀₀ of 0.5. Images were obtained using a Zeiss Axioplan 2 fluorescence microscope equipped with a cooled charge-coupled-device camera and 100× and 63× objective lenses. The distribution of SSA4 mRNA and the locations of nuclei were visualized in the same cells. Each experiment was repeated at least twice.

Ssa4p-GFP FACS assay.

A fluorescence-activated cell sorter (FACS) was used to measure the levels of Ssa4p-GFP produced following temperature stress. Yeast strains containing an integrated SSA4-GFP allele were grown in selective media overnight to a maximum OD₆₀₀ of 0.5. Cultures were shifted to 42°C for 30 min, collected by centrifugation at 2,000 rpm at 4°C for 2 min, and resuspended in ice-cold phosphate-buffered saline, followed by incubation on ice. Cell concentrations were approximately 10⁶ cells per ml.

For each sample, the GFP signal intensity of 10⁵ cells was measured at 4°C using a FACSTAR cell sorter (Becton Dickinson). Graphical plots showing the relative numbers of cells with various GFP signal intensities were obtained by using Cell Quest software (Becton Dickinson). Each experiment was repeated at least twice.

SSA4 mRNP IP experiments.

The experiments were performed based on a previously published protocol (17). Wild-type and mutant strains encoding hemagglutinin (HA)-tagged proteins were grown in 50 ml of YPD overnight at room temperature. The OD₆₀₀s of the cultures did not exceed 0.5. Cells were shifted to 42°C for 30 min, collected by centrifugation, and washed once with Tris-buffered saline. Cell pellets were resuspended in 1 ml of RNA-immunoprecipitation (IP) buffer (25 mM Tris-HCl [pH 7.5], 100 mM KCl, 0.2% Triton X-100, 0.2 mM phenylmethylsulfonyl fluoride) supplemented with a protease inhibitor mixture (Roche), 5 mM dithiothreitol, and 10 units/ml RNasin (Invitrogen). The cells were lysed by vortexing them with glass beads in the cold room for 10 min. The lysate obtained after centrifugation (15 min, 13,000 rpm) was used for immunoprecipitations. Thirty microliters of immunoglobulin G beads (EZview Red Protein G affinity gel; Sigma Chemical Co.) was used for each immunoprecipitation and was washed in RNA-IP buffer. Antibodies were added to the beads on ice. Thirty microliters of the bound antibody-bead fraction was added to 400 μl of lysate, RNA-IP buffer was added to a 1.5-ml total volume, and the samples were rotated overnight at 4°C. The beads were then washed five times with RNA-IP buffer. The RNAs in the immunoprecipitates and the total lysate samples were precipitated by adding a 1/10 volume of 3 M sodium acetate and 1 ml of 100% ethanol. The pellet was washed with 70% ethanol, dried, and resuspended in 40 μl of RNase-free water. One microliter of DNase (RNase free; Roche) was added to each sample and incubated at 37°C for 10 min. DNase was deactivated by incubation for 1 h at 65°C. A 12.5-μl volume of each sample and 2.5 μl from the total lysate sample were used for a 20-μl reverse transcription reaction mixture (Invitrogen). From this reaction mixture, 4 μl of each sample was used for a 50-μl PCR mixture. The same SSA4 primers used to produce the SSA4 in situ hybridization probe were used. Ten microliters of each PCR was loaded onto a 1% agarose gel. Photographs were taken following electrophoresis. Each experiment was repeated at least twice.

Dependence of heat shock mRNA export on export factors and mRNA-binding proteins.

We employed two assays to analyze heat shock gene expression: in situ hybridization to determine how various mutations affected the subcellular distribution of SSA4 mRNA (encoding Hsp70) and flow cytometry to analyze production of Ssa4p. In all cases, Ssa4p-GFP was expressed from the genomic SSA4 locus. Flow cytometry has a large dynamic range and can readily distinguish between a complete and a nearly complete defect in export of heat shock mRNA. In contrast, the in situ assay can readily distinguish between normal export of SSA4 mRNA and limited nuclear accumulation under conditions where some SSA4 mRNA is exported, but this approach is considerably less able to distinguish between a complete and nearly complete block in SSA4 mRNA export.

Using these assays, we compared mutant strains (the npl3-1, mex67-5, and rip1Δ strains) whose heat shock response we had analyzed previously but not by flow cytometry (Fig. 1). At 23°C in all strains, there was virtually no SSA4 mRNA detected by in situ hybridization and the very low Ssa4p-GFP signal seen using flow cytometry represents the background and is equal to the signal that would be obtained if GFP were not fused to heat shock mRNA. Heat shock in wild-type cells led to robust expression of Ssa4p-GFP; SSA4 mRNA could be detected throughout the cell, and a high level of Ssa4p-GFP was produced (Fig. 1A). rip1Δ cells served as a negative control since there is no export of any mRNA after heat shock at 42°C (28). SSA4 mRNA accumulated in nuclei, and there was no increase in the Ssa4p-GFP signal (Fig. 1B). In npl3-1 cells (Fig. 1C), there was a strong induction of SSA4 mRNA production. A small fraction of npl3-1 cells showed limited nuclear accumulation of SSA4 mRNA. Although Ssa4p-GFP was produced, a broad range of fluorescence intensities was seen and less Ssa4p-GFP was present than in wild-type cells. A very low level of SSA4 mRNA could be detected in the cytoplasm of mex67-5 cells shifted to 42°C, but all cells accumulated SSA4 mRNA in their nuclei and there was no production of Ssa4p-GFP (Fig. 1D). These results serve as a basis for evaluating the requirements for mRNA export following heat shock in other mutant yeast strains.

The FACS assay reflects both the amount of heat shock mRNA exported and the efficiency of its translation for production of Ssa4p-GFP. Because the translatability and stability of SSA4 mRNA can be affected by various mutations, the amount of Ssa4p-GFP can be lower than that in wild-type cells even when the same amount of SSA4 mRNA is exported.

yra1 mutant strains are able to export heat shock mRNA following heat shock.

Yra1p plays an important role in mRNA biogenesis. It is recruited to the mRNA during mRNA synthesis by the THO complex (21, 37) and then recruits Mex67p to the mRNP (36). Poly(A)⁺ mRNA accumulates in nuclei when Yra1p is depleted or when yra1-1 cells are shifted to 37°C. The central region of Yra1p is essential and contains RNA recognition motifs (RRMs). Deletions that remove the N or C terminus or the central RRMs are all viable, though some grow less well than the wild type at 37°C and show nuclear accumulation of poly(A)⁺ RNA at elevated temperatures (35, 44). Both the N- and the C-terminal domains of Yra1p can bind to Mex67p, and a mutant lacking both termini is inviable.

We examined export of heat shock mRNA at 42°C in several yra1 mutant strains (35). No defect in export of SSA4 mRNA at 42°C was seen in yra1 mutants lacking either the N terminus (yra1ΔN, lacking aa 1 to 17), the C terminus (yra1ΔC, lacking aa 184 to 209), or the RRM domain (yra1ΔRRM, lacking aa 76 to 183). Surprisingly, no defect in SSA4 mRNA export was seen with the temperature-sensitive yra1-1 allele, even though this mutant is quite defective in export of polyadenylated mRNA at 37°C. The FACS analyses showed that all four mutant strains produced approximately the same amount of Ssa4p-GFP (Fig. 2) as the wild type (Fig. 1). Yra2p is a nonessential homolog of Yra1p produced at a much lower level than Yra1p. Overexpression of Yra2p permits growth of cells lacking Yra1p. We constructed double-mutant strains by deleting YRA2 from each of the yra1 mutants under study. Deletion of YRA2 had no effect on the export of SSA4 mRNA following heat shock (data not shown). We conclude that neither Yra1p nor Yra2p is required for stress mRNA export.

Yra1p lacking the RNA-binding domain (RRM) is unable to bind to heat shock RNA.

We determined whether the proteins encoded by yra1 mutants were bound to SSA4 mRNA. Strains producing HA epitope-tagged forms of Yra1p as the only form of Yra1p present (44) were heat shocked at 42°C for 30 min. Yra1p-HA and associated RNA were immunoprecipitated, and reverse transcription (RT)-PCR was used to determine whether SSA4 mRNA had been coprecipitated (Fig. 2E). For controls, immunoglobulin G Sepharose beads were incubated with the HA antibody, either without any lysate (Fig. 2E, lane 1) or with a lysate from a wild-type strain producing untagged Yra1p. The SSA4 PCR product was barely detectable in the untagged control, indicating a very low level of nonspecific binding of SSA4 mRNA to the beads or antibody. In Fig. 2E, the even-numbered lanes show SSA4 in lysates and the odd-numbered lanes reflect SSA4 RNA pulled down in association with HA-tagged Yra1p. A prominent SSA4 band was observed in cells producing wild-type Yra1p-HA, indicating that Yra1p was bound to SSA4 mRNA at 42°C (Fig. 2E, lane 5). SSA4 mRNA was also pulled down with Yra1pΔC (Fig. 2E, lane 7) and Yra1pΔN (Fig. 2E, lane 9). Yra1p lacking the RRM did not interact with SSA4 mRNA in this assay (Fig. 2E, lane 11). Although Yra1p is not required for export of heat shock mRNA, the data indicate that wild-type Yra1p and some mutant Yra1ps were recruited to heat shock mRNA.

npl3-1 yra1 double mutants are capable of heat shock response.

We mated npl3-1 to yra1 mutants and attempted to isolate npl3 yra1 double mutants. yra1-1 and npl3-1 are synthetically lethal (data not shown), but we were able to isolate double mutants containing yra1ΔC, yra1ΔN, or yra1ΔRRM. Growth assays (Fig. 3A) indicate that the single yra1 mutants grew in nearly the same fashion at 30°C and 34°C, while the double mutants were temperature sensitive and grew very poorly or not at all at 34°C.

We next examined the abilities of the double mutants to produce SSA4 mRNA and Ssa4p-GFP following heat shock. In all cases, SSA4 mRNA was induced and exported after heat shock (Fig. 3B to D). In the npl3-1 yra1ΔC strain (Fig. 3C), most cells showed nuclear foci of SSA4 mRNA along with a cytoplasmic SSA4 mRNA signal with an intensity similar to that seen with the other mutants, but this is often seen even with wild-type cells. The levels of Ssa4p-GFP produced were similar to those seen with yra1 single mutants (Fig. 2).

npl3-1 cells are defective in recruiting Yra1p to SSA4 mRNA.

The npl3-1 mutant does not have an SSA4 mRNA export defect. We wondered whether the mutant protein could be recruited to SSA4 mRNA and whether recruitment of other mRNA-binding proteins would be affected by the npl3-1 mutation. We performed SSA4 mRNA pulldown experiments with npl3-1 yra1 double-mutant strains (encoding HA-tagged Yra1p), as described above for the single yra1 mutants. Interestingly, wild-type Yra1p was not bound to SSA4 mRNA after heat shock in npl3-1 mutant cells (Fig. 4A, lane 7) but was bound to SSA4 mRNA in the wild-type (NPL3) strain (Fig. 4A, lane 5). This is additional evidence that Yra1p is not required for export of SSA4 heat shock mRNA following heat shock, because the npl3-1 strain is capable of heat shock mRNA export. Moreover, it suggests that Npl3p may be involved in recruitment of Yra1p to SSA4 mRNA, although this could be indirect. In contrast, in npl3-1 cells under normal growth conditions (23°C, permissive for npl3-1 cells), Yra1p was bound to the mRNA of a normal housekeeping gene, ADH1 (Fig. 4A). The levels of ADH1 and other non-stress response mRNAs were too low following heat shock to determine what mRNA-binding proteins were bound to the mRNAs (data not shown). None of the mutant yra1 proteins was recruited to heat shock mRNA in npl3-1 cells (data not shown).

Because Yra1p was not detected on SSA4 mRNA after heat shock in npl3-1 cells, we wondered whether Npl3-1p or Nab2p was bound to SSA4 mRNA. The data in Fig. 4B indicate that Npl3p (Fig. 4B, lane 5) and mutant Npl3-1p (Fig. 4B, lane 9) were each bound to SSA4 mRNA. This finding was verified in a pulldown experiment using GFP-tagged Npl3p (data not shown). We also observed that Nab2p was bound to SSA4 mRNA in both the wild type (Fig. 4B, lane 7) and the npl3-1 mutant strain (Fig. 4B, lane 11). Taken together, the data indicate that Yra1p, Npl3p, and Nab2p become part of the SSA4 mRNP after stress in wild-type cells. In addition, Nab2p but not Yra1p can be recruited to SSA4 mRNP in npl3-1 cells, even though Npl3-1p itself associates with SSA4 mRNA.

The THO complex is not required for export of heat shock mRNA.

The THO complex of four proteins (Tho2p, Thp2p, Hpr1p, and Mft1p) is thought to function in transcription elongation and in recruitment of Yra1p and Sub2p to nascent mRNPs. Each of the genes encoding THO complex proteins is not essential, but cells lacking any of the four proteins are ts for growth and mRNA export (37). The phenotypes of mutants lacking two THO complex subunits are very similar to those of mutants lacking a single subunit (A. Aguilera, personal communication), suggesting that in the absence of any one subunit, the complex does not form (16). Since the complex is not essential under some conditions, alternate pathways for recruiting Yra1p and Sub2p are thought to exist (20). We analyzed export of SSA4 mRNA in the single-mutant strains each lacking one THO complex component. The data for thp2Δ and tho2Δ are shown in Fig. 5A and B, but the same phenotypes were seen for each of the mutant THO mutant strains (data not shown). Each of the THO mutants produced approximately wild-type levels of Ssa4p-GFP (compare Fig. 5 with Fig. 1). As is the case with wild-type cells, poly(A)⁺ mRNA accumulated in nuclei of THO complex mutants following heat shock (31, 37; also data not shown).

Defining the requirements for export of heat shock mRNA.

Yra1p is thought to recruit Mex67p to mRNA (36, 37, 44) and, unlike Npl3p, is believed to be removed from mRNPs prior to export. To examine whether Yra1p was required for heat shock mRNA export, we used both a ts point mutant, yra1-1, and three viable mutants whose deletions remove sequences from the N-terminal, C-terminal, or central RRM regions of the protein (35). None was defective for export of SSA4 mRNA, and the amount of Ssa4-GFP produced in the yra1 mutant strains was nearly the same as that produced in wild-type cells (compare Fig. 2A to D with Fig. 1A). Even though Yra1p does not appear to be required for heat shock mRNA export, wild-type Yra1p as well as Yra1p-ΔC and Yra1p-ΔN became associated with heat shock mRNA (Fig. 2E). We also found that Nab2p and mutant Npl3-1p associated with SSA4 mRNA in wild-type cells (Fig. 4B). Together, these findings indicate that while Npl3p and Yra1p may not be required for heat shock mRNA export, both wild-type and mutant Yra1p and Npl3p can associate with SSA4 mRNA if they are present.

npl3-1 yra1 double-mutant strains were also capable of exporting heat shock message and producing Ssa4p-GFP (Fig. 3). Interestingly, recruitment of Yra1p to heat shock mRNA did not occur in npl3-1 cells (Fig. 4). One possibility is that other mRNA-binding proteins perform the function of Yra1p in cells where it is not recruited to heat shock mRNA. Hrb1p and Gbp2p are two nonessential mRNA-binding proteins. Each could be recruited to heat shock mRNA, and heat shock mRNA export was unaffected in Δhrb1 Δgbp2 double-mutant cells (unpublished results).

None of the THO complex components of TREX are required for heat shock mRNA export (Fig. 5 and data not shown), but Sub2p, which is recruited by THO, is essential for mRNA export, both in growing cells and after heat shock (data not shown). Because THO complex components are not essential for growth or heat shock gene expression, it is likely that Sub2p has another way to associate with the nascent heat shock mRNP. The finding that overexpression of Sub2p suppresses the mRNA export defect of hpr1Δ is consistent with this (20).

Pathways for mRNA export.

Whether multiple distinct pathways for mRNA export exist is not known. At one extreme, export of all mRNAs might utilize all of the same factors. At the other, there could be two or more classes of mRNA, each using a different set of mRNA export and packaging factors. Whether NPCs that are specialized for export of a subset of mRNAs exist is also not known. Even though all NPCs contain the same nucleoporins, NPCs juxtaposed to the nucleolus are not associated with Mlp1p (8) and these NPCs are thought to be used primarily for export of ribosomal subunits. There is no evidence that any NPCs are specific for export of a subset of mRNAs.

Hieronymus and Silver showed that the set of mRNAs that could be coimmunoprecipitated with Mex67p overlapped only partially with the set for Yra1p (13), and some mRNAs were not coprecipitated with either protein. One interpretation of these results is that other proteins are present and provide to some mRNAs the functions of Mex67p and Yra1p. However, it is not known whether an mRNA with a single molecule of Mex67p or Yra1p can be coprecipitated efficiently, so it is possible that all species of mRNA utilize Mex67p and Yra1p. Furthermore, some mRNAs may be sufficiently inabundant in nuclei that very few of these mRNAs would be coprecipitated even if associated with Mex67p or Yra1p. Interestingly, Mex67p is not required for mRNA export in Schizosaccharomyces pombe (41). In contrast, S. pombe requires Rae1p whereas its orthologue in S. cerevisiae, Gle2p, is not required for mRNA export.

The DEAD box protein Dbp5p is essential, and mutations affecting Dbp5p result in very rapid onsets of accumulation of poly(A)⁺ mRNA in all nuclei at nonpermissive temperatures (32) and after heat shock (28). Although Dbp5 is also essential in mammalian cells (30), it does not appear to be required for mRNA export in Drosophila (10). Because so many of the proteins important for mRNA export are highly conserved, it is likely that mRNA export occurs by identical mechanisms in all nucleated cells. We think it most likely that some organisms contain two or more proteins able to perform the same function during mRNA export while others contain a single gene product.

We have conducted multiple screens to identify proteins important for heat shock mRNA export and have not identified any transport factors in yeast that are required only for export after stress. In the screen performed using the strain that produces Ssa4p-GFP under control of the SSA4 promoter, we mutagenized cells, sorted using flow cytometry, collected the darkest 1% of cells, and then identified those strains in this set that were temperature sensitive for growth (12). We also screened our original collection of temperature-sensitive strains for those that accumulated SSA4 mRNA in their nuclei following heat shock (12). In this screen, we studied further only those strains that were temperature sensitive, because this allowed us to identify the mutant genes by complementation. We attempted to identify novel nonessential genes important for heat shock mRNA export by performing transposon-mediated mutagenesis in the Ssa4p-GFP strain and again sorting cells to isolate the darkest 1%. Using a transposon as a tool to identify mutated genes, we studied dark strains that were not temperature sensitive for growth, and the only gene found to be inactivated by the transposon was SSA4-GFP (C. A. Heath and C. N. Cole, unpublished results). This suggests that there are very few or no nonessential genes whose inactivation would result in a defect specifically in heat shock mRNA export.

In summary, the data presented here indicate that under stress conditions there are more relaxed requirements for packaging mRNA for export, in that Yra1p and Npl3p are not required. This could lead to decreased accuracy in formation of export-competent heat shock mRNPs. However, the levels of Ssa4pGFP are approximately the same in wild-type cells and in many mutants that affect mRNA-binding proteins. Possibly, heat shock mRNAs are more readily packaged into an exportable configuration than are some normal cellular mRNAs. This could result from their having less potential for formation of secondary structure or from other features of heat shock mRNA biogenesis. There is at present no evidence for the existence of heat shock-specific mRNA export factors or for the existence of more than one mRNA export pathway.