Tho1, a novel hnRNP, and Sub2 provide alternative pathways for mRNP biogenesis in yeast THO mutants

MOLECULAR AND CELLULAR BIOLOGY, June 2006, p. 4387–4398 Vol. 26, No. 12 0270-7306/06/$08.00⫹0 doi:10.1128/MCB.00234-06 Copyright © 2006, American Society for Microbiology. All Rights Reserved. Tho1, a Novel hnRNP, and Sub2 Provide Alternative Pathways for mRNP Biogenesis in Yeast THO Mutants† Sonia Jimeno, Rosa Luna, Marı´a Garcı´a-Rubio, and Andre´s Aguilera* Departamento de Gene´tica, Facultad de Biologı´a, Universidad de Sevilla, Avd. Reina Mercedes 6, 41012 Sevilla, Spain Received 4 October 2005/Returned for modification 5 December 2005/Accepted 21 March 2006 THO is a protein complex that functions in cotranscriptional mRNP formation. Yeast THO1 and SUB2 (Saccharomyces cerevisiae) were identified as multicopy suppressors of the expression defects of the hpr1⌬ mutant of THO. Here we show that multicopy THO1 suppresses the mRNA accumulation and export defects and the hyperrecombination phenotype of THO mutants but not those of sub2⌬,thp1⌬,orspt4⌬. Similarly, Sub2 overexpression suppresses the RNA export defect of hpr1⌬. Tho1 is a conserved RNA binding nuclear protein that specifically binds to transcribed chromatin in a THO- and RNA-dependent manner and genetically interacts with the shuttling hnRNP Nab2. The ability of Tho1 to suppress hpr1⌬resides in its C-terminal half, which contains the RNA binding activity and is located after a SAP/SAF (scaffold-associated protein/scaffoldassociated factor) domain. Altogether, these results suggest that Tho1 is an hnRNP that, similarly to Sub2, assembles onto the nascent mRNA during transcription and participates in mRNP biogenesis and export. Overexpression of Tho1 or Sub2 may provide alternative ways for mRNP formation and export in the absence of a functional THO complex. Transcription is a central cellular function that occurs in the nucleus of eukaryotic cells in coordination with other nuclear processes. RNA polymerase II (RNAPII)-driven transcription takes place in tight association with different mRNA processing steps (46, 59). Indeed, a number of important proteins involved in mRNA processing, such as the capping enzymes, splicing factors, and the cleavage and polyadenylation stimulatory factor responsible for recognizing the poly(A) ⫹ signal, are loaded onto the nascent mRNA by interacting with the C-terminal domain (CTD) of the RNAPII (43). Thus, in cell extracts, a recombinant CTD stimulates poly(A) site cleavage and mRNA splicing (22, 23) and influences 3⬘-end formation and capping of mRNAs (50, 55). The result of the combined action of molecular processes is a mature mRNA-protein (mRNP) particle proficient for nuclear export. Any failure compromising the formation of an export-proficient mRNP would likely trigger mRNA degradation by the nuclear exosome (28, 54), a process which may also occur cotranscriptionally as shown in Drosophila spp. (6). The nuclear integration of the different processes taking place from transcription to mRNA export was recently observed at the cellular level by the nuclear pore colocalization of actively transcribed GAL genes (8). An intriguing connection of mRNP biogenesis with other nuclear processes is that observed with genetic stability. Even though there is accumulated evidence for the stimulation of both mutation and recombination by transcription, the functional and physical interconnection between mRNP biogenesis and genetic instability comes from studies with the THO complex of Saccharomyces cerevisiae (2, 3). THO is a four-protein complex composed of stoichiometric amounts of Tho2, Hpr1, Mft1, and Thp2 (10). Null mutations of any component of THO lead to similar phenotypes of transcription impairment, transcription-associated hyperrecombination and RNA export defects (9, 10, 29, 53), and increased levels of exosome-depen- dent mRNA instability (37, 58). THO has also been purified in humans and Drosophila, and the two major subunits, Tho2 and Hpr1, are present in all of the eukaryotes studied (10, 47, 53). Furthermore, THO forms, together with the RNA export proteins Sub2/UAP56 and Yra1/Aly, a larger complex termed TREX, as identified in yeast (Saccharomyces cerevisiae) and humans (53). In addition to the physical interaction, there is a functional interaction between Sub2, Yra1, and the THO complex. Thus, mutant alleles of SUB2 and YRA1 are synthetic lethal with THO mutations and lead to the same transcriptiondependent hyperrecombination and gene expression phenotypes as do THO mutations. In addition, Sub2 overexpression suppresses hpr1⌬(13, 29, 53). THO, Sub2, and Yra1 are recruited to active chromatin in a way in which THO seems to be recruited first to facilitate the subsequent recruitment of Sub2 and Yra1 (36, 53, 58). The connection between mRNP biogenesis factors and the origin of genetic instability is also provided by the yeast Thp1- Sac3 complex involved in RNA export and located in the proximity of the nuclear pore complex (14), the Mex67-Mtr2 export factor, or the Nab2 hnRNP. Mutations in these factors confer transcription, RNA export, and hyperrecombination phenotypes similar to those of mutants of the THO complex (15, 16, 29). One major cause for the transcription impairment and hyperrecombination phenotypes of THO mutants is the cotranscriptional accumulation of DNA-RNA hybrids formed behind the elongating RNAPII in the absence of functional THO. This indicates that one role of THO/TREX in mRNP assembly may be to maintain the nascent mRNA apart from the transcribed * Corresponding author. Mailing address: Departamento de Gene´tica, Facultad de Biologı´a, Universidad de Sevilla, Avd. Reina Mercedes 6, 41012 Sevilla, Spain. Phone: 34 954 557107. Fax: 34 954 5571204. E-mail: [email protected]. † Supplemental material for this article may be found at http://mcb .asm.org/. 4387 on June 29, 2017 by USE/BCTA.GEN UNIVERSITARIAhttp://mcb.asm.org/Downloaded from DNA, impeding undesired interactions of the nascent mRNA with the DNA template (24). One way of understanding the functional interaction between genetic instability and mRNP biogenesis in general, and the THO complex in particular, has been through the isolation and analysis of multicopy suppressors. Thus, the Sub2 RNA export factor, a putative RNA-dependent ATPase with homology to RNA helicases (32), was found as a multicopy suppressor of hpr1⌬(13, 29). In addition, a new gene of unknown function, THO1, was identified as a multicopy suppressor of hpr1⌬in two different screenings (29, 45). No phenotype has been associated so far with the null mutants of THO1 to provide clues about its in vivo function (45). Here we show that Tho1 is a novel nuclear hnRNP that binds RNA in vitro and is recruited to transcribed chromatin in vivo in a THO-complex- and RNA-dependent manner. Tho1 overexpression suppresses the recombination, RNA export, and gene expression defects of THO ⫺ mutants but not those of sub2⌬mutants. Multicopy SUB2 also suppresses the RNA export defect of hpr1⌬strains. The ability of Tho1 to suppress hpr1⌬resides in its RNA binding C-terminal region. Altogether, our study suggests that Tho1, like Sub2, assembles onto the nascent mRNA during transcription and that Tho1 and Sub2 can provide alternative ways for mRNP biogenesis in the absence of a functional THO complex. MATERIALS AND METHODS Strains. Yeast strains used were the previously described W303-1A (wild type) U678-1C (hpr1⌬::HIS3), AYW3-3C (hpr1⌬::HIS3), WRK-1C (tho2⌬::KanMX4), RK2-6C (tho2⌬::KanMX4), AW33-11Bu ⫺ (tho1⌬::TRP1) (45), sub2⌬(sub2⌬::HIS3) (27, 29, 38), Sub2-TAP (SUB2-TAP) (53), MFM67-13A (mft1⌬::KanMX4) (10) and WMK-2A (mft1⌬::KanMX4) (10), and the BY4741 (wild type) and its isogenic thp2⌬ mutant BY-HR167 (EUROSCARF, Frankfurt, Germany) (10), plus the strains constructed for this study, SUL-1A (MAT␣ade2 his3 trp1 ura3 sub2⌬HIS3 leu2-k:: URA3-ADE2::leu2-k), BSU-S2T-3B (MAT␣his3 trp1 ura3 leu2-k::URA3-ADE2:: leu2-k SUB2-TAP), BSU-S2T-6D (MATahis3 trp1 ura3 hpr1⌬::HIS3 SUB2-TAP), WWT1T and WWT2T (W303 carrying TAP-tagged Tho1 and Tho2, respectively), WHT1T (hpr1⌬::HIS3) carrying TAP-tagged Tho1, WWYL.1D (wild type) and WHYL.2A (hpr1⌬::HIS3) carrying the GAL1pr::YLR454c fusion (53), and NALT-1 and NALT-2 (mutant nab2-1and double mutant tho1⌬nab2-1), obtained by genetic crosses between AW33-11Bu ⫺ and nab2-1(20). The chromosomal THO1-TAP fusion was constructed by transforming W303-1A with the 1.17-kb, 5⬘-end TAP-tagged THO1 sequence obtained by PCR on plasmid pBS1479 (48) using the primers 5⬘-TCCAGAGTAAGTAAAAACAGGAGAGGC AACCGCTCTGGTTACAGAAGATCCATGGAAAAGAGAAG-3⬘and 5⬘AAA GAACCGAAACTAGAATGAAAAACTCCACCAAAACGGCTTGAGCCTTC GACTCACTATAGGG-3⬘. Constructions of SUB2-TAP and THO2-TAP were described previously (29, 53). Plasmids. Plasmids pCM184-LAUR and p416GAL1lacZ have been described previously (9, 29). Plasmid pT7-HA-THO1, containing an N-terminal His 6 - hemagglutinin (HA) translational fusion of the complete THO1 coding sequence under the control of the T7 RNA pol promoter, was constructed by inserting into the NheI-ClaI site of pT7-7-His 6 a 0.72-kb HA-THO1 fragment obtained by PCR using 5⬘-GGGTACATGTGCTAGCTATCCCTATGACGTTCCCGACTATGC TATGGCAGATTATTCTTC-3⬘and 5⬘-TACTCCAAGCTTATCGATGCCTT TATCTTCTGTAACCAG-3⬘as primers. Plasmids pT7-HA-5⬘THO1 and pT7- HA-3⬘THO1 are identical to pT7-HA-THO1 except for containing the 0.21-kb 5⬘-end and 0.54-kb 3⬘-end fragments of THO1, respectively. Such copies were obtained by PCR using 5⬘-GGGTACATGTGCTAGCTATCCCTATGACGTT CCCGACTA-3⬘and 5⬘-GCTAAGGCCCAAGCTTCCTGTGGAGAAACTTC ACTTTCAC-3⬘as primers for 5⬘THO1 and 5⬘-CGGCATGCAGCTAGCGAA CAAAACCAGGAACAAGGGTCAG-3⬘and 5⬘-TACTCCAAGCTTATCGAT GCCTTTATCTTCTGTAACCAGA-3⬘as primers for 3⬘THO1. Plasmids pGAL-THO1, pGAL-NTHO1, and pGAL-THO1C were constructed by inserting the respective XbaI-ClaI inserts from the pT7-7His 6 plasmids into p416GAL based on pRS416 containing the GAL1 promoter. Purification of recombinant Tho1 proteins from Escherichia coli.BL21(DE3) strains transformed with pT7-HA-THO1, pT7-HA-5⬘THO1, and pT7-HA- 3⬘THO1 were grown in LB medium with 100 ␮g/ml ampicillin to an optical density at 600 nm (OD 600 ) of 0.8. Expression of the T7 promoter was induced for 4 h by adding IPTG (isopropyl-␤-D-thiogalactopyranoside) to a final concentration of 1 mM. Proteins were purified from the lysates by affinity chromatography on a Ni-nitrilotriacetic acid column (QIAGEN) according to the manufacturer’s instructions and quantified by Lowry’s method. In situ mRNA export assays. In situ localization of poly(A) ⫹ RNA was performed as described previously (5) using a digoxigenin-labeled oligo(dT) 16 probe. In vitro RNA binding activity. RNA binding activity was determined in vitro by RNA band-shifting assays following previously published procedures. As 32 P- labeled and cold nucleic acid samples, we used the 90-mer RNA sequence from the polylinker of pBluescript-SK and the 60-mer single-stranded DNA (ssDNA) previously described (29). Chromatin immunoprecipitation. For chromatin immunoprecipitation (ChIP) analyses, either the yeast strains WW-T1T and WHT1T, WWT2T, and Sub2- TAP containing the THO1-TAP, THO2-TAP, and SUB2-TAP, respectively, or the yeast strains WWYL.1D and WHYL.2A carrying the GAL1pr::YLR454c fusion were used. For ChIP analyses with PMA1, strains were grown in yeast extract-peptone-dextrose at 30°C up to an OD 660 of 0.8. For ChIP with GAL1 and GAL1pr::YLR454c genes, strains were grown in synthetic complete (SC)–2% glycerol–2% lactate to an OD 660 of 0.5. The culture was split in two, one half was supplemented with 2% glucose (repressed transcription) and the other with 2% galactose (activated transcription). Samples were taken after 3 h. ChIP assays were performed as described previously (21). RNase treatment was made with 300 U of RNase T1 and 300 U of RNase A for2hat30°C before immunoprecipitation. For experiments with RNase treatment, the time for cross-linking was reduced from 15 to 4 min. Immunoprecipitations were performed by using immunoglobulin-Sepharose for tandem affinity purification (TAP)-tagged proteins and the monoclonal anti-Rpb1-CTD antibody 8WG16 (Berkeley Antibody Company) and protein A-Sepharose for RNAPII. The GFX purification system (Amersham) was used for the last DNA purification step. We used 20- to 27-bp oligonucleotides for the PCR amplification of four fragments of GAL1, five fragments of PMA1, and three fragments of YLR454c, as indicated, and of the bp 9716 to 9863 intergenic region of chromosome V that was used as a negative control. The PCR products obtained were electrophoresed in a 15% acrylamide gel, stained with ethidium bromide, and quantified in a Fuji FLA-3000. For quantification of each ChIP assay, the linear range of each PCR amplification was first established with serial dilutions of the total input (I) and immunoprecipitated (P) DNA. Calculations and normalization of each ChIP were carried out as previously described (4, 30, 33). Briefly, we first performed PCR amplification experiments using different DNA dilutions, so that conditions were established to perform all definitive PCR amplification experiments in the linear range (see Fig. S1 in the supplemental material). The enrichment of the PCR amplification of interest above background levels was calculated by dividing the ratio “gene-specific P signal/intergenic P signal” between the ratio “gene-specific I signal/intergenic I signal.” All experiments were performed at least three times within the linear range with similar results. Miscellaneous. Recombination frequencies were obtained from two to three different transformants for each genotype tested. The median frequency of 6 to 12 independent values for each transformant was determined as reported previously (29). Northern analyses were performed according to standard procedures. RESULTS Multicopy THO1 and SUB2 suppress all phenotypes of hpr1⌬mutants. For the isolation of multicopy suppressors of hpr1⌬, we used the LAUR gene expression system containing a 4.15-kb lacZ-URA3 translational fusion under the control of the tet promoter (29). Whereas wild-type cells can express this fusion construct properly, this is not the case for THO mutants. Thus, hpr1⌬cells carrying the LAUR system were unable to form colonies on synthetic medium lacking uracil (SC-Ura- Trp-Leu) and did not express ␤-galactosidase activity. With this assay, we obtained different overlapping clones that defined only two genes that suppressed hpr1⌬in multicopy, SUB2, whose analysis was previously reported (29), and THO1, a novel gene of unknown function that we have previously 4388 JIMENO ET AL. MOL.CELL.BIOL. on June 29, 2017 by USE/BCTA.GEN UNIVERSITARIAhttp://mcb.asm.org/Downloaded from found as a multicopy suppressor of the hpr1⌬thermosensitivity and gene expression phenotype (45). In silico analysis revealed that Tho1 has a N-terminal DNA binding domain, termed the SAP domain, that is conserved in eukaryotes and present in a number of scaffold-associated proteins involved in RNA and DNA metabolism (7) (see Fig. S2 in the supplemental material). Interestingly, our in silico analysis revealed that the human ortholog of Tho1 is likely the Hcc-1 protein (see Fig. S2 in the supplemental material), which was identified as a highly expressed protein in human hepatocarcinoma and which has been shown in two-hybrid assays to bind the two DEAD box putative RNA helicases, hSUB2/UAP56/BAT1 and DDX39 (12, 35). As previously shown for multicopy SUB2, we show that hpr1⌬cells carrying the LAUR system and THO1 in multicopy FIG. 1. Suppression of the gene expression, hyperrecombination, and RNA export phenotypes of hpr1⌬by multicopy THO1 and SUB2. (A) Analysis of the capacity of hpr1⌬(U678-4C) and sub2⌬(sub2⌬) cells carrying the P tet ::lacZ-URA3 construct to form colonies in SC lacking Ura (sub2⌬cells are slow-growing but viable null sub2 Ura ⫺ mutants (29) derived from W303 isogenic strain DLY23 (27, 38). The cells analyzed were transformants with the empty multicopy vector YEp351 (YEp, negative control) or YEp351-HPR1 (YEp-HPR1, positive control), YEp351-THO1 (YEp-THO1) or YEp351- SUB2 (YEp-SUB2) carrying the HPR1,THO1, and SUB2 genes, respectively. The top panel shows a scheme of the P tet ::lacZ-URA3 translational fusion used to study expression of lacZ and URA3. (B) Northern analyses of the same hpr1⌬transformant strains shown in panel A. RNA was isolated from mid-log-phase cultures grown in SC lacking Trp and Leu. As a 32 P-labeled DNA probe we used the 3-kb BamHI lacZ fragment and an internal 589-bp 25S rRNA fragment obtained by PCR. mRNA quantifications were performed in a Fuji FLA3000 and normalized with respect to the rRNA levels of each sample, and are given in arbitrary units (AU) (C) Recombination frequency of the intrachromosomal direct-repeat system leu2-k::URA3-ADE2:: leu2-k in hpr1⌬(U678-4C) and sub2⌬(SUL-1A) mutants transformed with multicopy vector YEp351 and YEp351-THO1. The top panel shows a scheme of the recombination assay used. Recombination frequencies are the median value of a total of 12 independent colonies obtained from SC lacking Leu. Ura ⫺ recombinants were selected in SC plus 5-fluoroorotic acid plates. (D) Subcellular localization of poly(A) ⫹ RNAs detected in situ with digoxigeninlabeled oligo(dT) 16 . Samples were taken from mid-log-phase cultures of hpr1⌬(U678-4C) cells transformed with either YEp351, YEp351-THO1 or YEp351-SUB2 and shifted to 37° for4hinSClacking Leu. Nuclei (DNA) were visualized by staining with 10 ␮g/ml DAPI (4⬘,6⬘-diamidino-2- phenylindole). Analyses were performed with an Olympus AHBT3 microscope. VOL. 26, 2006 Tho1 AND mRNP BIOGENESIS 4389 on June 29, 2017 by USE/BCTA.GEN UNIVERSITARIAhttp://mcb.asm.org/Downloaded from recovered the abilities to grow on SC-Ura-Trp-Leu and express ␤-galactosidase activity (Fig. 1A; data not shown). This suppression occurs at the level of mRNA accumulation, as shown by Northern analysis. As can be seen in Fig. 1B, hpr1⌬cells were able to significantly increase the levels of lacZ-URA3 mRNA accumulation when expressing THO1 in multicopy. In addition to the gene expression defect, two other hallmark phenotypes of THO mutants are the hyperrecombination and mRNA export defect phenotypes. If multicopy THO1 suppressed the absence of the THO complex in the cell directly and not just the gene expression defect of hpr1⌬cells in an indirect manner, the recombination and mRNA export phenotypes of hpr1⌬might also be suppressed. To test this, we determined the frequency of direct-repeat recombination in the chromosomal leu2-k::URA3-ADE2::leu2-k system and the subcellular localization of bulk poly(A) ⫹ RNA in mutant cells transformed with the YEp351-THO1 multicopy plasmid. As can be seen in Fig. 1C, recombination levels in hpr1⌬cells carrying THO1 in multicopy were sevenfold lower than those in hpr1⌬cells transformed with the empty multicopy vector FIG. 2. Suppression of different phenotypes of THO mutants by multicopy THO1. (A) Northern analyses of tho2⌬(RK2-6C), mft1⌬(WMK- 2A) cells carrying the P tet ::lacZ-URA3 construct, and thp2⌬(BY-HR167) carrying the GAL1pr::lacZ construct under transcription activation conditions. (B) Recombination frequency of the intrachromosomal direct-repeat system leu2-k::URA3-ADE2::leu2-k mft1⌬(MFM67-13A) and tho2⌬(WRK-1C) mutants transformed with multicopy vector YEp351 and YEp351-THO1. Details are as described for Fig. 1. (C) Suppression of the thermosensitivity (ts) phenotype of hpr1⌬,tho2⌬,mft1⌬and thp2⌬by multicopy THO1. Viability of hpr1⌬(U678-4C) tho2⌬(WRK-1C), mft1⌬(MFM67-13A), and thp2⌬(BY-HR167) cells transformed with either YEp351 or YEp351-THO1. As controls, we used wild-type (WT) W303-1A and BY4741. As they gave identical results, only one is shown. Transformants were spotted as 10-fold serial dilutions on selective medium. Photographs were taken after 3 or 5 days at 30 or 37°C, respectively, as indicated. 4390 JIMENO ET AL. MOL.CELL.BIOL. on June 29, 2017 by USE/BCTA.GEN UNIVERSITARIAhttp://mcb.asm.org/Downloaded from YEp351. In a previous study (45), we did not detect a significant suppression of hyperrecombination of the first THO1 clone obtained. Apparently, as we now found a sixfold reduction in the recombination of hpr1⌬using a different E. coliamplified clone from that study (data not shown), the E. coliamplified clone used in that previous study accumulated some mutation that limited its capacity to reduce hyperrecombination. In contrast, multicopy THO1 was able to suppress neither the gene expression defect (Fig. 1A) nor the high levels of recombination (Fig. 1C) of sub2⌬cells. It is important to note that we used slow-growing but viable sub2⌬strains isogenic to W303-1A that were previously shown to be viable (38). The ability of multicopy SUB2 to suppress the hyperrecombination of hpr1⌬cells was reported previously (13). In situ localization of poly(A) ⫹ RNA by digoxigenin-labeled oligo(dT) revealed that hpr1⌬cells accumulated poly(A) ⫹ RNA in the nucleus after4hofshifting to 37° if transformed with YEp351 and in both the nucleus and the cytoplasm if transformed with multicopy YEp351-THO1 (Fig. 1D). Similar results were obtained with multicopy YEp351-SUB2 carrying SUB2 (Fig. 1D). Therefore, the RNA export defect of hpr1⌬ cells was also suppressed by multicopy THO1 and SUB2. Multicopy THO1 suppresses all mutants of the THO complex. In order to determine whether THO1 suppression was specific to hpr1⌬or general to other mutants of the THO complex, we tested the ability of multicopy THO1 to suppress different phenotypes of the other three THO-null mutants, tho2⌬,mft1⌬, and thp2⌬. As can be seen in Fig. 2, multicopy THO1 partially suppressed the incapacity of tho2⌬and mft1⌬ to express the lacZ-URA3 construct to different degrees as determined by uracil prototrophy (data not shown) and Northern analysis (Fig. 2A). The suppression of gene expression phenotypes was also observed in thp2⌬strains (Fig. 2A), in which case Northern analyses were performed with the GAL1- lacZ system because the genetic markers of the thp2⌬strains were incompatible with those of the lacZ-URA3 construct. Hyperrecombination was reduced 19.4- and 16.6-fold in tho2⌬ and mft1⌬, respectively (Fig. 2B). Finally, the ability of multicopy THO1 to suppress all THO mutants was tested with the thermosensitivity (ts) phenotype. As can be seen in Fig. 2C, the ts phenotype of all THO-null mutants (hpr1⌬,tho2⌬,mft1⌬, and thp2⌬) was suppressed by multicopy THO1. These results are in agreement with the observation that THO is a structural and functional unit (10) as suppression is observed for all phenotypes. Nevertheless, the variability observed in the intensity of suppression may reflect a difference in the in vivo relevance of each subunit for each phenotype tested, in particular in the presence of multicopy THO1. The suppression effect of multicopy THO1 was specific to the THO complex mutants, since null mutations of THP1 and SPT4, involved in RNA export and/or transcription and with effects on mRNA accumulation similar to those of THO mutants, were not suppressed by multicopy THO1 (data not shown). The fact that multicopy THO1 suppresses the different phenotypes of THO mutants and not those of sub2⌬suggests a functional relationship of Tho1 with THO rather than with TREX. Nevertheless, in contrast to THO mutations and to sub2⌬cells, tho1⌬did not have any effect on direct-repeat recombination and gene expression or on mRNA export (45; data not shown). hpr1⌬mutants show reduced processivity of transcription elongation that is suppressed by multicopy THO1 or SUB2. Mutants of the THO complex are impaired in transcription elongation as determined in vivo and in vitro (9, 41, 49). To specifically analyze transcription impairment in THO mutants in vivo independently of any possible effect in RNA stability, FIG. 3. Analysis of the distribution of RNAPII along the GAL1pr:: YLR454c fusion construct located at the endogenous YLR454c chromosomal locus. (A) ChIP analysis in wild-type (WT) and hpr1⌬isogenic strains. (B) ChIP analysis in hpr1⌬cells transformed with YEp351-HPR1, YEp351, YEp351-THO1, or YEp351-SUB2. Strains used were WWYL-1D (WT) and WHYL-2A (hpr1⌬). The scheme of the gene analyzed and the PCR-amplified fragments are shown. All experiments were performed with two different transformants. ChIPs were repeated three times for each sample. One representative acrylamide electrophoresis for each experiment is shown. The other fragments correspond to the DNA regions studied. The relative abundance of each DNA fragment was calculated as the ratio between each DNA fragment and the intergenic region quantification results of the precipitated fractions (P) normalized with respect to the corresponding ratios of the input fractions (I). The recruitment data shown are referred to the value of the 5⬘region, taken as 100%. Error bars indicate standard deviations. The PCR fragment corresponding to the bp 9716 to 9863 intergenic region of chromosome V used as a control is indicated by asterisks. m, middle. VOL. 26, 2006 Tho1 AND mRNP BIOGENESIS 4391 on June 29, 2017 by USE/BCTA.GEN UNIVERSITARIAhttp://mcb.asm.org/Downloaded from we determined RNAPII recruitment by ChIP analysis. This was performed at the 8-kb-long YLR454c gene fused to the GAL1 promoter in three different regions (5⬘end, middle, and 3⬘end). Consistent with recent results (41), Fig. 3A shows that, whereas more than 95% of RNAPIIs reached the 3⬘end of the gene in wild-type cells, this value was reduced to 50% in hpr1⌬ mutants, indicating that the processivity of transcription elongation is decreased in THO mutants. Importantly, Tho1 and Sub2 overexpression suppressed this transcription processivity defect of hpr1⌬cells (Fig. 3B). Altogether, the ability of Tho1 to suppress all phenotypes of THO ⫺ mutants indicates that the Tho1 and the THO complex must function at the same molecular level. Tho1 is a conserved nuclear protein that binds to actively transcribed chromatin in a THO- and RNA-dependent manner. As expected from the nuclear localization of the THO complex (10), Tho1 is predominantly nuclear, as we have shown with an N-terminal Tho1-GFP fusion (data not shown), and as has been observed in a high-throughput analysis (18). We next wondered if Tho1, as previously shown for THO/ TREX, was present in actively transcribed chromatin. This was determined by ChIP assays using yeast strains containing the TAP-tagged version of THO1 at the chromosomal locus. We studied the distribution of Tho1 at four different positions of the GAL1 locus: the promoter 5⬘end of the coding region, the middle part of the open reading frame, the 3⬘end of the open reading frame, and the untranslated 3⬘-end region (UTR) of the gene. As can be seen in Fig. 4A, there was no amplification of the DNA sequences analyzed when GAL1 was repressed (2% glucose), indicating no recruitment of Tho1 to inactive chromatin. However, the recruitment of Tho1 to GAL1 chromatin was observed when transcription was active. Recruitment was poor at the promoter 5⬘-end region and high in the rest of the regions tested. The largest recruitment of Tho1 was observed at the 3⬘end of the coding region, with its presence being reduced at the UTR. To test whether the recruitment of Tho1 was specific for the regulated GAL genes, we extended the ChIP analysis to the constitutively transcribed PMA1 gene. Five different regions of PMA1 were analyzed: the promoter, the 5⬘end, the middle part, and the 3⬘end of the coding sequence, and the 3⬘UTR. As can be seen in Fig. 4B, poor but detectable recruitment of Tho1 was observed at the promoter, the 5⬘end of the coding sequence, and the UTR, and as in the GAL1 gene, recruitment of Tho1 was higher toward the 3⬘end of the coding sequence. Therefore, we can conclude that Tho1 is recruited to active chromatin regardless of whether it is regulated (GAL1)or constitutive (PMA1). As has been shown recently for Sub2 (1), we wondered whether Tho1 recruitment to transcribed chromatin was dependent on the THO complex and on the nascent RNA. To assay whether the THO complex was required or whether it FIG. 4. Tho1 binds to actively transcribed chromatin in a THO- and RNA-dependent manner. The scheme of the genes analyzed and the PCR fragments amplified by PCR are shown. One representative acrylamide electrophoresis for each experiment is shown, with the PCR fragment corresponding to the bp 9716 to 9863 intergenic region of chromosome V used as a control indicated with asterisks. (A) ChIP analysis of Tho1 at the endogenous GAL1 gene in the wild-type strain WWT1T. (B) ChIP analysis of Tho1 at the endogenous PMA1 gene in the wild-type (WT) WWT1T strains carrying the TAP-tagged Tho1. (C) Importance of the THO complex for the recruitment of Tho1 to chromatin at the PMA1 gene, as determined by comparative ChIP analysis in the wild type (WWT1T) and hpr1⌬mutant (WHT1T) (D) Effect of RNase treatment on Tho1 and Tho2 recruitment to active chromatin at the PMA1 gene, as determined by modified ChIP analyses. These were performed in the wild-type WWT1T and WWT2T strains, carrying TAP-tagged Tho1 and Tho2, respectively, treated with (⫹) or without (⫺) RNase before immunoprecipitation. Cross-linking times were reduced to 4 min in either treated or nontreated samples. (E) Northern analyses of the expression of PMA1 in WT and hpr1⌬congenic strains RNA was isolated from mid-log-phase cultures grown in rich media. As a 32 P-labeled DNA probe, we used the 2.7-kb PMA1 fragment and an internal 589-bp 25S rRNA fragment, both obtained by PCR. Other details are as described for Fig. 1 and 2. AU, arbitrary units. Error bars indicate standard deviations. 4392 JIMENO ET AL. MOL.CELL.BIOL. on June 29, 2017 by USE/BCTA.GEN UNIVERSITARIAhttp://mcb.asm.org/Downloaded from enhanced recruitment of Tho1 to active chromatin, we compared the results of the ChIP analysis of Tho1-TAP in wildtype and hpr1⌬strains. As can be seen in Fig. 4C, Tho1 recruitment was clearly reduced in hpr1⌬cells, suggesting that the THO complex is required for Tho1 recruitment. To rule out the possibility that the THO dependency of Tho1 recruitment to chromatin was due to an incapacity of hpr1⌬to transcribe PMA1, we analyzed the levels of PMA1 transcription by Northern analysis and found that they were only slightly reduced in hpr1⌬versus wild-type cells (Fig. 4E). This was expected, as we had previously shown that the effect of hpr1⌬on transcription is significant in highly transcribed long or GC- rich DNA sequences (11), whereas PMA1 does not belong to this class of genes, and is consistent with previous Northern analyses of PMA1 in hpr1⌬strains (58). To determine whether the Tho1 recruitment is RNA dependent, a ChIP analysis was performed with or without RNases A and T1 before immunoprecipitation. Given that Tho2 recruitment was reported to be independent of RNA (1), we used a TAP-tagged Tho2 protein as a negative control. As can be seen in Fig. 4D, in RNase-treated samples, Tho1 recruitment to the PMA1 gene was abolished, whereas Tho2 was strongly recruited. As expected in these experiments, Sub2 recruitment was also reduced after RNase treatment (data not shown). Tho1 recruitment to transcribed chromatin is therefore dependent on THO and RNA. FIG. 5. In vitro RNA binding activity of Tho1. (A) Twelve percent sodium dodecyl sulfate-polyacrylamide gel electrophoresis of a His 6 -tagged version of the complete Tho1 protein and its truncated N-Tho1 and Tho1-C forms obtained from E. coli and stained with Coomassie blue. Proteins were purified from the BL21 E. coli strain transformed with pT7-HA-THO1, pT7-HA-5⬘THO1, and pT7-HA-3⬘THO1. A scheme of each protein purified is shown on the right, with the SAP domain shown in gray. (B) Band shift assays of the RNA binding activity of Tho1 using 5 ng of the 90-bp, 32 P-labeled RNA from the polylinker of pBluescript-SK (0.05 M 10 ⫺7 ) and increasing amounts of purified Tho1. (C) Competition assays of RNA binding activity using increasing amounts of the same 90-mer RNA molecule used before and a 60-mer ssDNA as cold competitor. (D) RNA binding assays with increasing equimolecular amounts of the N-Tho1 and Tho1-C truncated forms of Tho1. Other details are as described for panel B. (E) Comparison of the ability to suppress the thermosensitivity (ts) phenotype of hpr1⌬by overexpression of truncated Tho1 versions. Viability of hpr1⌬(U678-4C) cells transformed with either pGAL-THO1, pGAL-NTHO1, and pGAL-THO1C or the empty vector p416GAL (pGAL). Transformants were spotted on selective medium with glucose to control for cell number and on galactose-containing medium to induce gene expression. Other details are as described for Fig. 2. WT, wild type. VOL. 26, 2006 Tho1 AND mRNP BIOGENESIS 4393 on June 29, 2017 by USE/BCTA.GEN UNIVERSITARIAhttp://mcb.asm.org/Downloaded from In vitro RNA binding activity of Tho1. Given all of the previous results, we wondered if Tho1 had the ability to bind RNA in vitro. Since Tho1 has two structurally different halves, the N-terminal region that contains the SAP/SAF domain and the C-terminal half, with no known domain, we constructed His 6 -tagged versions of the full 218-amino-acid (aa) Tho1 protein, the 50-aa N-terminal region containing the SAP domain (N-Tho1), and the 168-aa C-terminal part lacking the SAP domain (Tho1-C). Such proteins were expressed in E. coli from the T7 promoter and purified with Ni-nitrilotriacetic acid columns in order to determine their nucleic acid binding properties in vitro. The three polypeptides were purified (Fig. 5A), and RNA binding was determined by gel-shifting assays using a 32 P- labeled 90-nucleotide RNA fragment synthesized in vitro. As can be seen in Fig. 5B, the full Tho1 protein binds to RNA. Band shifting was directly proportional to the amount of Tho1 protein used. This binding could be competed with cold RNA in a manner proportional to the amount of cold RNA added, whereas it was not competed by higher amounts of cold ssDNA (Fig. 5C). Therefore, binding to RNA was specific rather than a manifestation of unspecific binding to any type of singlestranded nucleic acid. We therefore performed a more detailed analysis to determine which region was conferring the main ability of the protein to bind to RNA using the N-Tho1- and Tho1-C-truncated versions. Figure 5D shows that although the N-Tho1 version containing only the SAP domain had a weak ability to bind RNA, the major RNA binding activity resided in the C-terminal part of the protein outside of the SAP domain, although no known RNA binding motif appears in this region. As expected from its N-terminal SAP domain, Tho1 also binds double-stranded DNA (see Fig. S3 in the supplemental material). Given these results, we tested the capacities of both truncated Tho1 proteins to suppress the hpr1⌬mutation. For this, we constructed new plasmids in which the N-Tho1- and Tho1-C-encoding DNA sequences were fused to the regulated GAL1 promoter. As can be seen in Fig. 5E, analysis of yeast hpr1⌬transformed with these plasmids revealed that, whereas N-Tho1 overexpression was unable to suppress the thermosensitivity (ts)ofhpr1⌬, Tho1-C overexpression suppressed this ts phenotype with a much better efficiency than did the whole Tho1 protein. Multicopy THO1 does not enhance Sub2 recruitment to active chromatin. It has been shown that THO enhances the ability of Sub2 to bind active chromatin, suggesting that THO helps recruit Sub2 to the site of transcription (58). We therefore wondered whether the suppression of THO mutations by multicopy THO1 was due to the capacity of Tho1 to help recruit Sub2 to active chromatin. To test this possibility, we assayed, by ChIP analysis, the ability of TAP-Sub2 to be recruited to the PMA1 gene in hpr1⌬cells transformed with the empty YEp351 vector or with the multicopy YEp351-THO1 carrying THO1. As can be seen in Fig. 6, the recruitment of Sub2 to the two different PMA1 regions tested was reduced to 40 and 20% of the wild-type levels, regardless of whether the hpr1⌬strains expressed THO1 in multicopy. Therefore, Tho1 does not stabilize or help to recruit Sub2 to the site of transcription. Genetic interaction of Tho1 with the Nab2 RNA binding protein. Thus far, we were not able to identify a phenotype associated with tho1⌬mutants and our identification of a role of Tho1 in mRNP biogenesis derives from the effects caused by Tho1 overexpression. Consequently, by directly studying the effect of tho1⌬in double mutant combinations, we next determined whether THO1 could genetically interact with other functions of mRNP biogenesis and export. We analyzed the viability of double mutants of tho1⌬and mutations in genes encoding different protein complexes that may act subsequently in the mRNP biogenesis pathway in which THO par- FIG. 6. Effect of multicopy THO1 on Sub2 recruitment to active chromatin at the PMA1 gene in hpr1⌬and wild-type strains. ChIP analyses were performed with strains BSU-S2T.3B (WT) and BSU-S2T.6D (hpr1⌬) transformed with plasmid YEp351-THO1, carrying THO1 in multicopy or the empty vector. Other details are as described for Fig. 3. Error bars indicate standard deviations. The PCR fragment corresponding to the bp 9716 to 9863 intergenic region of chromosome V used as a control is indicated by asterisks. 4394 JIMENO ET AL. MOL.CELL.BIOL. on June 29, 2017 by USE/BCTA.GEN UNIVERSITARIAhttp://mcb.asm.org/Downloaded from ticipates, such as sub2⌬,mex67-5,nab2-1, and thp1⌬. These single mutants show similar gene expression and hyperrecombination phenotypes to those of THO mutants (16, 29). All double mutants constructed were viable at 30°C (data not shown), as was also shown for tho1⌬hpr1⌬mutants (45). Interestingly, although tho1⌬nab2-1had no phenotype at the permissive temperature (30°C), tho1⌬suppressed the cryosensitivity (cs) phenotype of nab2-1cells. Thus, nab2-1 tho1⌬cells formed colonies at 16°C, whereas nab2-1did not (Fig. 7A), and the nuclear poly(A) ⫹ RNA accumulation phenotype observed in nab2-1mutants at 16°C was reduced in tho1⌬nab2-1double mutants (Fig. 7B). This result provides a demonstration that the function of wild-type levels of Tho1 is related to RNP biogenesis and export and establishes a functional link between Tho1 and the Nab2 poly(A) ⫹ -RNA binding protein. DISCUSSION This study reveals that Tho1, a SAF-box/SAP domain protein, is a nuclear hnRNP with a strong capacity to bind RNA at its C-terminal half and that has a functional relationship to THO and Sub2 in cotranscriptional mRNP biogenesis. Multicopy THO1, like SUB2, suppresses all phenotypes of THO- complex mutants. As is the case for Sub2, Tho1 is specifically recruited to active chromatin in a RNA- and THO-dependent manner. The functional relationship of Tho1 with the THO complex suggests a function of Tho1 in mRNP biogenesis/processing that is supported by the ability of Tho1 to bind RNA in vitro. Despite the functional interconnection of Sub2 and Tho1, the primary functions of both proteins may be different. Thus, sub2⌬cells are either sick and thermosensitive or inviable, depending on genetic background (19, 38), and show gene expression and RNA export defects and a strong hyperrecombination phenotype (13, 29). tho1⌬mutants are viable and show no phenotypes and, whereas hpr1 sub2 mutants are inviable, hpr1 tho1 and sub2 tho1 are viable (45; data not shown). Either the function of Tho1 in mRNP metabolism is redundant or dispensable. Nonetheless, other genes with a putative role in mRNP biogenesis and export for which only their overexpression and not their null mutation have a phenotype have been defined. An example of that is Gbp2, which interacts physically with the TREX complex (25, 53, 57). Sub2 may be an RNA-dependent ATPase, as shown for the human ortholog UAP56 (51) that forms, with the major RNA binding protein Yra1, a heterodimer that physically associates with THO in a larger complex termed TREX. The requirement of Sub2-Yra1 in RNA export has been well established (52), and the associated phenotypes of gene expression defect and transcription-dependent hyperrecombination conferred by their mutations suggest that Sub2 function is essential for the cotranscriptional formation of RNA export-competent mRNPs. The association of Sub2 and THO with active chromatin is consistent with this idea. Although THO facilitates the recruitment of Sub2 to active chromatin, multicopy THO1 does FIG. 7. Deletion of THO1 rescues the cs phenotype and the mRNA export defects of nab2-1. (A) Tenfold serial dilutions of nab2-1(NALT-1) and tho1⌬nab2-1(NALT-2) were spotted on selective plates and incubated at 16°C and 30°C for 5 days. (B) Subcellular localization of poly(A) ⫹ RNA detected in situ with digoxigenin-labeled oligo(dT) 16 . Other details are as described for Fig. 1. VOL. 26, 2006 Tho1 AND mRNP BIOGENESIS 4395 on June 29, 2017 by USE/BCTA.GEN UNIVERSITARIAhttp://mcb.asm.org/Downloaded from