Both KH and non-KH domain sequences are required for polyribosome asso
http://www.100md.com
《核酸研究医学期刊》
Department of Human Genetics, 1 Graduate Program in Genetics and Molecular Biology and 2 Department of Biochemistry, Emory University School of Medicine, Atlanta, GA 30322, USA
* To whom correspondence should be addressed. Tel: +1 404 727 3924; Fax: +1 404 727 3949; Email: jfridov@emory.edu
ABSTRACT
Scp160p is a 160 kDa RNA-binding protein in yeast previously demonstrated to associate with specific messages as an mRNP component of both soluble and membrane-bound polyribosomes. Although the vast majority of Scp160p sequence consists of 14 closely spaced KH domains, comparative sequence analyses also demonstrate the presence of a potential nuclear localization sequence located between KH domains 3 and 4, as well as a 110 amino acid non-KH N-terminal region that includes a potential nuclear export sequence (NES). As a step toward investigating the structure/function relationships of Scp160p, we generated two truncated alleles, FLAG.SCP160N1, encoding a protein product that lacks the first 74 amino acids, including the potential NES, and FLAG.SCP160C1, encoding a protein product that lacks the final KH domain (KH14). We report here that the N-truncated protein, expressed as a green fluorescent protein fusion in yeast, remains cytoplasmic, with no apparent nuclear accumulation. Biochemical studies further demonstrate that although the N-truncated protein remains competent to form RNPs, the C-truncated protein does not. Furthermore, polyribosome association is severely compromised for both truncated proteins. Perhaps most important, both truncated alleles appear only marginally functional in vivo, as demonstrated by the inability of each to complement scp160/eap1 synthetic lethality in a tester strain. Together, these data challenge the notion that Scp160p normally shuttles between the nucleus and cytoplasm, and further implicate polyribosome association as an essential component of Scp160p function in vivo. Finally, these data underscore the vital roles of both KH and non-KH domain sequences in Scp160p.
INTRODUCTION
One of the fundamental questions in molecular biology concerns the role of RNA-binding proteins in mediating the processes of post-transcriptional gene expression in eukaryotes (1). Recent studies in a variety of experimental systems have identified large numbers of RNA-binding proteins, many of which have been conserved across broad expanses of evolutionary time (1,2). Nonetheless, the precise biological roles and structure/function relationships for many of these proteins remain unknown (2). We report here a significant step toward uncovering the structure/function relationships of Scp160p, a 160 kDa RNA-binding protein from Saccharomyces cerevisiae.
First identified in 1995, Scp160p was originally hypothesized to function in the maintenance of ploidy in yeast, due in large part to the null phenotype, which includes abnormal cell size and shape, increased DNA content and aberrant segregation of genetic markers through meiosis (3). Biochemical studies from our laboratory and others quickly challenged this hypothesis, however, demonstrating that Scp160p binds selectively in vitro to ribohomopolymers and to rRNA, but not to tRNA (4), and that it associates in vivo with a specific subset of mRNAs and other proteins to form large mRNP complexes (5–8). Some of these mRNPs remain free in the cytosol, but most co-fractionate with soluble or membrane-bound polyribosomes (5–8). In addition, loss or mutation of EAP1, which encodes an eIF4E binding protein involved in both translation initiation and spindle pole body function, is synthetically lethal in combination with loss of SCP160 (8). Finally, Gou et al. (9) recently identified Scp160p as a potential effector of G-mediated signal transduction in yeast, although the mechanism and the extent of this function remain unclear. Combined, these data implicate Scp160p for a role in mediating the post-transcriptional regulation of specific mRNAs, as well as the function of specific mRNP proteins. The scp160-null phenotype may therefore reflect the downstream effects of aberrant expression or function of Scp160p's target mRNAs and proteins, rather than a direct role of Scp160p in many distinct cellular processes.
Although little direct structural information regarding Scp160p has been reported, sequence alignment studies have revealed the presence of 14 predicted hnRNP K homology, or KH domains (4,10,11). KH domains, each spanning about 70 amino acids in length, and including a conserved pattern of hydrophobic residues, an invariant Gly-X-X-Gly segment, and a variable loop, are one of the most common motifs found in the RNA-binding proteins of both prokaryotes and eukaryotes, including humans (11,12). Of the 14 KH domains in Scp160p, seven are classical, containing the Gly-X-X-Gly sequence, and seven are divergent, with insertions or deletions interrupting the Gly-X-X-Gly sequence (4). Recently, Baum et al. (13) probed the significance of the final four KH domains of Scp160p by testing the functionality of alleles missing either KH domains 13 and14 or 11–14. Consistent with the results reported here, both truncated proteins failed to associate with polyribosomes in yeast.
In large part because of its KH domains, Scp160p demonstrates significant sequence homology (23% amino acid sequence identity and 40% similarity) to a growing family of multiple KH-domain proteins collectively known as vigilins. First identified in chicken (14), vigilin homologs have now been reported in species ranging from Neurospora crassa (Genpept #7899383) to humans (15). Although all vigilins appear to bind nucleic acid, the nature and specificity of the ligands, as well as the proposed cellular functions, remain diverse (14,16–24). Whether these disparities represent true evolutionary divergence, or the vagaries of different experimental systems and approaches, remains unclear.
In addition to its many KH domains, Scp160p also contains a potential nuclear localization sequence (NLS) positioned between KH domains 3 and 4, and a 110 amino acid non-KH domain N-terminal region including a potential nuclear export sequence (NES; amino acids 52–61) (4,10). Neither the NLS nor the NES has been functionally confirmed. Nonetheless, the potential presence of both an NLS and an NES raises the possibility that Scp160p may be a nuclear/cytoplasmic shuttle protein, as has been seen for a number of other KH domain RNA-binding proteins . Other than the 110 amino acid N-terminal region, the largest stretch of non-KH domain sequence in Scp160p is 8 amino acids.
As a step toward elucidating the structure/function relationships of Scp160p, we have generated two truncated alleles, FLAG.SCP160N1, whose protein product lacks the first 74 amino acids, including the putative NES, and FLAG.SCP160C1, whose protein product lacks the final KH domain (KH14), and expressed both proteins in yeast. Our results demonstrate that the N-truncated protein, expressed as a green fluorescent protein (GFP)-fusion, continues to localize to the cytoplasm, with no apparent nuclear accumulation. Considering that the putative NLS remains intact in this truncated protein, which now lacks an NES, these results clearly challenge the assumption that wild-type Scp160p is a nuclear-cytoplasmic shuttle protein.
Size fractionation studies demonstrate that the N-truncated protein remains competent to form RNPs of normal size, although the C-truncated protein does not. Interestingly, polyribosome association is severely compromised in both, with the N-truncated protein demonstrating no detectable polyribosome association, and the C-truncated protein demonstrating a biphasic distribution with only a small fraction remaining associated with polyribosomes. Perhaps most important, both truncated alleles appear only marginally functional, as measured by the inability of each to complement scp160/eap1 synthetic lethality in a tester strain. Together, these data further implicate polyribosome association as an essential component of Scp160p function, and underscore the vital roles of both the KH and non-KH domain sequences in Scp160p.
MATERIALS AND METHODS
Yeast strains and manipulation
All yeast manipulations were performed according to standard protocols (26). The strains utilized in this study are listed in Table 1 and as indicated, all were derived either from the haploid parent strain JJ52 or from the haploid parent strain W303 .
Table 1. Yeast strains used in this study
Construction and expression of the SCP160N1 allele
The FLAG-SCP160N1 allele was generated by cutting a wild-type subclone PstI–ApaI, dropping out the 200 bp N-terminal fragment, and ligating in the annealed oligonucleotides SCPATF1 (5'-GGATCCAAAATGGACTACAAGGACGACGACGACAAGGGCC-3') and SCPATR1 (5'-CTTGTCGTCGTCGTCCTTGTAGTCCATTTTGGATCCTGCA-3'). This manipulation resulted in loss of the first 74 codons from the SCP160 open reading frame (ORF), with the concomitant substitution of sequences encoding a starting methionine followed by a FLAG epitope (Asp-Tyr-Lys-Asp-Asp-Asp-Asp-Lys). The PstI restriction site immediately upstream of the ORF had been introduced by prior manipulation of the wild-type sequence. Strains of yeast expressing either FLAG-tagged wild-type Scp160p or FLAG-tagged Scp160N1p in place of the wild-type protein were generated by two-step gene replacement, as described previously (5,6,26). As indicated in Table 1, strain JFy4619, which expresses FLAG.Scp160N1p from a genomic allele, was functionally covered by a URA3 plasmid-borne wild-type allele, which was eliminated from the cells by 5FOA selection just prior to analysis.
C-terminal GFP tags were also added to both the wild-type and Scp160N1p sequences, as described previously (6). The GFP-tagged alleles (JF2140 and JF2592, respectively) were confirmed by DNA sequencing, linearized with BstEII, and integrated into the SCP160 locus for expression. Strains expressing GFP-tagged isoforms of Scp160p were further confirmed by western-blot analysis using an affinity-purified anti-GFP polyclonal antibody (SIGMA G1544) at 0.1 μg/ml, as recommended by the manufacturer. Yeast expressing FLAG-Scp160p.GFP or FLAG.Scp160N1p.GFP were used for microscopy (Figure 2), but not for any of the other experiments reported here.
Figure 2. Fluorescence microscopy of yeast expressing either full-length FLAG.Scp160p.GFP or FLAG.Scp160N1p.GFP. All images were viewed at 100x magnification. The left-most panel in each row presents the GFP signal, indicating distribution of the tagged Scp160p protein, the middle panel in each row presents the corresponding Hoescht dye signal, indicating nuclear location, and the right-most panel in each row presents the DIC image. As discussed in the text, no nuclear accumulation was observed in the truncated Scp160p protein despite loss of the putative NES.
Construction and expression of the SCP160C1 allele
The FLAG.SCP160C1 allele was generated initially using one-step gene replacement in the haploid strain JFy4493, which already encoded a FLAG-tag at the N-terminus of SCP160. In brief, the C-terminus of this allele was modified by insertion of a PCR-amplified kanMX cassette flanked by the appropriate SCP160 sequences (26). The oligonucleotides scpdelKH14kanf2 (5'-TCGATACTGCTGTTAAGTTGATTAAAGAAAGAATTGCCAAGGCACCATCTGCTACATAGgtttagcttgcctcgtcccc-3') and scpdel KH14kanr2 (5'-TATATAAGTAAGTAAAAGCCAAAATCTATATTGAAAAAAATTGGTTTCAAAGAGCTTGTtggatggcggcgttagtatcg-3') were used as primers to generate this integrating fragment. Following selection on G418 plates, colonies were confirmed by genomic PCR using the primers scpKH10f1 (5'-TGGAGAACGGACATAAGATGGT-3') and scpR5 (5'-CACCGCCTTATAACGAAGAC-3'), followed by direct sequencing of the PCR product. The resultant strain was designated JFy4834 (Table 2).
Table 2. Relative plasmid loss as a measure of in vivo FLAG.SCP160N1 and FLAG. SCP160C1 function in yeast
To obtain a plasmid-borne allele, we applied a gap-repair strategy using as gapped backbone the plasmid pYCPlac.SCP160 (JF4468) cut AflII/EagI, and as insert a PCR fragment generated using the primers scpKH10f1 (5'-TGGAGAACGGACATAAGATGGT-3') and scpR5 (5'-CACCGCCTTATAACGAAGAC-3') with genomic DNA from yeast JFy4834 as template. As above, positive clones were confirmed by PCR followed by DNA sequencing of relevant regions. The confirmed plasmid clone that was used in subsequent studies was designated JF4544.
Fluorescence microscopy
Prior to microscopy, cells expressing either FLAG.Scp160p.GFP or FLAG.Scp160N1p.GFP were grown to mid-log phase in synthetic medium with 2% dextrose. Cells were stained by incubation in 5 μM Hoescht dye (SIGMA# B2261, Bis Benzamide Hoechste #33342) for 30 min. Cells were viewed using an Olympus BX-60 epifluorescence microscope with a GFP optimized filter set (Chroma). Images viewed at 100x magnification in the GFP, Hoescht and DIC channels were captured in digital form using IPLab Spectrum software. The DIC images presented in Figure 2 were brightened using Adobe Photoshop, but no there was no digital enhancement of any of the other images.
Biochemical analysis of Scp160p
All cell lysis, subcellular fractionation and anti-FLAG affinity isolation procedures were performed as described previously (6,7). Western-blot analyses were also performed as described previously (6) using the monoclonal antibodies M2 (Roche) to detect FLAG-tagged proteins, HA-11 (Covance) to detect hemagglutinin (HA)-tagged proteins and the anti-GFP affinity-purified polyclonal antibody from Sigma (#G1544) to detect GFP-tagged proteins, as cited in the text.
Analysis of in vivo function of the FLAG.SCP160N1 and FLAG.SCP160C1 alleles
The test of SCP160 in vivo function was performed as described previously (8), using the scp160/eap1 double null tester strain JFy4247, transformed with the test plasmids pJF2542, which encodes FLAG.Scp160N1p, and pJF4544, which encodes FLAG.Scp160C1p. As described previously (8), plasmids encoding wild-type Scp160p (pJF2803 for 2μ, pJF4468 for CEN) versus no Scp160p (pJF1102 for 2μ, pJF1096 for CEN) were also included as positive and negative controls, respectively.
RESULTS
Expression of FLAG-tagged Scp160N1p and Scp160C1p in yeast
As a first step toward defining a structure/function map for Scp160p, we generated both an N-terminally truncated allele, FLAG.SCP160N1, in which the first 74 codons, including the potential NES (open inverted triangles, Figure 1A), were removed and replaced by a FLAG tag, and a C-terminally truncated allele, FLAG.SCP160C1, in which the final KH domain was removed. All 14 KH domains (shaded squares, Figure 1A), as well as the potential NLS(filled inverted triangles, Figure 1A), remained intact in FLAG.SCP160N1.
Figure 1. Expression of FLAG.Scp160N1p and FLAG.Scp160C1p in yeast. (A) Cartoons illustrating the domain structures of FLAG.Scp160p, FLAG-Scp160N1p and FLAG-Scp160C1p. The asterisk (*) represents the FLAG epitope, the circle represents the 110 amino acid N-terminal region, the open triangle represents the potential NES and the filled triangle represents the potential NLS. Each box represents one KH domain. (B) Anti-FLAG western-blot analysis of lysates from yeast expressing either untagged Scp160p (control), FLAG-Scp160p, FLAG.Scp160N1p or FLAG-Scp160C1p, as indicated. The filled arrow indicates the full-length and truncated Scp160p proteins; the open arrow indicates a 100 kDa anti-FLAG cross-reacting protein of unknown identity (5,6) that served as an internal control for loading of lanes.
Western-blot analysis of yeast expressing FLAG-tagged wild-type Scp160p, FLAG-tagged Scp160N1p, or FLAG-tagged Scp160C1p demonstrated comparable levels of protein, although as expected, the migration patterns of both truncated proteins differed slightly from that of the full-length protein (Figure 1B, filled arrow). Yeast expressing native, untagged Scp160p served as a negative control, and an 100 kDa cross-reacting band of unknown identity (Figure 1B, open arrow) served as a convenient internal control for the loading of lanes.
Subcellular localization of FLAG.Scp160N1p
To explore the subcellular localization of FLAG.Scp160N1p, we generated a C-terminally GFP-tagged isoform, and integrated the corresponding allele into the normal genomic locus, as described previously for the wild-type allele (6). As reported previously (3,6,27), GFP signal in the wild-type cells was predominantly if not completely cytoplasmic, with clear enrichment around the nuclear periphery, the site of the yeast endoplasmic reticulum (ER). Hoescht staining of these cells confirmed the position of the nucleus, and the corresponding DIC images revealed the overall outline of each cell (Figure 2). Like their wild-type counterparts, yeast expressing the truncated FLAG.Scp160N1p.GFP fusion protein showed GFP signal that was predominantly, if not completely cytoplasmic. No evidence of nuclear accumulation of the truncated protein was seen, suggesting that the potential NES in the deleted region is not normally responsible for the cytoplasmic localization of the full-length protein. These data therefore further suggest that wild-type Scp160p may not normally shuttle between the nucleus and the cytoplasm of yeast.
FLAG.Scp160N1p remains competent to form large RNPs in yeast, but FLAG.Scp160C1p does not
To probe the biochemical interactions of Scp160N1p and Scp160C1p, as compared with wild-type Scp160p, yeast expressing FLAG-tagged isoforms of all three proteins were grown to mid-log phase (OD600 1.0), lysed and the supernatants size-fractionated by S-300 size-exclusion column chromatography. The resultant fractions were subjected to western-blot analysis using the anti-FLAG monoclonal M2 as the primary antibody. As shown in Figure 3 (upper two panels), the majority of signal representing both the wild-type and N-truncated Scp160p proteins migrated at or near the void volume of the column, indicating that both proteins exist as components of large complexes (1300 kDa). As we have reported previously for wild-type Scp160p (5,6), pre-treatment of each lysate with RNase prior to size-fractionation reduced the apparent size of each protein to 500 kDa (Figure 3, bottom two panels), which is consistent with the previously demonstrated migration of purified native Scp160p (5). Interestingly, the C-truncated protein migrated as the smaller size even in the context of untreated lysates (Figure 3, third panel), demonstrating that this protein did not remain competent to form normal RNP complexes.
Figure 3. Size estimation of FLAG.Scp160p, FLAG.Scp160N1p and FLAG.Scp160C1p-containing complexes using S-300 gel filtration column chromatography. Lysates of cells expressing the indicated alleles of FLAG-tagged full-length or truncated SCP160 were subjected to S-300 gel filtration column chromatography, as described in Materials and Methods. The resultant fractions were subjected to western-blot analysis using the anti-FLAG-tag monoclonal antibody M2. Each panel presents the corresponding sections of two gels juxtaposed, one loaded with samples of column elution fractions from 30 to 50 ml, and the other with samples of fractions from 50 to 70 ml. The position of the void volume for this column, representing complexes of 1300 kDa, is indicated. The upper three panels represent cells lysed in the absence of RNase; the lower two panels represent lysates treated with RNase prior to chromatography. Although the results of only single representative experiments are presented, all procedures were repeated at least three times with indistinguishable results.
As a next step to probe the composition of the FLAG.Scp160N1p containing RNP complexes, we applied a previously described anti-FLAG affinity purification procedure (5–8) to isolate these complexes from yeast also expressing HA-tagged Bfr1p. Prior studies identified Bfr1p as one of many proteins that co-purify with Scp160p-containing mRNP complexes in yeast (6). As shown in Figure 4 (upper panels), and as reported previously for full-length FLAG.Scp160p (6), HA.Bfr1p co-purified with FLAG.Scp160N1p. Also like its wild-type counterpart, the association between FLAG.Scp160N1p and HA-Bfr1p was sensitive to RNase pre-treatment (Figure 4, middle set of panels). Finally, no HA-Bfr1p signal was detected following the mock-isolation of complexes from yeast expressing HA-tagged Bfr1p in conjunction with native, untagged Scp160p, demonstrating the specificity of the isolation procedure (Figure 4, lower panels). These data demonstrate that the FLAG.Scp160N1p protein retained the ability to form at least a subset of its normal macromolecular interactions.
Figure 4. FLAG.Scp160N1p associates with HA-Bfr1p in an RNA-dependent manner. Upper two panels: Scp160p-containing complexes were affinity isolated as described in Materials and Methods from lysates of yeast co-expressing FLAG.Scp160N1p and HA.Bfr1p, and subjected to western-blot analysis with both anti-FLAG (upper panel) and anti-HA monoclonal antibodies (bottom panel). As illustrated, although the affinity isolation procedure was directed against the FLAG.Scp160N1p protein alone, HA-Bfr1p protein co-eluted from the affinity column. Middle two panels: Scp160p-containing complexes were affinity isolated and characterized as described above, except that samples were pretreated with RNase prior to the affinity isolation procedure. As illustrated, the association between FLAG.Scp160N1p and HA.Bfr1p was RNase-sensitive. Bottom two panels: as a control for specificity of the isolation and elution procedure, a parallel isolation was performed on lysates of yeast co-expressing HA.Bfr1p with native, untagged Scp160p. As illustrated, although HA.Bfr1p protein was present in the lysates (void and flow through fractions), no HA.Bfr1p signal eluted from the anti-FLAG affinity column. Although the results of only single representative experiments are presented, both test and control experiments were repeated at least three times with indistinguishable results.
Both FLAG.Scp160N1p and FLAG.Scp160C1p are severely compromised in their ability to associate with polyribosomes
To probe the ability of each truncated protein to associate with polyribosomes, soluble cell lysates were size-fractionated by sucrose gradient centrifugation, as described previously (5,6). Lysates from yeast expressing the FLAG-tagged wild-type protein were analyzed in parallel as a control. By their nature, sucrose gradients enable separation of much larger complexes than do S300 columns, allowing the distinction of free RNPs from ribosomal subunits, monosomes and polysomes, as shown in Figure 5. Western-blot analyses of the fractions from these gradients (Figure 5) demonstrated that although the wild-type FLAG.Scp160p protein localized predominantly to the bottom of the gradient, co-fractionating as expected with polyribosomes, the FLAG.Scp160N1p protein localized predominantly to the lighter fractions. Furthermore, the FLAG.Scp160C1p protein localized predominantly to the very top of the gradient, although a small amount of this protein continued to co-fractionate with polyribosomes.
Figure 5. Analysis of FLAG.Scp160p, FLAG.Scp160N1p and FLAG.Scp160C1p-containing complexes using sucrose gradient fractionation. Upper panel: a representative tracing of the UV (254 nm) absorption profile from gradient fractions of wild-type cells, with the positions of the small and large ribosomal subunits, monosomes and polyribosomes indicated. The corresponding profiles from yeast expressing FLAG.Scp160N1p or FLAG.Scp160C1p were essentially indistinguishable. Bottom three panels: lysates of yeast expressing the indicated alleles of SCP160 were subjected to sucrose gradient fractionation, as described in Materials and Methods, followed by western-blot analysis of the fractions using the anti-FLAG antibody M2. Although the results of only single representative experiments are presented, experiments were repeated at least three times with indistinguishable results.
Differential centrifugation of cell lysates to separate soluble from membrane-associated pools also demonstrated a notable distinction between the full-length and truncated Scp160p proteins. As reported previously (4,6,27), wild-type FLAG.Scp160p localized predominantly to the membrane pellet, while both the FLAG.Scp160N1p and FLAG.Scp160C1p proteins were almost evenly distributed between the soluble and membrane-associated fractions (Figure 6A and B). This significant shift of both the FLAG.Scp160N1p and FLAG.Scp160C1p signals from the membrane-associated pool to the soluble pool is fully consistent with our previous observation that treatments which disrupt Scp160p association with polyribosomes also shift signal from the membrane-associated pellet to the soluble fraction (6).
Figure 6. Distribution of FLAG.Scp160p, FLAG.Scp160N1p and FLAG.Scp160C1p between the soluble and membrane-associated subcellular pools. (A) Lysates of cells expressing the indicated alleles of SCP160 were fractionated into total, soluble and membrane-associated pools by differential centrifugation, as described in Materials and Methods, followed by western-blot analysis of the fractions using the anti-FLAG antibody M2. (B) Fraction of Scp160p signal sedimenting with cell membranes. Averages ± SD (n) are plotted for each allele. The values obtained for both truncated proteins were significantly different from the corresponding wild-type value (t-test).
In vivo function of both truncated FLAG.SCP160N1 and FLAG.SCP160C1 alleles is severely compromised
To assess functional capacity of both the FLAG.SCP160N1 and FLAG.SCP160C1 alleles in vivo, we quantified the ability of each to complement scp160/eap1 synthetic lethality in a double null tester strain, as described previously (8). In brief, plasmid-borne copies of each truncated SCP160 allele were introduced into yeast genomically null for both scp160 and eap1, covered by a URA3 maintenance plasmid encoding wild-type SCP160. Following growth for a fixed number of generations in the absence of uracil selection, transformants were plated to medium with and without 5FOA. The fraction of cells that could grow in the presence of 5FOA, which kills only URA3+ cells, versus in its absence, served as an indicator of how many cells had lost their URA3 plasmid. Yeast transformed with either empty plasmid backbone, or plasmid carrying a second wild-type copy of SCP160, served as negative and positive controls, respectively. For comparison, the degree of URA3 plasmid loss in each strain was scaled relative to that seen in the corresponding positive control, which was set to 100%. The final values calculated, therefore, no longer directly reflected plasmid retention or loss, but rather indicated how effectively, relative to the wild-type allele, the test sequence in question could enable cells to remain viable despite loss of their URA3+ SCP160+ maintenance plasmid. Since this maintenance plasmid was the same in all strains, issues of plasmid replication efficiency or distribution, independent of SCP160 sequence, should have been cancelled out. As shown in Table 2, according to this procedure, the FLAG.SCP160N1 allele retained <10% wild-type function, and the corresponding FLAG.SCP160C1 allele retained <5% wild-type function.
DISCUSSION
The experiments presented here were designed to ask three questions: (i) is the N-terminal region of Scp160p, including the potential NES, responsible for the cytoplasmic localization of Scp160p; (ii) is the non-KH domain N-terminal region important for Scp160p macromolecular interaction or function; and (iii) is the final KH domain important for Scp160p macromolecular interaction or function? In effect, we have compared the relative functional contributions of KH and non-KH domain sequences in SCP160. Our results clearly demonstrate that the N-terminal region and NES are not required for cytoplasmic localization of Scp160p, or for assembly of Scp160p into RNPs, although this region is required for association of Scp160p with polyribosomes. In contrast, the final KH domain is required for both RNP assembly and association of Scp160p with polyribosomes. Perhaps most important, both the non-KH domain N-terminus and the C-terminal KH domain are essential for function of Scp160p in vivo. Together, these data underscore the importance of both KH and non-KH domain sequences in Scp160p, and further implicate polyribosome association as a key component of Scp160p function.
Subcellular localization
All studies reported to date addressing the subcellular localization of Scp160p have placed the protein firmly in the cytoplasm, with enrichment around the nuclear envelope/ER (3–8,24,27). Nonetheless, the identification of both putative NLS and NES sequences in Scp160p (4,10) raised the possibility that Scp160p might be a nuclear/cytoplasmic shuttle protein with a steady-state distribution favoring the cytoplasm. The results presented here, demonstrating that loss of the putative NES does not result in nuclear accumulation of the truncated Scp160p, clearly challenge this hypothesis, and further suggest that the remaining putative NLS also may be non-functional. Where in the cell, and at what stage of complex assembly the Scp160p protein finds and associates with its target mRNAs, therefore remains an open question, although presumably these interactions occur within the cytoplasm.
Macromolecular interactions
The data presented here demonstrate that the FLAG.Scp160N1p truncated protein was able to form a subset of its normal biochemical interactions, including association with HA-Bfr1p and assembly into RNPs, although it was severely compromised in its ability to associate with polyribosomes. Whether these data reflect a direct versus indirect role of the N-terminal sequence in polyribosome association remains unclear, as does whether the full N-terminal region is required for polyribosome association, versus a smaller sub-domain. Furthermore, beyond the presence of Bfr1p, the precise nature and composition of the RNP complexes formed by FLAG.Scp160N1p remains to be explored. At minimum, however, the fact that at least some normal biochemical interactions involving FLAG.Scp160N1p remain intact implies that the truncated protein is not globally misfolded. The observation that RNP formation and polyribosome association can be uncoupled further implies that polyribosome association is not an obligate first step in the process of Scp160p mRNP formation.
The data presented here also demonstrate that the FLAG.Scp160C1p truncated protein is compromised with regard to both RNP formation and polyribosome association. This result is striking for at least three reasons. First, it demonstrates either that Scp160p requires all 14 KH domains to function properly, or else that KH14 is special in some way. This is a testable hypothesis that is currently under investigation. Second, although Baum et al. (13) also reported recently that deletion of the final two KH domains from Scp160p disrupted polyribosome association, those authors concluded that sequences within KH13 were essential for this process, and never checked the impact of KH14-loss alone. Our results clearly demonstrate that the loss of KH14 alone is sufficient to disrupt Scp160p-polyribosome association, although the mechanism of that impact remains unclear. While the possibility of partial or global unfolding of the C-truncated protein cannot formally be excluded, the fact that this protein remains largely soluble (Figure 6) and is detected in lysates at levels comparable to that of the wild-type protein (Figure 1B), argues against this possibility. Finally, these results demonstrate that although the N-terminus may be essential for polyribosome association of Scp160p, it is not sufficient.
Function in vivo
One of the most important observations reported here is that both the FLAG.SCP160N1 and FLAG.SCP160C1 were significantly compromised with regard to function in vivo (Table 2). These data demonstrate that although the non-KH domain sequences in Scp160p may be minimal, they are not insignificant. These data further demonstrate that despite the presence of 13 other KH domains, KH14 is essential. Finally, as mentioned earlier, these data implicate polyribosome association as an essential component of Scp160p function in vivo.
ACKNOWLEDGEMENTS
We are grateful to Kristen Riehman, who helped to generate the FLAG-SCP160N1 allele, to Brian Lang, who performed some of the early characterization of this allele, and to Alice Watson, who helped to generate earlier versions of some of the strains and plasmids utilized in this study. This work was supported by funds from the National Science Foundation (Award 0112911 to J.L.F.-K.).
REFERENCES
Burd,G.G. and Dreyfuss,G. ( (1994) ) Conserved structures and diversity of functions or RNA-binding proteins. Science, , 265, , 615–621.
Dreyfuss,G., Kim,V.N. and Kataoka,N. ( (2002) ) Messenger-RNA-binding proteins and the messages they carry. Nature Rev. Mol. Cell Biol., , 3, , 195–205.
Wintersberger,U., Kuhne,C. and Karwan,A. ( (1995) ) Scp160p, a new yeast protein associated with the nuclear membrane and the endoplasmic reticulum, is necessary for maintenance of exact ploidy. Yeast, , 11, , 929–944.
Weber,V., Wernitznig,A., Hager,G., Harata,M., Frank,P. and Wintersberger,U. ( (1997) ) Purification and nucleic-acid-binding properties of a Saccharomyces cerevisiae protein involved in the control of ploidy. Eur. J. Biochem., , 249, , 309–317.
Lang,B.D. and Fridovich-Keil,J.L. ( (2000) ) Scp160p, a multiple KH-domain protein, is a component of mRNP complexes in yeast. Nucleic Acids Res., , 28, , 1576–1584.
Lang,B.D., Li,A.-M., Black-Brewster,H.D. and Fridovich-Keil,J.L. ( (2001) ) The brefeldin A resistance protein Bfr1p is a component of polyribosome-associated mRNP complexes in yeast. Nucleic Acids Res., , 29, , 2567–2574.
Li,A.-M., Watson,A. and Fridovich-Keil,J.L. ( (2003) ) Scp160p associates with specific mRNAs in yeast. Nucleic Acids Res., , 31, , 1830–1837.
Mendelsohn,B.A., Li,A.M., Vargas,C.A., Riehman,K., Watson,A. and Fridovich-Keil,J.L. ( (2003) ) Genetic and biochemical interactions between SCP160 and EAP1 in yeast. Nucleic Acids Res., , 31, , 5838–5847.
Guo,M., Aston,C., Burchett,S.A., Dyke,C., Fields,S., Rajarao,S.J., Uetz,P., Wang,Y., Young,K. and Dohlman,H.G. ( (2003) ) The yeast G protein alpha subunit Gpa1 transmits a signal through an RNA binding effector protein Scp160. Mol. Cells, , 12, , 517–524.
Currie,J.R. and Brown,W.T. ( (1999) ) KH domain-containing proteins of yeast: absence of a fragile X gene homologue. Am. J. Med. Genet., , 84, , 272–276.
Siomi,H., Siomi,M.C., Nussbaum,R.L. and Dreyfuss,G. ( (1993) ) The protein product of the fragile X gene, FMR1, has characteristics of an RNA-binding protein. Cell, , 74, , 291–298.
Gibson,T.J., Thompson,J.D. and Heringa,J. ( (1993) ) The KH domain occurs in a diverse set of RNA-binding proteins that include the antiterminator NusA and is probably involved in binding to nucleic acid. FEBS Lett., , 324, , 361–366.
Baum,S., Bittins,M., Frey,S. and Seedorf,M. ( (2004) ) Asc1p, a WD40-domain containing adaptor protein, is required for the interaction of the RNA-binding protein Scp160p with polysomes. Biochem. J., , 390, , 823–830.
Schmidt,C., Henkel,B., Poschl,E., Zorbas,H., Purschke,W.G., Gloe,T.R. and Muller,P.K. ( (1992) ) Complete cDNA sequence of chicken vigilin, a novel protein with amplified and evolutionary conserved domains. Eur. J. Biochem., , 206, , 625–634.
Plenz,G., Kugler,S., Schnittger,S., Rieder,H., Fonatsch,C. and Muller,P.K. ( (1994) ) The human vigilin gene: identification, chromosomal localization and expression pattern. Hum. Genet., , 93, , 575–582.
Cortes,A., Huertas,D., Fanti,L., Pimpinelli,S., Marsellach,F.X., Pina,B. and Azorin,F. ( (1999) ) DDP1, a single-stranded nucleic acid-binding protein of Drosophila, associates with pericentric heterochromatin and is functionally homologous to the yeast Scp160p, which is involved in the control of cell ploidy. EMBO J., , 18, , 3820–3833.
Cortes,A. and Azorin,F. ( (2000) ) DDP1, a heterochromatin-associated multi-KH-domain protein of Drosophila melanogaster, interacts specifically with centromeric satellite DNA sequences. Mol. Cell. Biol., , 20, , 3860–3869.
Cunningham,K.S., Dodson,R.E., Nagel,M.A., Shapiro,D.J. and Schoenberg,D.R. ( (2000) ) Vigilin binding selectively inhibits cleavage of the vitellogenin mRNA 3' untranslated region by the mRNA endonuclease polysomal ribonuclease 1. Proc. Natl Acad. Sci. USA, , 97, , 12498–12502.
Dodson,R.E. and Shapiro,D.J. ( (1997) ) Vigilin, a ubiquitous protein with 14 K homology domains, is the estrogen-inducible vitellogenin mRNA 3'-untranslated region-binding protein. J. Biol. Chem., , 272, , 12249–12252.
Dodson,R.E. and Shapiro,D.J. ( (1994) ) An estrogen-inducible protein binds specifically to a sequence in the 3' untranslated region of estrogen-stabilized vitellogenin mRNA. Mol. Cell. Biol., , 14, , 3130–3138.
Goolsby,K.M. and Shapiro,D.J. ( (2003) ) RNAi-mediated depletion of the 15 KH domain protein, vigilin, induces death of dividing and non-dividing human cells but does not initially inhibit protein synthesis. Nucleic Acids Res., , 31, , 5644–5653.
Grunweller,A., Purschke,W.G., Kugler,S., Kruse,C. and Muller,P.K. ( (1997) ) Chicken vigilin gene: a distinctive pattern of hypersensitive sites is characteristic for its transcriptional activity. Biochem. J., , 326, , 601–607.
Kugler,S., Grunweller,A., Probst,C., Klinger,M., Muller,P.K. and Kruse,C. ( (1996) ) Vigilin contains a functional nuclear localisation sequence and is present in both the cytoplasm and the nucleus. FEBS Lett., , 382, , 330–334.
Kugler,S., Plenz,G. and Muller,P.K. ( (1996) ) Two additional 5' exons in the human vigilin gene distinguish it from the chicken gene and provide the structural basis for differential routes of gene expression. Eur. J. Biochem., , 238, , 410–417.
Feng,Y., Gutekunst,C.A., Eberhart,D.E., Yi,H., Warren,S.T. and Hersch,S.M. ( (1997) ) Fragile X mental retardation protein: nucleocytoplasmic shuttling and association with somatodendritic ribosomes. J. Neurosci., , 17, , 1539–1547.
Guthrie,C. and Fink,G.R. ( (1991) ) Guide to yeast genetics and molecular biology. Meth. Enzymol., , 194, , 3–21, 281–302.
Frey,S., Pool,M. and Seedorf,M. ( (2001) ) Scp160p, an RNA-binding, polysome-associated protein, localizes to the endoplasmic reticulum of Saccharomyces cerevisiae in a microtubule-dependent manner. J. Biol. Chem., , 276, , 15905–15912.(Ai-min Li, Claudia A. Vargas, Melissa A.)
* To whom correspondence should be addressed. Tel: +1 404 727 3924; Fax: +1 404 727 3949; Email: jfridov@emory.edu
ABSTRACT
Scp160p is a 160 kDa RNA-binding protein in yeast previously demonstrated to associate with specific messages as an mRNP component of both soluble and membrane-bound polyribosomes. Although the vast majority of Scp160p sequence consists of 14 closely spaced KH domains, comparative sequence analyses also demonstrate the presence of a potential nuclear localization sequence located between KH domains 3 and 4, as well as a 110 amino acid non-KH N-terminal region that includes a potential nuclear export sequence (NES). As a step toward investigating the structure/function relationships of Scp160p, we generated two truncated alleles, FLAG.SCP160N1, encoding a protein product that lacks the first 74 amino acids, including the potential NES, and FLAG.SCP160C1, encoding a protein product that lacks the final KH domain (KH14). We report here that the N-truncated protein, expressed as a green fluorescent protein fusion in yeast, remains cytoplasmic, with no apparent nuclear accumulation. Biochemical studies further demonstrate that although the N-truncated protein remains competent to form RNPs, the C-truncated protein does not. Furthermore, polyribosome association is severely compromised for both truncated proteins. Perhaps most important, both truncated alleles appear only marginally functional in vivo, as demonstrated by the inability of each to complement scp160/eap1 synthetic lethality in a tester strain. Together, these data challenge the notion that Scp160p normally shuttles between the nucleus and cytoplasm, and further implicate polyribosome association as an essential component of Scp160p function in vivo. Finally, these data underscore the vital roles of both KH and non-KH domain sequences in Scp160p.
INTRODUCTION
One of the fundamental questions in molecular biology concerns the role of RNA-binding proteins in mediating the processes of post-transcriptional gene expression in eukaryotes (1). Recent studies in a variety of experimental systems have identified large numbers of RNA-binding proteins, many of which have been conserved across broad expanses of evolutionary time (1,2). Nonetheless, the precise biological roles and structure/function relationships for many of these proteins remain unknown (2). We report here a significant step toward uncovering the structure/function relationships of Scp160p, a 160 kDa RNA-binding protein from Saccharomyces cerevisiae.
First identified in 1995, Scp160p was originally hypothesized to function in the maintenance of ploidy in yeast, due in large part to the null phenotype, which includes abnormal cell size and shape, increased DNA content and aberrant segregation of genetic markers through meiosis (3). Biochemical studies from our laboratory and others quickly challenged this hypothesis, however, demonstrating that Scp160p binds selectively in vitro to ribohomopolymers and to rRNA, but not to tRNA (4), and that it associates in vivo with a specific subset of mRNAs and other proteins to form large mRNP complexes (5–8). Some of these mRNPs remain free in the cytosol, but most co-fractionate with soluble or membrane-bound polyribosomes (5–8). In addition, loss or mutation of EAP1, which encodes an eIF4E binding protein involved in both translation initiation and spindle pole body function, is synthetically lethal in combination with loss of SCP160 (8). Finally, Gou et al. (9) recently identified Scp160p as a potential effector of G-mediated signal transduction in yeast, although the mechanism and the extent of this function remain unclear. Combined, these data implicate Scp160p for a role in mediating the post-transcriptional regulation of specific mRNAs, as well as the function of specific mRNP proteins. The scp160-null phenotype may therefore reflect the downstream effects of aberrant expression or function of Scp160p's target mRNAs and proteins, rather than a direct role of Scp160p in many distinct cellular processes.
Although little direct structural information regarding Scp160p has been reported, sequence alignment studies have revealed the presence of 14 predicted hnRNP K homology, or KH domains (4,10,11). KH domains, each spanning about 70 amino acids in length, and including a conserved pattern of hydrophobic residues, an invariant Gly-X-X-Gly segment, and a variable loop, are one of the most common motifs found in the RNA-binding proteins of both prokaryotes and eukaryotes, including humans (11,12). Of the 14 KH domains in Scp160p, seven are classical, containing the Gly-X-X-Gly sequence, and seven are divergent, with insertions or deletions interrupting the Gly-X-X-Gly sequence (4). Recently, Baum et al. (13) probed the significance of the final four KH domains of Scp160p by testing the functionality of alleles missing either KH domains 13 and14 or 11–14. Consistent with the results reported here, both truncated proteins failed to associate with polyribosomes in yeast.
In large part because of its KH domains, Scp160p demonstrates significant sequence homology (23% amino acid sequence identity and 40% similarity) to a growing family of multiple KH-domain proteins collectively known as vigilins. First identified in chicken (14), vigilin homologs have now been reported in species ranging from Neurospora crassa (Genpept #7899383) to humans (15). Although all vigilins appear to bind nucleic acid, the nature and specificity of the ligands, as well as the proposed cellular functions, remain diverse (14,16–24). Whether these disparities represent true evolutionary divergence, or the vagaries of different experimental systems and approaches, remains unclear.
In addition to its many KH domains, Scp160p also contains a potential nuclear localization sequence (NLS) positioned between KH domains 3 and 4, and a 110 amino acid non-KH domain N-terminal region including a potential nuclear export sequence (NES; amino acids 52–61) (4,10). Neither the NLS nor the NES has been functionally confirmed. Nonetheless, the potential presence of both an NLS and an NES raises the possibility that Scp160p may be a nuclear/cytoplasmic shuttle protein, as has been seen for a number of other KH domain RNA-binding proteins . Other than the 110 amino acid N-terminal region, the largest stretch of non-KH domain sequence in Scp160p is 8 amino acids.
As a step toward elucidating the structure/function relationships of Scp160p, we have generated two truncated alleles, FLAG.SCP160N1, whose protein product lacks the first 74 amino acids, including the putative NES, and FLAG.SCP160C1, whose protein product lacks the final KH domain (KH14), and expressed both proteins in yeast. Our results demonstrate that the N-truncated protein, expressed as a green fluorescent protein (GFP)-fusion, continues to localize to the cytoplasm, with no apparent nuclear accumulation. Considering that the putative NLS remains intact in this truncated protein, which now lacks an NES, these results clearly challenge the assumption that wild-type Scp160p is a nuclear-cytoplasmic shuttle protein.
Size fractionation studies demonstrate that the N-truncated protein remains competent to form RNPs of normal size, although the C-truncated protein does not. Interestingly, polyribosome association is severely compromised in both, with the N-truncated protein demonstrating no detectable polyribosome association, and the C-truncated protein demonstrating a biphasic distribution with only a small fraction remaining associated with polyribosomes. Perhaps most important, both truncated alleles appear only marginally functional, as measured by the inability of each to complement scp160/eap1 synthetic lethality in a tester strain. Together, these data further implicate polyribosome association as an essential component of Scp160p function, and underscore the vital roles of both the KH and non-KH domain sequences in Scp160p.
MATERIALS AND METHODS
Yeast strains and manipulation
All yeast manipulations were performed according to standard protocols (26). The strains utilized in this study are listed in Table 1 and as indicated, all were derived either from the haploid parent strain JJ52 or from the haploid parent strain W303 .
Table 1. Yeast strains used in this study
Construction and expression of the SCP160N1 allele
The FLAG-SCP160N1 allele was generated by cutting a wild-type subclone PstI–ApaI, dropping out the 200 bp N-terminal fragment, and ligating in the annealed oligonucleotides SCPATF1 (5'-GGATCCAAAATGGACTACAAGGACGACGACGACAAGGGCC-3') and SCPATR1 (5'-CTTGTCGTCGTCGTCCTTGTAGTCCATTTTGGATCCTGCA-3'). This manipulation resulted in loss of the first 74 codons from the SCP160 open reading frame (ORF), with the concomitant substitution of sequences encoding a starting methionine followed by a FLAG epitope (Asp-Tyr-Lys-Asp-Asp-Asp-Asp-Lys). The PstI restriction site immediately upstream of the ORF had been introduced by prior manipulation of the wild-type sequence. Strains of yeast expressing either FLAG-tagged wild-type Scp160p or FLAG-tagged Scp160N1p in place of the wild-type protein were generated by two-step gene replacement, as described previously (5,6,26). As indicated in Table 1, strain JFy4619, which expresses FLAG.Scp160N1p from a genomic allele, was functionally covered by a URA3 plasmid-borne wild-type allele, which was eliminated from the cells by 5FOA selection just prior to analysis.
C-terminal GFP tags were also added to both the wild-type and Scp160N1p sequences, as described previously (6). The GFP-tagged alleles (JF2140 and JF2592, respectively) were confirmed by DNA sequencing, linearized with BstEII, and integrated into the SCP160 locus for expression. Strains expressing GFP-tagged isoforms of Scp160p were further confirmed by western-blot analysis using an affinity-purified anti-GFP polyclonal antibody (SIGMA G1544) at 0.1 μg/ml, as recommended by the manufacturer. Yeast expressing FLAG-Scp160p.GFP or FLAG.Scp160N1p.GFP were used for microscopy (Figure 2), but not for any of the other experiments reported here.
Figure 2. Fluorescence microscopy of yeast expressing either full-length FLAG.Scp160p.GFP or FLAG.Scp160N1p.GFP. All images were viewed at 100x magnification. The left-most panel in each row presents the GFP signal, indicating distribution of the tagged Scp160p protein, the middle panel in each row presents the corresponding Hoescht dye signal, indicating nuclear location, and the right-most panel in each row presents the DIC image. As discussed in the text, no nuclear accumulation was observed in the truncated Scp160p protein despite loss of the putative NES.
Construction and expression of the SCP160C1 allele
The FLAG.SCP160C1 allele was generated initially using one-step gene replacement in the haploid strain JFy4493, which already encoded a FLAG-tag at the N-terminus of SCP160. In brief, the C-terminus of this allele was modified by insertion of a PCR-amplified kanMX cassette flanked by the appropriate SCP160 sequences (26). The oligonucleotides scpdelKH14kanf2 (5'-TCGATACTGCTGTTAAGTTGATTAAAGAAAGAATTGCCAAGGCACCATCTGCTACATAGgtttagcttgcctcgtcccc-3') and scpdel KH14kanr2 (5'-TATATAAGTAAGTAAAAGCCAAAATCTATATTGAAAAAAATTGGTTTCAAAGAGCTTGTtggatggcggcgttagtatcg-3') were used as primers to generate this integrating fragment. Following selection on G418 plates, colonies were confirmed by genomic PCR using the primers scpKH10f1 (5'-TGGAGAACGGACATAAGATGGT-3') and scpR5 (5'-CACCGCCTTATAACGAAGAC-3'), followed by direct sequencing of the PCR product. The resultant strain was designated JFy4834 (Table 2).
Table 2. Relative plasmid loss as a measure of in vivo FLAG.SCP160N1 and FLAG. SCP160C1 function in yeast
To obtain a plasmid-borne allele, we applied a gap-repair strategy using as gapped backbone the plasmid pYCPlac.SCP160 (JF4468) cut AflII/EagI, and as insert a PCR fragment generated using the primers scpKH10f1 (5'-TGGAGAACGGACATAAGATGGT-3') and scpR5 (5'-CACCGCCTTATAACGAAGAC-3') with genomic DNA from yeast JFy4834 as template. As above, positive clones were confirmed by PCR followed by DNA sequencing of relevant regions. The confirmed plasmid clone that was used in subsequent studies was designated JF4544.
Fluorescence microscopy
Prior to microscopy, cells expressing either FLAG.Scp160p.GFP or FLAG.Scp160N1p.GFP were grown to mid-log phase in synthetic medium with 2% dextrose. Cells were stained by incubation in 5 μM Hoescht dye (SIGMA# B2261, Bis Benzamide Hoechste #33342) for 30 min. Cells were viewed using an Olympus BX-60 epifluorescence microscope with a GFP optimized filter set (Chroma). Images viewed at 100x magnification in the GFP, Hoescht and DIC channels were captured in digital form using IPLab Spectrum software. The DIC images presented in Figure 2 were brightened using Adobe Photoshop, but no there was no digital enhancement of any of the other images.
Biochemical analysis of Scp160p
All cell lysis, subcellular fractionation and anti-FLAG affinity isolation procedures were performed as described previously (6,7). Western-blot analyses were also performed as described previously (6) using the monoclonal antibodies M2 (Roche) to detect FLAG-tagged proteins, HA-11 (Covance) to detect hemagglutinin (HA)-tagged proteins and the anti-GFP affinity-purified polyclonal antibody from Sigma (#G1544) to detect GFP-tagged proteins, as cited in the text.
Analysis of in vivo function of the FLAG.SCP160N1 and FLAG.SCP160C1 alleles
The test of SCP160 in vivo function was performed as described previously (8), using the scp160/eap1 double null tester strain JFy4247, transformed with the test plasmids pJF2542, which encodes FLAG.Scp160N1p, and pJF4544, which encodes FLAG.Scp160C1p. As described previously (8), plasmids encoding wild-type Scp160p (pJF2803 for 2μ, pJF4468 for CEN) versus no Scp160p (pJF1102 for 2μ, pJF1096 for CEN) were also included as positive and negative controls, respectively.
RESULTS
Expression of FLAG-tagged Scp160N1p and Scp160C1p in yeast
As a first step toward defining a structure/function map for Scp160p, we generated both an N-terminally truncated allele, FLAG.SCP160N1, in which the first 74 codons, including the potential NES (open inverted triangles, Figure 1A), were removed and replaced by a FLAG tag, and a C-terminally truncated allele, FLAG.SCP160C1, in which the final KH domain was removed. All 14 KH domains (shaded squares, Figure 1A), as well as the potential NLS(filled inverted triangles, Figure 1A), remained intact in FLAG.SCP160N1.
Figure 1. Expression of FLAG.Scp160N1p and FLAG.Scp160C1p in yeast. (A) Cartoons illustrating the domain structures of FLAG.Scp160p, FLAG-Scp160N1p and FLAG-Scp160C1p. The asterisk (*) represents the FLAG epitope, the circle represents the 110 amino acid N-terminal region, the open triangle represents the potential NES and the filled triangle represents the potential NLS. Each box represents one KH domain. (B) Anti-FLAG western-blot analysis of lysates from yeast expressing either untagged Scp160p (control), FLAG-Scp160p, FLAG.Scp160N1p or FLAG-Scp160C1p, as indicated. The filled arrow indicates the full-length and truncated Scp160p proteins; the open arrow indicates a 100 kDa anti-FLAG cross-reacting protein of unknown identity (5,6) that served as an internal control for loading of lanes.
Western-blot analysis of yeast expressing FLAG-tagged wild-type Scp160p, FLAG-tagged Scp160N1p, or FLAG-tagged Scp160C1p demonstrated comparable levels of protein, although as expected, the migration patterns of both truncated proteins differed slightly from that of the full-length protein (Figure 1B, filled arrow). Yeast expressing native, untagged Scp160p served as a negative control, and an 100 kDa cross-reacting band of unknown identity (Figure 1B, open arrow) served as a convenient internal control for the loading of lanes.
Subcellular localization of FLAG.Scp160N1p
To explore the subcellular localization of FLAG.Scp160N1p, we generated a C-terminally GFP-tagged isoform, and integrated the corresponding allele into the normal genomic locus, as described previously for the wild-type allele (6). As reported previously (3,6,27), GFP signal in the wild-type cells was predominantly if not completely cytoplasmic, with clear enrichment around the nuclear periphery, the site of the yeast endoplasmic reticulum (ER). Hoescht staining of these cells confirmed the position of the nucleus, and the corresponding DIC images revealed the overall outline of each cell (Figure 2). Like their wild-type counterparts, yeast expressing the truncated FLAG.Scp160N1p.GFP fusion protein showed GFP signal that was predominantly, if not completely cytoplasmic. No evidence of nuclear accumulation of the truncated protein was seen, suggesting that the potential NES in the deleted region is not normally responsible for the cytoplasmic localization of the full-length protein. These data therefore further suggest that wild-type Scp160p may not normally shuttle between the nucleus and the cytoplasm of yeast.
FLAG.Scp160N1p remains competent to form large RNPs in yeast, but FLAG.Scp160C1p does not
To probe the biochemical interactions of Scp160N1p and Scp160C1p, as compared with wild-type Scp160p, yeast expressing FLAG-tagged isoforms of all three proteins were grown to mid-log phase (OD600 1.0), lysed and the supernatants size-fractionated by S-300 size-exclusion column chromatography. The resultant fractions were subjected to western-blot analysis using the anti-FLAG monoclonal M2 as the primary antibody. As shown in Figure 3 (upper two panels), the majority of signal representing both the wild-type and N-truncated Scp160p proteins migrated at or near the void volume of the column, indicating that both proteins exist as components of large complexes (1300 kDa). As we have reported previously for wild-type Scp160p (5,6), pre-treatment of each lysate with RNase prior to size-fractionation reduced the apparent size of each protein to 500 kDa (Figure 3, bottom two panels), which is consistent with the previously demonstrated migration of purified native Scp160p (5). Interestingly, the C-truncated protein migrated as the smaller size even in the context of untreated lysates (Figure 3, third panel), demonstrating that this protein did not remain competent to form normal RNP complexes.
Figure 3. Size estimation of FLAG.Scp160p, FLAG.Scp160N1p and FLAG.Scp160C1p-containing complexes using S-300 gel filtration column chromatography. Lysates of cells expressing the indicated alleles of FLAG-tagged full-length or truncated SCP160 were subjected to S-300 gel filtration column chromatography, as described in Materials and Methods. The resultant fractions were subjected to western-blot analysis using the anti-FLAG-tag monoclonal antibody M2. Each panel presents the corresponding sections of two gels juxtaposed, one loaded with samples of column elution fractions from 30 to 50 ml, and the other with samples of fractions from 50 to 70 ml. The position of the void volume for this column, representing complexes of 1300 kDa, is indicated. The upper three panels represent cells lysed in the absence of RNase; the lower two panels represent lysates treated with RNase prior to chromatography. Although the results of only single representative experiments are presented, all procedures were repeated at least three times with indistinguishable results.
As a next step to probe the composition of the FLAG.Scp160N1p containing RNP complexes, we applied a previously described anti-FLAG affinity purification procedure (5–8) to isolate these complexes from yeast also expressing HA-tagged Bfr1p. Prior studies identified Bfr1p as one of many proteins that co-purify with Scp160p-containing mRNP complexes in yeast (6). As shown in Figure 4 (upper panels), and as reported previously for full-length FLAG.Scp160p (6), HA.Bfr1p co-purified with FLAG.Scp160N1p. Also like its wild-type counterpart, the association between FLAG.Scp160N1p and HA-Bfr1p was sensitive to RNase pre-treatment (Figure 4, middle set of panels). Finally, no HA-Bfr1p signal was detected following the mock-isolation of complexes from yeast expressing HA-tagged Bfr1p in conjunction with native, untagged Scp160p, demonstrating the specificity of the isolation procedure (Figure 4, lower panels). These data demonstrate that the FLAG.Scp160N1p protein retained the ability to form at least a subset of its normal macromolecular interactions.
Figure 4. FLAG.Scp160N1p associates with HA-Bfr1p in an RNA-dependent manner. Upper two panels: Scp160p-containing complexes were affinity isolated as described in Materials and Methods from lysates of yeast co-expressing FLAG.Scp160N1p and HA.Bfr1p, and subjected to western-blot analysis with both anti-FLAG (upper panel) and anti-HA monoclonal antibodies (bottom panel). As illustrated, although the affinity isolation procedure was directed against the FLAG.Scp160N1p protein alone, HA-Bfr1p protein co-eluted from the affinity column. Middle two panels: Scp160p-containing complexes were affinity isolated and characterized as described above, except that samples were pretreated with RNase prior to the affinity isolation procedure. As illustrated, the association between FLAG.Scp160N1p and HA.Bfr1p was RNase-sensitive. Bottom two panels: as a control for specificity of the isolation and elution procedure, a parallel isolation was performed on lysates of yeast co-expressing HA.Bfr1p with native, untagged Scp160p. As illustrated, although HA.Bfr1p protein was present in the lysates (void and flow through fractions), no HA.Bfr1p signal eluted from the anti-FLAG affinity column. Although the results of only single representative experiments are presented, both test and control experiments were repeated at least three times with indistinguishable results.
Both FLAG.Scp160N1p and FLAG.Scp160C1p are severely compromised in their ability to associate with polyribosomes
To probe the ability of each truncated protein to associate with polyribosomes, soluble cell lysates were size-fractionated by sucrose gradient centrifugation, as described previously (5,6). Lysates from yeast expressing the FLAG-tagged wild-type protein were analyzed in parallel as a control. By their nature, sucrose gradients enable separation of much larger complexes than do S300 columns, allowing the distinction of free RNPs from ribosomal subunits, monosomes and polysomes, as shown in Figure 5. Western-blot analyses of the fractions from these gradients (Figure 5) demonstrated that although the wild-type FLAG.Scp160p protein localized predominantly to the bottom of the gradient, co-fractionating as expected with polyribosomes, the FLAG.Scp160N1p protein localized predominantly to the lighter fractions. Furthermore, the FLAG.Scp160C1p protein localized predominantly to the very top of the gradient, although a small amount of this protein continued to co-fractionate with polyribosomes.
Figure 5. Analysis of FLAG.Scp160p, FLAG.Scp160N1p and FLAG.Scp160C1p-containing complexes using sucrose gradient fractionation. Upper panel: a representative tracing of the UV (254 nm) absorption profile from gradient fractions of wild-type cells, with the positions of the small and large ribosomal subunits, monosomes and polyribosomes indicated. The corresponding profiles from yeast expressing FLAG.Scp160N1p or FLAG.Scp160C1p were essentially indistinguishable. Bottom three panels: lysates of yeast expressing the indicated alleles of SCP160 were subjected to sucrose gradient fractionation, as described in Materials and Methods, followed by western-blot analysis of the fractions using the anti-FLAG antibody M2. Although the results of only single representative experiments are presented, experiments were repeated at least three times with indistinguishable results.
Differential centrifugation of cell lysates to separate soluble from membrane-associated pools also demonstrated a notable distinction between the full-length and truncated Scp160p proteins. As reported previously (4,6,27), wild-type FLAG.Scp160p localized predominantly to the membrane pellet, while both the FLAG.Scp160N1p and FLAG.Scp160C1p proteins were almost evenly distributed between the soluble and membrane-associated fractions (Figure 6A and B). This significant shift of both the FLAG.Scp160N1p and FLAG.Scp160C1p signals from the membrane-associated pool to the soluble pool is fully consistent with our previous observation that treatments which disrupt Scp160p association with polyribosomes also shift signal from the membrane-associated pellet to the soluble fraction (6).
Figure 6. Distribution of FLAG.Scp160p, FLAG.Scp160N1p and FLAG.Scp160C1p between the soluble and membrane-associated subcellular pools. (A) Lysates of cells expressing the indicated alleles of SCP160 were fractionated into total, soluble and membrane-associated pools by differential centrifugation, as described in Materials and Methods, followed by western-blot analysis of the fractions using the anti-FLAG antibody M2. (B) Fraction of Scp160p signal sedimenting with cell membranes. Averages ± SD (n) are plotted for each allele. The values obtained for both truncated proteins were significantly different from the corresponding wild-type value (t-test).
In vivo function of both truncated FLAG.SCP160N1 and FLAG.SCP160C1 alleles is severely compromised
To assess functional capacity of both the FLAG.SCP160N1 and FLAG.SCP160C1 alleles in vivo, we quantified the ability of each to complement scp160/eap1 synthetic lethality in a double null tester strain, as described previously (8). In brief, plasmid-borne copies of each truncated SCP160 allele were introduced into yeast genomically null for both scp160 and eap1, covered by a URA3 maintenance plasmid encoding wild-type SCP160. Following growth for a fixed number of generations in the absence of uracil selection, transformants were plated to medium with and without 5FOA. The fraction of cells that could grow in the presence of 5FOA, which kills only URA3+ cells, versus in its absence, served as an indicator of how many cells had lost their URA3 plasmid. Yeast transformed with either empty plasmid backbone, or plasmid carrying a second wild-type copy of SCP160, served as negative and positive controls, respectively. For comparison, the degree of URA3 plasmid loss in each strain was scaled relative to that seen in the corresponding positive control, which was set to 100%. The final values calculated, therefore, no longer directly reflected plasmid retention or loss, but rather indicated how effectively, relative to the wild-type allele, the test sequence in question could enable cells to remain viable despite loss of their URA3+ SCP160+ maintenance plasmid. Since this maintenance plasmid was the same in all strains, issues of plasmid replication efficiency or distribution, independent of SCP160 sequence, should have been cancelled out. As shown in Table 2, according to this procedure, the FLAG.SCP160N1 allele retained <10% wild-type function, and the corresponding FLAG.SCP160C1 allele retained <5% wild-type function.
DISCUSSION
The experiments presented here were designed to ask three questions: (i) is the N-terminal region of Scp160p, including the potential NES, responsible for the cytoplasmic localization of Scp160p; (ii) is the non-KH domain N-terminal region important for Scp160p macromolecular interaction or function; and (iii) is the final KH domain important for Scp160p macromolecular interaction or function? In effect, we have compared the relative functional contributions of KH and non-KH domain sequences in SCP160. Our results clearly demonstrate that the N-terminal region and NES are not required for cytoplasmic localization of Scp160p, or for assembly of Scp160p into RNPs, although this region is required for association of Scp160p with polyribosomes. In contrast, the final KH domain is required for both RNP assembly and association of Scp160p with polyribosomes. Perhaps most important, both the non-KH domain N-terminus and the C-terminal KH domain are essential for function of Scp160p in vivo. Together, these data underscore the importance of both KH and non-KH domain sequences in Scp160p, and further implicate polyribosome association as a key component of Scp160p function.
Subcellular localization
All studies reported to date addressing the subcellular localization of Scp160p have placed the protein firmly in the cytoplasm, with enrichment around the nuclear envelope/ER (3–8,24,27). Nonetheless, the identification of both putative NLS and NES sequences in Scp160p (4,10) raised the possibility that Scp160p might be a nuclear/cytoplasmic shuttle protein with a steady-state distribution favoring the cytoplasm. The results presented here, demonstrating that loss of the putative NES does not result in nuclear accumulation of the truncated Scp160p, clearly challenge this hypothesis, and further suggest that the remaining putative NLS also may be non-functional. Where in the cell, and at what stage of complex assembly the Scp160p protein finds and associates with its target mRNAs, therefore remains an open question, although presumably these interactions occur within the cytoplasm.
Macromolecular interactions
The data presented here demonstrate that the FLAG.Scp160N1p truncated protein was able to form a subset of its normal biochemical interactions, including association with HA-Bfr1p and assembly into RNPs, although it was severely compromised in its ability to associate with polyribosomes. Whether these data reflect a direct versus indirect role of the N-terminal sequence in polyribosome association remains unclear, as does whether the full N-terminal region is required for polyribosome association, versus a smaller sub-domain. Furthermore, beyond the presence of Bfr1p, the precise nature and composition of the RNP complexes formed by FLAG.Scp160N1p remains to be explored. At minimum, however, the fact that at least some normal biochemical interactions involving FLAG.Scp160N1p remain intact implies that the truncated protein is not globally misfolded. The observation that RNP formation and polyribosome association can be uncoupled further implies that polyribosome association is not an obligate first step in the process of Scp160p mRNP formation.
The data presented here also demonstrate that the FLAG.Scp160C1p truncated protein is compromised with regard to both RNP formation and polyribosome association. This result is striking for at least three reasons. First, it demonstrates either that Scp160p requires all 14 KH domains to function properly, or else that KH14 is special in some way. This is a testable hypothesis that is currently under investigation. Second, although Baum et al. (13) also reported recently that deletion of the final two KH domains from Scp160p disrupted polyribosome association, those authors concluded that sequences within KH13 were essential for this process, and never checked the impact of KH14-loss alone. Our results clearly demonstrate that the loss of KH14 alone is sufficient to disrupt Scp160p-polyribosome association, although the mechanism of that impact remains unclear. While the possibility of partial or global unfolding of the C-truncated protein cannot formally be excluded, the fact that this protein remains largely soluble (Figure 6) and is detected in lysates at levels comparable to that of the wild-type protein (Figure 1B), argues against this possibility. Finally, these results demonstrate that although the N-terminus may be essential for polyribosome association of Scp160p, it is not sufficient.
Function in vivo
One of the most important observations reported here is that both the FLAG.SCP160N1 and FLAG.SCP160C1 were significantly compromised with regard to function in vivo (Table 2). These data demonstrate that although the non-KH domain sequences in Scp160p may be minimal, they are not insignificant. These data further demonstrate that despite the presence of 13 other KH domains, KH14 is essential. Finally, as mentioned earlier, these data implicate polyribosome association as an essential component of Scp160p function in vivo.
ACKNOWLEDGEMENTS
We are grateful to Kristen Riehman, who helped to generate the FLAG-SCP160N1 allele, to Brian Lang, who performed some of the early characterization of this allele, and to Alice Watson, who helped to generate earlier versions of some of the strains and plasmids utilized in this study. This work was supported by funds from the National Science Foundation (Award 0112911 to J.L.F.-K.).
REFERENCES
Burd,G.G. and Dreyfuss,G. ( (1994) ) Conserved structures and diversity of functions or RNA-binding proteins. Science, , 265, , 615–621.
Dreyfuss,G., Kim,V.N. and Kataoka,N. ( (2002) ) Messenger-RNA-binding proteins and the messages they carry. Nature Rev. Mol. Cell Biol., , 3, , 195–205.
Wintersberger,U., Kuhne,C. and Karwan,A. ( (1995) ) Scp160p, a new yeast protein associated with the nuclear membrane and the endoplasmic reticulum, is necessary for maintenance of exact ploidy. Yeast, , 11, , 929–944.
Weber,V., Wernitznig,A., Hager,G., Harata,M., Frank,P. and Wintersberger,U. ( (1997) ) Purification and nucleic-acid-binding properties of a Saccharomyces cerevisiae protein involved in the control of ploidy. Eur. J. Biochem., , 249, , 309–317.
Lang,B.D. and Fridovich-Keil,J.L. ( (2000) ) Scp160p, a multiple KH-domain protein, is a component of mRNP complexes in yeast. Nucleic Acids Res., , 28, , 1576–1584.
Lang,B.D., Li,A.-M., Black-Brewster,H.D. and Fridovich-Keil,J.L. ( (2001) ) The brefeldin A resistance protein Bfr1p is a component of polyribosome-associated mRNP complexes in yeast. Nucleic Acids Res., , 29, , 2567–2574.
Li,A.-M., Watson,A. and Fridovich-Keil,J.L. ( (2003) ) Scp160p associates with specific mRNAs in yeast. Nucleic Acids Res., , 31, , 1830–1837.
Mendelsohn,B.A., Li,A.M., Vargas,C.A., Riehman,K., Watson,A. and Fridovich-Keil,J.L. ( (2003) ) Genetic and biochemical interactions between SCP160 and EAP1 in yeast. Nucleic Acids Res., , 31, , 5838–5847.
Guo,M., Aston,C., Burchett,S.A., Dyke,C., Fields,S., Rajarao,S.J., Uetz,P., Wang,Y., Young,K. and Dohlman,H.G. ( (2003) ) The yeast G protein alpha subunit Gpa1 transmits a signal through an RNA binding effector protein Scp160. Mol. Cells, , 12, , 517–524.
Currie,J.R. and Brown,W.T. ( (1999) ) KH domain-containing proteins of yeast: absence of a fragile X gene homologue. Am. J. Med. Genet., , 84, , 272–276.
Siomi,H., Siomi,M.C., Nussbaum,R.L. and Dreyfuss,G. ( (1993) ) The protein product of the fragile X gene, FMR1, has characteristics of an RNA-binding protein. Cell, , 74, , 291–298.
Gibson,T.J., Thompson,J.D. and Heringa,J. ( (1993) ) The KH domain occurs in a diverse set of RNA-binding proteins that include the antiterminator NusA and is probably involved in binding to nucleic acid. FEBS Lett., , 324, , 361–366.
Baum,S., Bittins,M., Frey,S. and Seedorf,M. ( (2004) ) Asc1p, a WD40-domain containing adaptor protein, is required for the interaction of the RNA-binding protein Scp160p with polysomes. Biochem. J., , 390, , 823–830.
Schmidt,C., Henkel,B., Poschl,E., Zorbas,H., Purschke,W.G., Gloe,T.R. and Muller,P.K. ( (1992) ) Complete cDNA sequence of chicken vigilin, a novel protein with amplified and evolutionary conserved domains. Eur. J. Biochem., , 206, , 625–634.
Plenz,G., Kugler,S., Schnittger,S., Rieder,H., Fonatsch,C. and Muller,P.K. ( (1994) ) The human vigilin gene: identification, chromosomal localization and expression pattern. Hum. Genet., , 93, , 575–582.
Cortes,A., Huertas,D., Fanti,L., Pimpinelli,S., Marsellach,F.X., Pina,B. and Azorin,F. ( (1999) ) DDP1, a single-stranded nucleic acid-binding protein of Drosophila, associates with pericentric heterochromatin and is functionally homologous to the yeast Scp160p, which is involved in the control of cell ploidy. EMBO J., , 18, , 3820–3833.
Cortes,A. and Azorin,F. ( (2000) ) DDP1, a heterochromatin-associated multi-KH-domain protein of Drosophila melanogaster, interacts specifically with centromeric satellite DNA sequences. Mol. Cell. Biol., , 20, , 3860–3869.
Cunningham,K.S., Dodson,R.E., Nagel,M.A., Shapiro,D.J. and Schoenberg,D.R. ( (2000) ) Vigilin binding selectively inhibits cleavage of the vitellogenin mRNA 3' untranslated region by the mRNA endonuclease polysomal ribonuclease 1. Proc. Natl Acad. Sci. USA, , 97, , 12498–12502.
Dodson,R.E. and Shapiro,D.J. ( (1997) ) Vigilin, a ubiquitous protein with 14 K homology domains, is the estrogen-inducible vitellogenin mRNA 3'-untranslated region-binding protein. J. Biol. Chem., , 272, , 12249–12252.
Dodson,R.E. and Shapiro,D.J. ( (1994) ) An estrogen-inducible protein binds specifically to a sequence in the 3' untranslated region of estrogen-stabilized vitellogenin mRNA. Mol. Cell. Biol., , 14, , 3130–3138.
Goolsby,K.M. and Shapiro,D.J. ( (2003) ) RNAi-mediated depletion of the 15 KH domain protein, vigilin, induces death of dividing and non-dividing human cells but does not initially inhibit protein synthesis. Nucleic Acids Res., , 31, , 5644–5653.
Grunweller,A., Purschke,W.G., Kugler,S., Kruse,C. and Muller,P.K. ( (1997) ) Chicken vigilin gene: a distinctive pattern of hypersensitive sites is characteristic for its transcriptional activity. Biochem. J., , 326, , 601–607.
Kugler,S., Grunweller,A., Probst,C., Klinger,M., Muller,P.K. and Kruse,C. ( (1996) ) Vigilin contains a functional nuclear localisation sequence and is present in both the cytoplasm and the nucleus. FEBS Lett., , 382, , 330–334.
Kugler,S., Plenz,G. and Muller,P.K. ( (1996) ) Two additional 5' exons in the human vigilin gene distinguish it from the chicken gene and provide the structural basis for differential routes of gene expression. Eur. J. Biochem., , 238, , 410–417.
Feng,Y., Gutekunst,C.A., Eberhart,D.E., Yi,H., Warren,S.T. and Hersch,S.M. ( (1997) ) Fragile X mental retardation protein: nucleocytoplasmic shuttling and association with somatodendritic ribosomes. J. Neurosci., , 17, , 1539–1547.
Guthrie,C. and Fink,G.R. ( (1991) ) Guide to yeast genetics and molecular biology. Meth. Enzymol., , 194, , 3–21, 281–302.
Frey,S., Pool,M. and Seedorf,M. ( (2001) ) Scp160p, an RNA-binding, polysome-associated protein, localizes to the endoplasmic reticulum of Saccharomyces cerevisiae in a microtubule-dependent manner. J. Biol. Chem., , 276, , 15905–15912.(Ai-min Li, Claudia A. Vargas, Melissa A.)