7The Yeast mRNA-binding Protein Npl3p Interacts with the Cap-binding Complex*

A number of RNA-binding proteins are associated with mRNAs in both the nucleus and the cytoplasm. One of these, Npl3p, is a heterogeneous nuclear ribonucleoprotein-like protein with some similarity to SR proteins and is essential for growth in the yeastS. cerevisiae. Temperature-sensitive alleles have defects in the export of mRNA out of the nucleus (1). In this report, we define a genetic relationship between NPL3 and the nonessential genes encoding the subunits of the cap-binding complex (CBP80 and CBP20). Deletion ofCBP80 or CBP20 in combination with certain temperature-sensitive npl3 mutant alleles fail to grow and thus display a synthetic lethal relationship. Further evidence of an interaction between Npl3p and the cap-binding complex was revealed by co-immunoprecipitation experiments; Cbp80p and Cbp20p specifically co-precipitate with Npl3p. However, the interaction of Npl3p with Cbp80p depends on both the presence of Cbp20p and RNA. In addition, we show that Cbp80p is capable of shuttling between the nucleus and the cytoplasm in a manner dependent on the ongoing synthesis of RNA. Taken together, these data support a model whereby mRNAs are co-transcriptionally packaged by proteins including Npl3p and cap-binding complex for export out of the nucleus.

A number of RNA-binding proteins are associated with mRNAs in both the nucleus and the cytoplasm. One of these, Npl3p, is a heterogeneous nuclear ribonucleoprotein-like protein with some similarity to SR proteins and is essential for growth in the yeast S. cerevisiae. Temperature-sensitive alleles have defects in the export of mRNA out of the nucleus (1). In this report, we define a genetic relationship between NPL3 and the nonessential genes encoding the subunits of the cap-binding complex (CBP80 and CBP20). Deletion of CBP80 or CBP20 in combination with certain temperature-sensitive npl3 mutant alleles fail to grow and thus display a synthetic lethal relationship. Further evidence of an interaction between Npl3p and the cap-binding complex was revealed by co-immunoprecipitation experiments; Cbp80p and Cbp20p specifically co-precipitate with Npl3p. However, the interaction of Npl3p with Cbp80p depends on both the presence of Cbp20p and RNA. In addition, we show that Cbp80p is capable of shuttling between the nucleus and the cytoplasm in a manner dependent on the ongoing synthesis of RNA. Taken together, these data support a model whereby mRNAs are co-transcriptionally packaged by proteins including Npl3p and capbinding complex for export out of the nucleus.
While in the nucleus, mRNA precursors, referred to as pre-mRNAs or heterogeneous nuclear RNAs undergo a series of processing events before entering the cytoplasm. These maturation events include co-transcriptional capping at the 5Ј-end, splicing, and cleavage and polyadenylation at the 3Ј-end. The proper execution of these steps affects the export of mRNA (reviewed in Refs. [2][3][4]. Thus, the process of mRNA export commences long before the RNA actually reaches the nuclear membrane. Pre-mRNAs and snRNAs 1 are first modified co-transcriptionally at the 5Ј-end by a monomethyl cap after reaching a length of 20 -30 nucleotides (5,6). This modification consists of a guanosine residue methylated at the N-7 position joined to the first encoded nucleotide of the RNA via a triphosphate linkage. The cap structure appears to be important for efficient splicing and export of RNAs (reviewed in Ref. 2).
Evidence has suggested that the effects of the cap structure on RNA metabolism are protein-mediated (7). A cap binding activity was purified from HeLa cell nuclear extracts and was found to consist of two proteins termed cap-binding protein 80 (Cbp80p) and cap-binding protein 20 (Cbp20p) (8 -10). Neither Cbp80p nor Cbp20p has been found to bind capped RNA alone, suggesting that a complex of the two proteins is required for binding of a cap (10). This complex is referred to as the capbinding complex (CBC). Homologues of both CBC proteins have been identified in all eukaryotes examined thus far (e.g. see Refs. 10 -13).
Although the 5Ј cap structure is not required for successful export, it has been shown to enhance the rate of mRNA export from the nucleus (7,14). Studies of the Balbiani ring pre-mRNP particle in the insect Chironomus tentans show that Cbp20p binds the pre-mRNA nascent transcript and remains bound to the pre-mRNA throughout splicing and as the mRNA is translocating through the nuclear pore complex (13). Because the 5Ј-end is in the lead as the mRNA exits the nucleus, these data raised the possibility that the CBC mediates the effect of the cap on mRNA export.
In the budding yeast, Saccharomyces cerevisiae, CBP80 has been isolated through both genetic and biochemical approaches (15,16). CBP80 is not an essential gene (15). However, deletion of the gene results in a severe growth defect. Yeast CBP20 was originally identified as MUD13, a gene that when mutated causes synthetic lethality in combination with a mutant form of U1 snRNA (12). Like CBP80, CBP20 is not an essential gene in yeast. A strain carrying null alleles of both CBP80 and CBP20 is also viable (17).
From the time that they leave the transcription complex, pre-mRNAs are also associated with proteins in complexes referred to as heterogeneous nuclear ribonucleoprotein particles (hnRNPs) (for a review, see Ref. 18). The hnRNP proteins are among the most abundant proteins found in the nucleus (19) and are proposed to function in nearly every maturation step of mRNA including splicing, polyadenylation, and export (for reviews, see Refs. 20 and 21). One of the most studied hnRNPs in mammalian cells is hnRNPA1, which belongs to the class of hnRNP proteins that shuttles between the nucleus and cytoplasm (22). Along with other hnRNPs in its class, hnRNPA1 contains two RNA recognition motifs (RRMs) and a glycine-rich region at the carboxyl terminus (23). Because hnRNPA1 is bound to poly(A) ϩ RNA in both the nucleus and the cytoplasm and is able to shuttle, it has been proposed to play a role in mRNA export (for a review, see Ref. 20). Shuttling proteins like hnRNPA1 could be escorting the RNA to the cytoplasm, where other RNA-binding proteins take over.
Several hnRNP proteins have been identified in S. cerevisiae, allowing for a genetic approach to the study of their functions.
One of the most studied hnRNP proteins is Npl3p (24,25). Cells bearing mutant npl3 alleles accumulate poly(A) ϩ RNA in their nuclei at the nonpermissive temperature, consistent with a block in mRNA export (1,(25)(26)(27). Npl3p shares structural features found in some mammalian hnRNPs such as hnRNPA1. These include two RRMs and a glycine-rich domain (RGG box) in the form of 15 RGG repeats (23). Npl3p is methylated at arginine residues in this glycine-rich domain by the major arginine methyltransferase, Hmt1p/Rmt1p, in yeast cells (28 -30). In addition, Npl3p contains a domain characterized by a series of serine-arginine repeats that overlaps with the RGG box. This domain, referred to as the RS domain, is found in splicing factors known as SR proteins (for a review, see Ref. 31). Npl3p has been shown to cross-link to poly(A) ϩ RNA (32,33). Although localized in the nucleus at steady state, Npl3p has been shown to shuttle between the nucleus and the cytoplasm (1,34). This ability to shuttle depends on the presence of RNA polymerase II activity such that in a strain carrying a mutation in RNA polymerase II, exit of Npl3p out of the nucleus is impaired (1). Furthermore, the ability of Npl3p to shuttle appears important for its function, because a strain carrying a mutation in an RRM that abolishes its ability to shuttle is temperature-sensitive (1). These data have led to a model where Npl3p plays a role as a carrier of mRNA for export out of the nucleus.
We have previously reported genetic interactions between NPL3 and HMT1 as well as between HMT1 and CBP80 (28,35). These results prompted us to explore the possible interaction between NPL3 and CBP80. We now report that CBP80 and CBP20 share a genetic relationship with NPL3. Furthermore, we show that Cbp80p and Cbp20p physically interact with Npl3p in vivo. Our results suggest that these RNA-binding proteins may work together to promote the export of mRNAs.

MATERIALS AND METHODS
Plasmids and Strains-Many of the plasmids carrying temperaturesensitive alleles of NPL3 were published earlier (1). The remaining mutant alleles come from the same screen for temperature-sensitive alleles. Briefly, a polymerase chain reaction mutagenesis and plasmid shuffling protocol was used, using the strain MHY132 and the plasmid pMHY3 (1,27). The plasmid carrying NPL3-Myc was the same plasmid used in previous studies (pPS1356; Ref. 35). Briefly, the Myc epitope was inserted into the PmlI site at the C terminus of NPL3 in a URA3 CEN plasmid (24,34). This construct was digested with ScaI. The 3.5-kilobase fragment containing NPL3-Myc was inserted into SmaIdigested pRS315 (36) to create pPS1356.
Yeast strains used in this study are listed in Table I. PSY905 was created by transforming MHY132 (1) with YCp50-NPL3-LEU2 and streaking the resulting transformants on 5-fluoroorotic acid (5-FOA) to select against the YCp50-NPL3-URA3 plasmid in the original strain. To disrupt the CBP20 gene, a polymerase chain reaction strategy was used (37). The resulting cbp20⌬ strain, PSY1942 was then crossed to PSY817, a npl3⌬ strain carrying a plasmid expressing Npl3p. The resulting diploid was sporulated, and tetrads were analyzed to obtain the strain PSY1944. Similarly, PSY865 and PSY817 were crossed as well as PSY1120 and PSY814. The diploids were sporulated, and tetrads were analyzed to obtain PSY1943 and PSY1946, respectively.
Export Assay-The nuclear export assay was performed exactly as described earlier (1). Briefly, expression of the reporter protein was induced for 2 h by growth of the cells in medium containing galactose. Expression of the reporter protein was then repressed by growth in glucose-containing medium for 2 h before shifting half of the culture to 37°C for 5 h and leaving the other half at 25°C. Cells were then visualized for green fluorescent protein (GFP) fluorescence.
Immunoprecipitations-Lysates for immunoprecipitations were prepared from cultures grown at 25°C in SC medium lacking leucine to an A 600 of 0.8 -1.0. Cells were washed in water, and glass beads (425-600 m; Sigma) were added such that the ratio of glass beads to size of pellet was approximately 1-2:1. Lysis buffer (137 mM NaCl, 1.76 mM KH 2 PO 4 , 5.4 mM Na 2 HPO 4 , 5.7 mM KCl, pH 7.2, 2.5 mM MgCl 2 , 0.5% Triton X-100; PBSMT) with protease inhibitors (1 mM phenylmethylsulfonyl fluoride and 3.125 g/ml each of pepstatin A, leupeptin, aprotinin, and chymostatin (all from Sigma)) was then added such that the buffer just covered the pellet and beads. Cells were lysed using the FastPrep Cell Disruptor (Bio101/Savant Instruments) for 30 s at a speed of 6.5. Lysis buffer was added to a final volume of 1 ml. Lysates were then clarified by centrifugation for 10 min at 14,000 ϫ g at 4°C. Protein concentrations of the lysates were determined using a protein assay kit (Bio-Rad). 1.5 mg of total protein of each lysate was incubated with anti-Myc beads (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) in a total volume of 500 l for 2 h at 4°C. For the RNase sensitivity experiment, lysates were pretreated with 0.5 l of 10 mg/ml RNase A (Sigma) for 15 min at 25°C. The beads were washed two times for 10 min each at 4°C with lysis buffer and once for 10 min at 4°C with lysis buffer containing only 0.1% Triton X-100. The beads were resuspended in lysis buffer without detergent. Bound proteins were eluted by the addition of SDS buffer (50 mM Tris-HCl (pH 6.8), 100 mM dithiothreitol, 2% SDS, 0.1% bromphenol blue, 10% glycerol) and separated on a 12% gel by SDS-polyacrylamide gel electrophoresis.
Silver Staining-For silver staining, the gel was fixed for 15 min in destain (30% methanol, 10% acetic acid) and then washed in water overnight. The gel was soaked for 15 min in a 50-ml solution consisting of a mixture of 0.4 g of AgNO 3 dissolved in 2 ml of water added slowly to 10.5 ml of 0.36% NaOH and 1.25 ml of 21% NH 4 OH. The gel was then washed three times for 3 min in water and then developed using developer solution consisting of 50 ml of water with 25 l of 10% citric acid and 25 l of 37% formaldehyde. Silver staining was stopped by soaking the gel in destain.
Immunoblots-For immunoblotting, gels were transferred to nitrocellulose membranes (Protran; Schleicher & Schuell). The membranes were incubated for 30 min at 25°C with PBS (137 mM NaCl, 1.76 mM KH 2 PO 4 , 5.4 mM Na 2 HPO 4 , 2.7 mM KCl, pH 7.2), 0.25% Tween 20 (Bio-Rad), and 2.5% milk powder. To detect proteins, membranes were then incubated for 1 h at 25°C with antibodies diluted in the same solution. This was followed by incubation with horseradish peroxidaseconjugated secondary antibodies (Jackson Immunoresearch Laboratories) and detection using enhanced chemiluminescence (Amersham Pharmacia Biotech). To detect the proteins of interest, we diluted corresponding antibodies as follows: anti-GFP antibody, 1:5000; anti-Myc antibody (Santa Cruz Biotechnology), 1:1000; anti-Cbp80p antibody, 1:10,000; and anti-Cbp20p antibody, 1:10,000. Cbp80p Is Able to Exit the Nucleus-Once bound to mRNAs, the cap-binding complex has been proposed to move out of the nucleus together with the RNA, as has been demonstrated for Npl3p (1). In yeast, Cbp80p localizes predominantly to the nucleus at steady state (16). To determine if Cbp80p is able to exit the nucleus, we utilized the in vivo nuclear export assay (1). This assay takes advantage of a temperature-sensitive mutant allele of the nucleoporin NUP49 (38). At the nonpermissive temperature of 37°C, nup49 -313 cells are defective in the nuclear import of proteins. However, no defect in export of proteins or RNAs has been observed in these cells. Therefore, if a protein that localizes in the nucleus at steady state is able to exit the nucleus, it will accumulate in the cytoplasm after the temperature shift in the absence of new protein synthesis.
To monitor the localization of Cbp80p, GFP was fused to the C terminus of the CBP80 coding region. The resulting fusion protein was placed under the control of the regulatable GAL1 promoter. This construct is functional because it rescues the synthetic lethality between hmt1⌬ and cbp80 -1 (data not shown). Expression of Cbp80p-GFP was induced by growth in galactose-containing medium. A fusion protein of the correct size, approximately 120 kDa, is seen by Western blot with anti-GFP antibody, with no free GFP detected (Fig. 1A). When cells expressing Cbp80p-GFP were viewed by fluorescence microscopy, all of the Cbp80p-GFP was in the nucleus (Fig. 1B).
In the nuclear export assay, expression of Cbp80p-GFP was induced for 2 h. Following repression of new protein synthesis by growth in glucose-containing medium, the cells were either left at the permissive temperature of 25°C or shifted to the nonpermissive temperature of 37°C for 5 h. As expected, Cbp80p-GFP localizes to the nucleus in nup49 -313 cells that were grown at the permissive temperature (Fig. 2, A and B). However, when the nup49 -313 cells were shifted to the nonpermissive temperature, Cbp80p-GFP is seen accumulating in the cytoplasm (Fig. 2, C and D), indicating that Cbp80p-GFP is able to exit the nucleus. Immunoblotting of cell lysates made after the temperature shift confirm that there is no degradation of Cbp80p-GFP (data not shown). Therefore, the cytoplasmic signal is due to the intact Cbp80p-GFP.
To test if the ability of Cbp80p-GFP to exit the nucleus is dependent on RNA polymerase II transcription, we performed the nuclear export assay in a strain carrying both the nup49 -313 and rpb1-1 mutant alleles (1). At the nonpermissive temperature, Cbp80p-GFP is no longer able to exit the nucleus, suggesting that active RNA polymerase II transcription is required (Fig. 2, G and H).
CBC Interacts with NPL3 Genetically-We have previously reported that a strain containing a null allele of CBP80 (cbp80⌬ strain) is synthetically lethal with a strain containing a null allele of HMT1 (35). In addition, a strain containing a null allele of HMT1 is synthetically lethal with the temperaturesensitive (ts) allele of NPL3, npl3-1 (28). To see if CBP80 also interacted genetically with NPL3, we crossed the cbp80⌬ strain to the npl3-1 strain. Tetrad analysis revealed that progeny containing both npl3-1 and cbp80⌬ were unable to germinate, indicating a synthetic lethal relationship.
To further test synthetic lethality of the null allele of CBP80 with other ts mutants of NPL3, a strain was constructed that contained null alleles of both CBP80 and NPL3 as well as a plasmid carrying a wild type copy of NPL3 and the URA3 gene. This strain was transformed with a series of plasmids that carry a ts allele of NPL3 and the LEU2 gene (1). Transformants were then streaked on plates lacking leucine and containing 5-FOA, a drug that selects against the presence of the URA3 gene. Therefore, if the null allele of CBP80 and the temperature-sensitive allele of interest are synthetically lethal, the transformants will not be able to grow on a leu Ϫ 5-FOA plate due to the lack of a wild type copy of NPL3. For instance, when the double mutant strain is transformed with a LEU2 plasmid carrying the npl3-1 allele, it is unable to survive on a leu Ϫ 5-FOA plate (Fig. 3A). Similarly, the double mutant strain transformed with the LEU2 plasmid is unable to survive on a leu Ϫ 5-FOA plate, whereas the strain is able to survive when it carries a LEU2 plasmid expressing wild type Npl3p. The latter strain gives rise to heterogeneous colonies. Analysis of this strain confirmed that both the large and small sized colonies contain both the null alleles of NPL3 and CBP80 (data not shown). The results from the leu Ϫ 5-FOA plate tests are summarized in Table II.
To further characterize the genetic interaction between NPL3 and the CBC, we decided to test whether the npl3 alleles that were also synthetic lethal with cbp80⌬ were also synthetic lethal with a deletion of CBP20. Therefore, we also constructed strains containing the null allele of CBP20 with the NPL3 null allele as well as the URA3 plasmid carrying NPL3. These strains were also transformed with plasmids carrying various ts alleles of NPL3. Results of the genetic analyses are summarized in Table II.
Although the mutations in NPL3 lie throughout the gene (Fig. 3B), the majority are found in the two RRMs. Several of the temperature-sensitive alleles result in a change in more than one amino acid. Therefore, in these alleles, it is not known which mutation causes the temperature sensitivity. For all mutants except npl3-27, the mutated Npl3p is located in the nucleus (1, 33), 2 and all mutants have some degree of defect in mRNA export (1). While no clear relationship between the  25°C (A, B, E, and F), or at 37°C (C, D, G, and H). Cells were photographed using Nomarski optics (A, C, E, and G) or for GFP fluorescence (B, D, F, and H).
location of the mutation in NPL3 and synthetic lethality with the null allele of CBP80, or CBP20 arose, several observations can be made. First, a mutation in a residue of NPL3 that is conserved in both RRMs and SR proteins is either dead in combination with both null alleles (e.g. npl3- 38), or alive in combination with all null alleles (e. g. npl3-48). Second, there are no npl3 alleles with mutations in the RGG box that result in synthetic lethality in combination with either of the null alleles. Finally, if the mutation in Npl3p is in an unconserved residue that is next to other unconserved residues, as is the case for npl3-15 and npl3-26, the temperature-sensitive allele is not synthetically lethal with any of the null alleles.
CBC Interacts with Npl3p In Vivo-Because of the genetic relationship between CBP80 and NPL3, we wanted to test if the proteins could physically interact. We constructed a gene fusion under the control of the NPL3 promoter that encodes Npl3p with the c-Myc epitope (39) at the carboxyl terminus (35). The encoded protein is recognized by the monoclonal antibody 9E10, and is functional in that it rescues a strain containing a null allele of NPL3. Strains were transformed with the NPL3-Myc plasmid, and immunoprecipitations were performed using anti-Myc beads to isolate Npl3p-Myc along with any interacting proteins from cell lysates. Samples were analyzed by silver staining (Fig. 4A). A predominant band of a size between 80 and 90 kDa was seen in all of the samples containing Npl3p-Myc. Mass spectrometry identified the protein as the major coat protein of the L-A double-stranded RNA virus of S. cerevisiae (40). This protein has been shown to bind the m 7 GMP from 5Ј capped mRNAs in vivo (41).
A second major protein of a size between 90 and 100 kDa was seen in samples from immunoprecipitations using cell lysates from a wild type strain (data not shown) or a strain containing the null allele of NPL3 (Fig. 4A, lane 3). This protein was not observed in the sample from the immunoprecipitation using cell lysate from a strain containing a null allele of CBP80 (Fig.  4A, lane 4). Mass spectrometry analysis revealed that the protein is Cbp80p. To ensure that the presence of Cbp80p was not due to nonspecific binding to the anti-Myc beads, immunoprecipitation was also performed using cell lysate from a strain FIG. 3. A, the null allele of CBP80 and npl3-1 are synthetically lethal. The cbp80⌬ npl3⌬ strain carrying NPL3 on a URA3 plasmid was transformed with the LEU2 vector alone or with the vector carrying wild type NPL3 or the npl3-1 allele. The transformants were then streaked on a leu Ϫ 5-FOA plate and assessed for ability to grow. B, schematic diagram of npl3 temperature-sensitive mutations. The protein sequence of NPL3 consists of an N terminus that contains repeats of the amino acid sequence APQE, two RRMs, and a C-terminal RGG box. The ts mutations above the diagram of Npl3p show no synthetic lethality with cbp80⌬ and cbp20⌬. The ts mutations below the diagram of Npl3p show synthetic lethality with both null alleles.
containing a null allele of NPL3 and a plasmid expressing untagged Npl3p (Fig. 4A, lane 1). Cbp80p was not observed in this sample. Therefore, Cbp80p interacts with Npl3p-Myc in vivo. Because Cbp80p and Cbp20p form a tight complex, the immunoprecipitations were repeated to check for the presence of Cbp20p. As above, cell lysates containing Npl3p-Myc (Fig. 4B,  left panel, lanes 2-4 and 6) or a control plasmid (Fig. 4B, left  panel, lanes 1 and 5) were subjected to immunoprecipitation with anti-Myc antibody. The resulting immunoprecipitates were probed with anti-Myc antibody to confirm the presence of Myc-Npl3p (Fig. 4B, left panel, lanes 6 -10). The immunoprecipitates were also probed with anti-Cbp80p antibody to confirm the results in Fig. 4A and with antibodies against Cbp20p. As in Fig. 4A, Cbp80p co-precipitated with Npl3p-Myc in cells lacking the genomic copy of NPL3 (Fig. 4B, right panel, lane 7). Similarly, Cbp20p co-precipitates with Npl3p-Myc and Cbp80p (Fig. 4B, right panel, lane 7). A Myc-tagged fragment of Npl3p containing only RRM2 and the C-terminal RGG domain still coprecipitated with Cbp80p and Cbp20p (data not shown). Unfortunately, it was not possible to stably express a protein lacking the RRMs and the RGG domain, which can promote RNA binding.
Expression of Cbp20p is highly decreased in the strain containing the null allele of CBP80 (Fig. 4B, right panel, lanes 3  and 17) such that we cannot determine if Cbp80p is required for the interaction between Cbp20p and Npl3p-Myc. In contrast, Cbp80p is present in cell lysate from cbp20⌬ cells (Fig.  4B, right panel, lanes 4 and 5). However, it is not detected in the sample from the immunoprecipitation using this lysate (Fig. 4B, right panel, lane 9), indicating that Cbp20p is required for the interaction between Cbp80p and Npl3p.
The Interaction between CBC and Npl3p Is Dependent on RNA-To test if the interaction between the cap-binding complex and Npl3p is RNA-dependent, we treated the lysate with RNase at room temperature before performing the immunoprecipitation. The amount of Npl3p-Myc immunoprecipitated remained the same in comparison with lysate that received no RNase treatment (Fig. 5A, compare lanes 2 and 3). However, the amount of Cbp80p and Cbp20p co-precipitated with Npl3p-Myc was dramatically diminished when the lysates were treated with RNase (Fig. 5B, compare lanes 1 and 2). To rule out the possibility that Cbp80p and Cbp20p were degraded during RNase treatment, lysates were immunoblotted for FIG. 4. Interaction between the CBC and Npl3p-Myc. A, anti-Myc beads were used to immunoprecipitate complexes containing Npl3p-Myc from yeast lysates from wild type, npl3⌬, and cbp80⌬cells carrying a plasmid expressing Npl3p-Myc (lanes 3 and 4). Samples were separated on a 12% gel by SDS-polyacrylamide gel electrophoresis and analyzed by silver staining. As a control, lysate from npl3⌬ cells carrying a plasmid expressing untagged Npl3p was also used (lanes 1 and 2). B, anti-Myc beads were used to precipitate complexes from lysates of npl3⌬, cbp80⌬, and cbp20⌬ cells carrying a plasmid expressing Npl3p-Myc (both panels, lanes [7][8][9]. These samples, along with the corresponding lysates (both panels, lanes [2][3][4], were separated on a 12% gel by SDS-polyacrylamide gel electrophoresis and transferred onto nitrocellulose. The left panel shows a blot probed with anti-Myc antibodies. The right panel shows a blot with identical samples. The upper portion was probed with anti-Cbp80p antibodies, and the bottom half was probed with anti-Cbp20p antibodies. As controls, lysates from npl3⌬ and cbp20⌬ cells carrying a plasmid expressing untagged Npl3p were also used (both panels, lanes 1, 5, 6, and 10). The asterisk denotes the location of the antibody heavy chain, while the carat denotes the location of the antibody light chain. Untreated lysate (lane 1) was separated along with samples of the immunoprecipitations on a 12% gel, transferred to nitrocellulose membrane, and blotted with anti-Myc antibodies. B, a blot of the identical immunoprecipitation samples in A was cut in half, and the upper portion was probed with anti-Cbp80p antibodies while the bottom half was probed with anti-Cbp20p antibodies. C, a blot of equal amounts of untreated and RNase-treated lysates from npl3⌬ (lanes 1 and 2) expressing Npl3p-Myc was cut in half as in B. Once again, the upper half was probed with anti-Cbp80p antibodies, and the bottom half was probed with anti-Cbp20p antibodies.
Cbp80p and Cbp20p. Both proteins were present at the same level before and after treatment (Fig. 5C, lanes 1 and 2). In conclusion, the lower levels of Cbp80p and Cbp20p observed in immunoprecipitations using RNase-treated lysates appear to be due to a weakened ability of these proteins to bind Npl3p-Myc, suggesting that the interactions are mediated at least in part by RNA. All attempts to reconstitute a direct interaction between Npl3p and CBC with recombinant proteins have thus far been negative, further suggesting a need for RNA to support the interaction. DISCUSSION In this report, we have demonstrated a genetic and biochemical interaction between one of the major yeast mRNA-binding proteins, Npl3p, and the proteins of the cap-binding complex. Mutations in NPL3 show allele-specific synthetic lethal relationships with both CBP80 and CBP20. Moreover, a complex containing Npl3p, Cbp80p, and Cbp20p can be isolated from yeast and is dependent on the presence of RNA.
Deletion of either component of the cap-binding complex, CBP80 or CBP20, is dead when combined with certain temperature-sensitive npl3 mutations. Moreover, CBP80 and CBP20 show the same allele-specific synthetic lethality with the npl3 ts mutants, supporting the idea that Cbp80p and Cbp20p work as a complex with relation to Npl3p. Taken together, these genetic data support the idea that the CBC works with Npl3p to ensure proper RNA metabolism.
The genetic interaction between the genes that encode the CBC and NPL3 is further supported by our finding that a complex containing both the CBC and Npl3p can be isolated from yeast. This complex requires Cbp20p because binding of Cbp80p to Npl3p is not detected in a strain lacking CBP20. It is possible that Cbp80p interacts with Npl3p only after Cbp20p binds the capped RNA. However, it appears that neither Cbp80p nor Cbp20p is able to bind capped RNA alone (10). Perhaps the interaction between the CBC and Npl3p can take place only after the CBC has bound the capped RNA. We could not determine if Cbp80p is necessary for the binding of Cbp20p to Npl3p, because expression of Cbp20p is barely detectable in a strain lacking CBP80 (17). 2 We have found that the interaction between the CBC and Npl3p is sensitive to RNase treatment, supporting the idea that capped RNA may also need to be present for the interaction. A direct interaction between the CBC and Npl3p has not yet been found.
In addition to binding the cap structure of mRNAs, the cap-binding complex is also known to bind the cap structure of snRNAs. Interestingly, Npl3p has also been associated with snRNAs in that it has been found as a potential component of the yeast U1 snRNP (42). In the same study, it was shown that the association of Npl3p with the U1 snRNP is weak, saltsensitive, and only moderately specific. Similarly, Mattaj and colleagues (43) have found that Npl3p and Cbp80p can be co-immunoprecipitated with the U1 snRNP protein Luc7. These interactions with Luc7 have also been shown to be weak and easily dissociated. However, the formal possibility still exists that interaction between the CBC and Npl3p exists in the context of snRNAs and not mRNAs.
We found that, like Npl3p and other mRNA-binding proteins, Cbp80p shuttles between the nucleus and the cytoplasm. This is consistent with the observation that Cbp20p is bound to the Balbiani ring RNP as it traverses across the nuclear pore (13). We also report that, like Npl3p, export of Cbp80p depends on ongoing mRNA synthesis. This result suggests that Cbp80p may need to be bound to mRNA in order to be exported. It has been previously shown that Cbp20p binds to nascent transcripts before they are released from the chromosome (13). Because of the interdependence shown between CBP20 and CBP80, Cbp80p is also most likely bound at this time. Capping itself has been shown to occur co-transcriptionally (5). This occurs through direct interactions between the capping machinery and the phosphorylated C-terminal domain of RNA polymerase II (44,45). Perhaps Cbp80p export depends on RNA polymerase II because binding of the CBC to capped RNA is also occurring co-transcriptionally.
Taken together, we propose that these data lend further support to a model where mRNAs are packaged co-transcriptionally by proteins such as Npl3p and Cbp80/20p. It is possible that the binding of the CBC to the cap structure of mRNAs as they are being synthesized promotes the binding of other mRNA-binding proteins such as Npl3p (Fig. 6). These and other proteins then leave the nucleus together with the mRNA and are released in the cytoplasm, where they rapidly return to the nucleus for another round of packaging and export. and P. Ferrigno for critical reading of the manuscript and for many helpful discussions.